Julien Amendola, Pierre L. Roubertoux, Bernard Verrier and Jacques Durand. Altered sensorimotor development in SOD1G85R transgenic mice, a model of amyotrophic lateral sclerosis

ADDRESS : CNRS, Marseille Plasticité et Physiopathologie de la Motricité, UMR 6196 CNRS / Université de la Méditerranée 31, Chemin Joseph Aiguier 13402 Marseille Cedex 20, France¹ E-mail : julien.amendola@libertysurf.fr

Amyotrophic lateral sclerosis (ALS) is a late-onset neurodegenerative disease affecting motoneurons in the cortex, brainstem and spinal cord. The SOD1^G85R transgenic mice, expressing familial amyotrophic lateral sclerosis (ALS)-linked mutation G85R in the human gene encoding Cu-Zn superoxide dismutase 1 (SOD1) develop an ALS-like disease characterized by an extremely rapid clinical course resulting from the loss of motor neurons. Despite several cellular mecanisms related to the degeneration have been proposed, the early alterations triggering the toxic pathways remain poorly documented. We addressed the question of whether the SOD1 mutation affects sensorimotor processes during the maturation of spinal motor networks. Behavioural tests revealed that appearance of some sensorimotor reflexes are significantly shifted in SOD1^G85R newborn pups compared to wild-type (WT) mice (C57Bl/6J). Furthermore, following bath-application of N-methyl-DL-aspartate (NMA) and serotonin (5-HT), the motor output extracellularly recorded from lumbo-sacral ventral roots in an in vitro brainstem/spinal cord preparation obtained from newborn mice demonstrated a hypoexcitability in SOD1^G85R lumbar motor networks. The results also suggests that sacral networks in which some motoneurons are spared in ALS are not affected in this postnatal period. Taken together, these results strongly support the idea that the G85R mutation affects the maturation of lumbar spinal motor networks in which motoneurons degenerate massively in adult animals.

Juko Ando¹, Ryoko Nakajima¹, Yutaka Ono ², Kimio Yoshimura³, Nobuhiko Kijima⁴, Kato Rumi Price⁵. Genetic influences on smoking and drinking behaviors in Japanese adolescence and young adulthood: A twin study ⁶

¹ Faculty of Letters, Keio University, Minato-ku Tokyo, Japan, ² Health Center, Keio University, Yokohama Kanagawa, Japan, ³ National Cancer Center, Chuo-ku Tokyo, Japan, ⁴ Psychological Laboratory, Keio University, Yokohama Kanagawa, Japan, ⁵ Washington University School of Medicine, St. Louis, USA, 6 Supported by a Grant-in Aid for Scientific Research (A) from the Ministry of Education, Science, Sports and Culture.

Address : :Faculty of Letters, Keio University, 2-15-45, Mita, Minatob-ku, Tokyo, 108-8345, Japan Telehone: 81 3 3453 4511 Fax: 81 3 5427 1578 Email: juko@msa.biglobe.ne.jp

Genetic influences on smoking and drinking behaviors were investigated by 335 Japanese twin pairs (157 MZf, 41 DZf, 72 MZm, 24 DZm, and 41 DZo; mean age = 20.4 yrs (SD=4.2)). Smoking status (current smoker / past smoker / non smoker) and frequency of drinking showed greater MZ similarity than DZ similarity (rMZ=.46 and rDZ=.24 for smoking status; rMZ=.59 and rDZ=.16 for frequency of drinking), indicating genetic contribution. For smoking status, however, there might be gender difference indicating that genetic contribution is found only for female (rMZf=.53, rDZf=.07, rMZm=.30, rDZm=.44). For frequency of drinking, both female and male showed substantial genetic influences (rMZf=.52, rDZf=.14, rMZm=.72, rDZm=.45, rDZo=.14), and these genetic influences were found even for minors(under 20 yrs old) who are not allowed to drink legally. There is no genetic comobidity between smoking status and drinking frequency.

Andrey P. Anokhin¹, Andrew C. Heath¹, and Erin Myers¹. Genetic influences on neurocognitive mechanisms of inhibitory control: a twin study of event-related brain potentials (ERPs) in a Go/No-Go task².

¹Washington University School of Medicine, St. Louis MO USA, ²Supported by the grants DA00421 from the National Institute on Drug Abuse and a pilot grant from the Missouri Alcoholism Research Center (P50 AA11998).

Address : Washington University School of Medicine, Department of Psychiatry, 18 S. Kingshighway, Suite 2T, St.Louis, MO 63108 USA Telephone: 314-286-2201; FAX: 314-286-0092; email: andrey@matlock.wustl.edu

Inhibition of prepotent responses plays a key role in cognitive control of goal-directed behavior and can be studied experimentally using the Go/No-Go paradigm which requires a speeded response to the Go stimuli and withholding a prepotent response when a No-Go stimulus is presented. Response inhibition in Go-NoGo tasks elicits a distinct mid-frontal ERP component, the N2, localized by recent electrophysiological and neuroimaging studies to the anterior cingulate cortex, and a strong enhancement of the frontal P3 component known as "P3 anteriorization". The N2 effect is believed to reflect the processing by the anterior cingulate of the conflict between competing but incompatible action tendencies. We assessed heritability of the No-Go N2 and the succeeding positive P3 component in 194 young female twins (52 monozygotic and 45 dizygotic pairs) who completed a cued version of the Continuous Performance Test. The ERPs were computed separately for Go and No-Go trials. Genetic model-fitting analysis showed that about 60% of variance in the amplitude of No-Go N2 and P3 components can be attributed to genetic factors. The results suggest that frontal No-Go N2 and P3 components are indicative of genetically transmitted individual differences in the neural substrates of conflict monitoring and response inhibition. These electrophysiological markers can potentially serve as endophenotypes for genetic studies of psychopathologies characterized by executive deficits and behavioral disinhibition.

Andrey P. Anokhin¹, Andrew C. Heath¹, and Erin Myers¹. Heritability of Behavioral Approach and Inhibition Systems (BIS/BAS) scales in twins².

¹Washington University School of Medicine, St. Louis MO USA, ²Supported by the grants DA00421 from the National Institute on Drug Abuse and a pilot grant from the Missouri Alcoholism Research Center (P50 AA11998).

Address : Washington University School of Medicine, Department of Psychiatry, 18 S. Kingshighway, Suite 2T, St.Louis, MO 63108 USA

Telephone: 314-286-2201; FAX: 314-286-0092; email: andrey@matlock.wustl.edu

We examined the genetic and environmental etiology of individual differences in the strength of hypothesized Behavioral Approach and Inhibition Systems (BAS and BIS) based on work by Gray (J.A.Gray, 1990, Cognition and Emotion, 4, 269-288). A modified self-report measure of BIS/BAS (C.S.Carver and T.L.White, 1994, J. Pers. Soc. Psychol. 67, 319-333) was administered to 212 young adult female twins (age 18-28) including 55 MZ and 51 DZ pairs. A biometrical genetic analysis using structural equation modeling showed significant heritability of BIS and BAS scores, suggesting that about 50% of variance in both measures can be explained by genetic factors. The balance between the two systems as measured by BAS-BIS difference score was also heritable. Two of the three individual subscales composing the BAS scale, Fun Seeking and Drive, showed significant heritability, whereas Reward Responsiveness did not. These preliminary results suggest a substantial contribution of genetic factors to individual differences in the sensitivity to signals of reward and punishment as assessed by BIS/BAS scales.

Laura A. Baker, Adrian Raine¹ and Kristen Jacobson. Genetic and environmental bases of pre-adolescent antisocial behavior: A multi-trait multi-method twin study.

¹Department of Psychology, University of Southern California, Los Angeles, CA USA. NIMH #MH58354

Address : University Park Campus, Los Angeles, CA 90089 Telephone: 213 740 2261 Fax: 213 746 9082 E-mail: lbaker@usc.edu

This paper describes the first wave of assessment in a new twin study of normal variation in antisocial and aggressive behavior (ASB). Data are presented from 600 twin pairs (both male and female) measured at age 9-10 years old, and their primary caregivers (over 90% biological mothers). Measures of ASB and aggression include symptom counts for conduct disorder and oppositional defiant disorder, as well as ratings of proactive and reactive aggression, relational aggression, CBCL scales for delinquency and aggression, a child psychopathy checklist, and a child delinquency interview. A multi-informant approach is used, based on child self reports, as well as teacher and caregiver ratings. Boys appeared significantly more aggressive and antisocial than girls for all measures across raters. The various ASB measures were moderately correlated within informants, but less so across informants, suggesting the possibility of generalized antisocial behavior factor with significant variation across raters. First principal component measures of the ASB measures within each rater all showed significant genetic and shared environmental influences in both boys and girls. Heritability estimates for individual ASB measures varied considerably, however, both within and between raters, suggesting specificity of etiologies for different definitions of ASB and the informants who provide the ASB ratings.

David A. Blizard,¹ David J. Vandenbergh,^{1, 2} Arimantas Lionikas,¹ Glenn S. Gerhard³, James W. Griffith,⁴ Laura C. Klein ^1,2, Joseph T. Stout,¹ Holly A. Mack^1,2, Joan M. Lakoski,⁵ Lars Larsson,⁶ Jeanne M. Spicer¹, George P. Vogler^1,2 and Gerald E. McClearn^1,2 . QTL influencing standard deviation of heart rate and blood pressure in the mouse.

¹Center for Developmental & Health Genetics,and ²Department of Biobehavioral Health, The Pennsylvania State University, University Park, Pennsylvania USA, ³Geisinger Medical Center, Weis Center for Research, Danville, Pennsylvania USA, ⁴ Department of Comparative Medicine, Pennsylvania State College of Medicine, Hershey, Pennsylvania USA, ⁵University of Pittsburgh, Department of Health Sciences, Pittsburgh, Pennsylvania USA, ⁶Department of Clinical Neurophysiology, Uppsala University, Uppsala, Sweden.

This work was supported by grants P01 AG14731 and T32 AG00276 from the National Institute on Aging of the National Institutes of Health.

ADDRESS : Center for Developmental and Health Genetics, 201, Research Bldg D, Pennsylvania State University, University Park, PA, 16802.

Telephone: 814-865-3429 Fax 814-863-4768 E-mail : dab22@psu.edu

Using indirect tail-cuff determinations systolic blood pressure and heart rate were recorded from nearly 400 male and female F₂s derived from a cross of C57BL/6J and DBA/2J mice and from 22 BXD RI strains at approximately 150 days of age. Recordings were obtained on 7 days with 3 blocks of 8 measurements per day for a maximum of 154 readings per F₂ mouse (5 days of readings in RIs). Standard deviations (SDs) were calculated for each 8 trial block for each mouse and mean SD for each mouse over all days used as primary datum for QTL analysis. Analysis with QTL Cartographer revealed the presence of three QTL contributing to the magnitude of within animal HR and BP SDs (SBP, peak at 33 cMs on Chr 9, 1-LOD support interval, 27-41 cms; HR, two peaks at 28 and 50 cMs on Chr 8 with support intervals, 17-37 and 39-60 cM, respectively; peak at 40 cM on Chr18, support interval, 18-55 cM). The HR SD on Chr 8 (proximal QTL) was verified by analysis of RI strains. Analysis of variance of a struggling index (SI) at markers near the QTL peaks revealed a significant effect of D18Mit123 (the marker at the HR SD peak on Chr 18; F, 2, 368 = 7.2, p<0.001). Animals homozygous for the B6 allele at this marker had higher SI than heterozygotes or D2D2 homozygotes, exactly the same pattern as that exhibited for the HR SD QTL on Chr 18. These QTL influencing within animal HR and SBP variability were on different chromosomes than QTL that influenced mean level of these functions (reported elsewhere). Physiological mechanisms that could account for differences in HR and BP variability will be discussed as well as possible pathophysiological consequences of individual differences in the magnitude of cardiovascular variability

Dorett I. Boomsma¹, C.E.M. van Beijsterveldt¹, E.M. Derks¹, M. Bartels¹, and J.J. Hudziak². Shared environmental factors involved in Anxiety/Depression during childhood: real or inflated by rater bias? A study from The Netherlands Twin Register.³

¹Department of Biological Psychology, Vrije Universiteit, Amsterdam, The Netherlands, ²Department of Psychiatry and Medicine (Division of Human Genetics), Center for Children, Youth and Families, and University of Vermont, College of Medicine, Burlington, USA, ³Supported by NWO Spinoza Grant numbers SPI-56-464 (Boomsma, P.I.) and by NIMH Grant number MH58799 (Hudziak, P.I.)

Address : Vrije Universiteit, Department of Biological Psychology, van der Boechorststraat 1, 1081 BT Amsterdam, The Netherlands, Telephone: 0031(0)20-4448787 Fax: 0031(0)20-4448832 Email: dorret@psy.vu.nl

In our longitudinal studies of problem behavior we find evidence for an increasing role of shared environment influencing Anxiety/Depression (A/D) during childhood. Because these results were based on a single rater, it is possible that rater bias may have inflated the role of shared environmental influences. To disentangle the effects of rater bias and unreliability from that of shared and nonshared environmental factors, we used a psychometric model incorporating both paternal and maternal ratings. The psychometric model assumes that part of the assessment of the child=s behavior is rater-specific and part is common to both raters. The part that is common to both parents contains only reliable variance. Rater bias can only confound the shared environmental influences specific to one parent. At ages 3, 5, 7, 10 and 12 years indices of A/D were obtained from maternal and paternal CBCL ratings (at age 5, anxiety was obtained from Devereux Child Behavior rating scale items) as part of a large ongoing longitudinal study of the Netherlands Twin Register (NTR). At ages 3 and 5 years data are available for around 9000 twin pairs; at 7, 10 and 12 years for around 7300, 4400 and 2400 pairs. Rater-specific shared environment, including rater bias, accounted for 0-19% of the total variance and was nearly absent at age 3 and strongest at age 7. When only the reliable part of A/D is taken into account, the role of environmental factors decreased at all ages, but the initial pattern of increasing influence during childhood remained. Heritability became smaller (around 70% at age 3 and around 40% at age 12 years) but remained the most important factor in explaining the variance of A/D across ages.

Dorret I. Boomsma¹, Gonneke Willemsen¹, Eco J.C. de Geus¹, Louise C. Hawkley², John T. Cacioppo², Danielle Posthuma¹. Genetics of Loneliness: A study from The Netherlands Twin Register.

¹Vrije Universiteit, Amsterdam, The Netherlands; ²University of Chicago, USA. This research was supported by NWO grants 575-25-006, 904-61-090,

985-10-002, 904-61-193 (Netherlands Organization for Scientific Research) and the National Institute of Aging Grant No. PO1 AG18911 (Social isolation, loneliness, health, and the aging process). Danielle Posthuma was supported by GenomEUtwin, European Union Contract No. QLG2-CT-2002-01254. Genotyping was carried out by the Center for Medical Genetics in Marshfield (research.marshfieldclinic.org/ genetics/).

Address : Biological Psychology, Vrije Universiteit, Van der Boechorststraat 1,

1081BT Amsterdam, The Netherlands, fax 31-20-4448832, Telephone 31-20-4448787, Email: dorret@psy.vu.nl

A measure of loneliness was obtained by factor analyses of YASR items (Achenbach, 1990, Young Adult Self Report. Univ Vermont, Dept Psychiatry, Burlington, VT). YASR items were assessed longitudinally in participants in the survey studies of the Netherlands Twin Register. The longitudinal stability of the loneliness measure ranged from 0.4 to 0.62 (2 to 9 year stability) in both males and females. Data on loneliness from 7,665 adolescent and (young) adult Dutch twins (average age 24 years) were analyzed with genetic structural equation models. The estimate of the genetic contribution to variation in loneliness was 47%, with the remaining variance explained by unique environmental factors. There was no evidence for sex differences in genetic architecture. A complete genome scan in a subsample of participants found evidence for 2 QTLs (LOD scores of > 3 and > 2).

Tanya M. M. Button,¹ Jane Scourfield,² Neilson Martin,³ Shaun Purcell^1,4 and Peter McGuffin^1. Family dysfunction interacts with genes in the causation of antisocial symptoms⁵. "

¹ Social, Genetic and Developmental Psychiatry Centre, Institute of Psychiatry, King=s College, London UK, ² Department of Psychological Medicine, University of Wales College of Medicine, Cardiff UK, ³ School of Psychology, Curtin University, Perth, Western Australia, ⁴ Whitehead Institute, MIT, Cambridge, MA, USA, ⁵ Supported by the MRC via a training fellowship to JS and studentships to NM and TB.

MAIL: Social, Genetic, and Developmental Research Centre, Institute of Psychiatry, King's College London, De Crespigny Park, Denmark Hill, London SE5 8AF, England Telephone: +44 (0) 207 848 5415 Fax: +44 (0) 207 848 0575 E.mail: t.button@iop.kcl.ac.uk

There is emerging evidence of gene-environment interaction effects on antisocial behavior, both from adoption studies and from a study using a measured genotype. An association between non-violent family dysfunction and antisocial behavior has also been reported, although not in the context of gene-environment interaction studies. The aim of this study was to examine the interaction of genes and family dysfunction in contributing to antisocial behavior in young people. Parents of 278 monozygotic and 378 dizygotic twin pairs, aged 5-18, from the CaStANET birth cohort twin register were questioned about zygosity, antisocial behavior and family environment. Using structural equation modelling we tested for main and interactive effects of genes and family dysfunction modelled as an environmental >moderator variable=. Both main and gene-environment interaction effects were highly significant . It was concluded that a risk genotype conferring susceptibility to family dysfunction is responsible for most of the variance in antisocial symptoms in childhood and adolescence.

Desmond Campbell, Harvey E. Wickham, Pak C. Sham. Simulation-based estimation of genetic, environmental and overall liability scores from pedigree affection data for multifactorial disorders.

SGDP, Institute of Psychiatry, KCL, London, UK.

Address: Room C0.29, SGDP Centre, Institute of Psychiatry, De Crespigny Park, London SE5 8AF Telephone +44 (0) 20 7848 0236 Fax: +44 (0) 20 7848 0866 Email: D.Campbell@iop.kcl.ac.uk

In contrast to Mendelian disorders, there is presumed to be an underlying continuum of liability to categorically defined multifactorial polygenic disorders (D. S. Falconer, T. F. C. Mackay, 1989, Introduction to Quantitative Genetics). For such disorders, the use of diagnostic categories when searching for biological disease markers leads to a loss of power. This is due to a loss of information; pedigree members can be qualitatively healthy despite being carriers of disease susceptibility alleles. A more powerful approach would be to use continuous measures reflecting underlying genetic, environmental and overall liabilities to the disorder. A software program has been developed which, given a polygenic disease=s heritability, estimates these liabilities for the individuals in pedigrees with the disease by Gibb=s sampling. The program models age of onset effects and different prevalences in subpopulations (defined by age groups, gender, marital status, or any known risk factor) in a user-specified way. The resulting liability estimates can be related to endophenotypes (or putative risk factors) by a generalized least-squares regression framework. A general hypothesis testing framework allows the user to specify any linear null-hypothesis based on the estimated regression coefficients and their covariance matrix. This facilitates determining whether the discovered liability-risk factor relationships are statistically significant. Program operation has been verified using simulated pedigree data.

Gregory Carey¹. Cholesky Problems.

¹ Department of Psychology and Institute for Behavioral Genetics, University of Colorado, Boulder CO USA. Supported in part by NIH grant M01 RR00051.

Address : Department of Psychology, University of Colorado, Boulder CO, USA 80309-0345 Telephone: 303-492-1658 Fax: 303-492-2967 Email: gregory.carey@colorado.edu

Behavioral geneticists commonly parameterize a genetic or environmental covariance matrix as the product of a nonsingular, lower diagonal matrix postmultiplied by its transposeCa technique commonly referred to as Afitting a Cholesky.@ Here, simulations demonstrate that this procedure: (1) may not produce likelihood ratio test statistics that are distributed as a ?²; (2) if the distribution of the test statistic appears to be ?², then the degrees of freedom are not always the difference between the number of parameters in the general model less the number of parameters in the constrained model; and (3) confidence limits on parameters may be inaccurate. It is hypothesized that the problem is related to the fact that the Cholesky parameterization requires that the covariance matrix formed by its product be positive definite. Even though a population covariance matrix must be positive definite, the combination of sampling error and the derivedCas opposed to directly observedCnature of some matrices in behavioral genetics allow matrices that are not positive definite. Hence, fitting a Cholesky constrains the area of search and compromises maximum likelihood theory. Until the reason for this phenomenon is understood and a satisfactory solution is developed, the Cholesky parameterization should be used with caution. An alternate strategy of fitting a lower diagonal matrix to data that avoids the Cholelsky problem is proposed.

Michèle Carlier¹, Silvia Stefanini², Arianna Bello² and Virginia Volterra³. Prehension and Laterality in Children with Williams-Beuren Syndrome

¹ Laboratory PsyCLÉ, University of Provence, Aix en Provence, France, ²Department of Neuroscience, University of Parma, Italy, ³Institute of Cognitive Science and Technologies, National Research Council (CNR), Roma, Italy.

The study was supported by the FIRB/MIUR AAction and Perception in the construction of the cognitive world@ (RBNE01SZB4), the ESF EUROCORES program AThe Origin of Man, Language and Languages@, and the Fondation Jérôme Lejeune.

Address M. Carlier: Laboratory PsyCLÉ, UFR PSE, University of Provence, 29 Avenue Robert Schuman13621 Aix en Provence, France E-mail: michele.carlier@up.univ-aix.fr

Individuals with Williams-Beuren syndrome (WBS) show a characteristic cognitive profile with weakness in visuo-spatial cognition assessed for example with the Block Design from the Wechsler Intelligence Scales. Spatial working memory, flexibility in the use of spatial properties and in the hierarchical organization of objects are supposed to be involved in this task. However a special difficulty in fine motor abilities and in the development of object prehension could explain part of the low performance of these children. Here we investigate object prehension abilities and manual laterality in children with WBS, aspects which were not clearly described by previous studies. Prehension abilities and degree of laterality was investigated in children with WBS. Nine children with WBS, within a restricted age range (9-12 years) participated to the study. All were right handers for the writing hand. Perceptual and motor abilities were assessed with the Developmental Test of Visual-Motor Integration (VMI) ; manual dexterity was assessed with the Movement Assessment Battery for Children (M-ABC). Two tasks of the M-ABC afford a special opportunity to analyze the type of prehension adopted by children and to evaluate the direction and degree of manual laterality. Fourteen typically developing children constituted the comparison groups (all were right handers for the writing hand) : 6 were matched with WBS children for mental age and 8 for chronological age. All children except one in the group matched for mental age were right-handers (i.e., the right hand was better than the left). Children with WBS performed like younger typically developing children as for degree of laterality, but some of them adopted a peculiar type of prehension, never observed in the control groups. These findings on atypical prehension could provide a partial explanation of visual-motor difficulties exhibited by the children.

Xiangning Chen^1,3, Jaqueline M. Vink², Michael C. Neal¹, Kenneth S. Kendler¹ and Dorret I. Boomsma^2,3. A candidate study of the EPAC gene for nicotine dependence using a Dutch twin sample

¹ Virginia Institute for Psychiatric and Behavioral Genetics and Department of Psychiatry, USA, ² Department of Biological Psychiatry, Vrije Universiteit Amsterdam, 1081 BT Amsterdam, The Netherlands, ³ Corresponding authors. Email: xchen@vcu.edu (XC) and DI.Boomsma@fpp1.psy.vu.nl (DIB)

Address : Virginia Institute for Psychiatric and Behavioral Genetics and Department of Psychiatry, Virginia Commonwealth University, 800 E. Leigh Street, Suite 1-110, Richmond, VA 23298, USA. Telephone: 804 828 8124 Fax: 804 828 1471 Email: xchen@vcu.edu

The exchange protein directly activated by cAMP (EPAC) is a rap1 guanine-nucleotide exchange factor that involves in inter- and intracellular signal transduction pathway through the regulation of GTPase activity. It has been reported to have altered gene expression in rat in a microarray study after nicotine administration. In a case-control study, we recently investigated the human ortholog of the EPAC gene and found a modest association between the gene and nicotine dependence in a Caucasian sample collected in Virginia. In the Virginia sample we typed 5 SNPs and found that 3 SNPs and a haplotype were associated with nicotine dependence. To verifying our results, we are currently performing a replication study using a Dutch twin sample. The Dutch sample was ascertained similarly using the Fagerstrom Tolerance Questionnaires and other instruments as the Virginia sample. We are typing the 3 associated SNPs and will perform single marker and haplotype analyses to evaluate their association with nicotine dependence. The results of this replication will be presented in the meeting.

R. P. Corley¹, M. C. Stallings¹, J. K. Hewitt¹, S. E. Young¹, and J. Zeiger¹. Robustness of Genome Scan Results on Adolescent Dependence Vulnerability from the Colorado CADD².

¹Institute for Behavioral Genetics, University of Colorado, Boulder, CO USA, ²Supported by grants DA-05131 and DA-11015, HD-010333, HD-36773, and MH-43899

Address: Institute for Behavioral Genetics, Campus Box 447, University of Colorado, Boulder, CO 80309-0447 USA Telephone: 303 492 5189 Fax: 303 492 8063 Email: Robin.Corley@colorado.EDU

We previously presented (R. P. Corley, M. C. Stallings, J. K. Hewitt, & S. E. Young, 2001, Behavior Genetics, 31, 450) a comparative behavioral genetic analysis of ten potential phenotypic definitions of dependence vulnerability across multiple substances during adolescence and young adulthood. Based on the results from 3676 interviews of community samples of adopted & non-adopted siblings and twins, we chose one phenotypic definition on an a priori basis as our best initial choice of phenotype for a genome wide search for quantitative trait loci influencing substance dependence vulnerability in adolescent treatment probands and their siblings (M. C. Stallings, et al., 2003, Drug and Alcohol Dependence, 70, 295-307), in which regions of interest were identified on chromosomes 3 and 9. Through 2003, an additional 1613 interviews with 12- to 25- year olds have been completed from community samples. We use our combined community samples to explore the additive genetic contribution to six additional phenotypic definitions of dependence vulnerability, the similarity of age and gender effects in the additional adolescents with those found for the original sample, the effect of controlling for age and gender effects through multiple threshold and standard regression approaches, and whether comparable peaks on Chromosomes 3 and 9 are found across the spectrum of dependence vulnerability phenotypes in the treatment sample.

Victoria E. Cosgrove^1,2, Blake Buhlig^1,2, Soo Hyun Rhee^1,2, Susan E. Young¹, Andrew Smolen¹, Robin P. Corley¹, John K. Hewitt^1,2. Sibling-based association analyses of the serotonin transporter polymorphism and Generalized Anxiety Disorder in adolescents³.

¹Institute for Behavioral Genetics and ²Department of Psychology, University of Colorado, Boulder, CO, ³Supported by NIH grants MH-01865, DA-11015, HD-18426, MH-43899, and DA-13956 and funding from the John D. and Catherine T. MacArthur Foundation.

Address : Institute for Behavioral Genetics Campus Box 447 University of Colorado, Boulder, CO 80309-0447 Telephone: (303) 735-2428 Fax: (303) 492-8063 Email: victoria.cosgrove@colorado.edu

Generalized Anxiety Disorder (GAD) is a commonly occurring anxiety disorder with a lifetime prevalence rate of 4-7%. Early-onset GAD tends to be more chronic and have associations with anxious personality. The serotonergic system has long been implicated in the regulation of mood and anxiety. Specific research has targeted the serotonin transporter (5-HTT), which functions to stop serotonin=s synaptic action by aiding in its reuptake into the presynaptic membrane. 5-HTT has a 44 base-pair repeat element (5-HTTLPR) in the 5= region of the gene. Its polymorphism leads to long (L, 528bp) and short (S, 484 bp) alleles; S is dominant over L and has half its transcriptional activity. Recent research has demonstrated association between anxiety and the S allele, however the specific association between 5-HTTLPR and early-onset GAD in adolescents has not yet been examined. This study seeks to investigate the possible association of 5-HTTLPR variants to GAD symptom counts in adolescents in order to investigate the relationship of the polymorphism to early-onset GAD. Participants were 711 12-year-old children from the Colorado Longitudinal Twin Study (LTS). GAD symptom counts were derived from the Diagnostic Interview Schedule for Children (DISC-IV). Controlling for population stratification, a sibling-based methodology for estimating allelic association with quantitative traits was applied (D.W. Fulker, S.S. Cherny, P.C. Sham, J.K Hewitt, 1999, Am. J. Hum. Genet. 64, 259-267). No association was found between 5-HTTLPR and DISC-IV GAD symptom counts in adolescents. Although results did not support the hypothesis that the 5-HTTLPR contributes to early-onset GAD symptomatology, it remains a worthwhile target of investigation since its role in anxious and depressive disorders is still unclear.

François X. Coudéo, Claire Mignot5, Stanislas Lyonneto and Arnold Munnicho. Academic impairment is the most frequent complication of neurofibromatose type-1(NF1) in children;.

oDépartement de Génétique, Hôpital Necker-Enfants Malades, Paris, France, 5Centre de Pédiatrie * les Collines de Cuques +, Résidence les Collines de Cuques, 6 av. de l=Armée d=Afrique, 13100 Aix-en-Provence, ;Supported in part by a grant from Association * Neurofibromatoses et Recklinghausen +, 34 Vieux chemin de Grenade, 31700 Blagnac, France

Address : Centre de Pédiatrie * les Collines de Cuques +, Résidence les Collines de Cuques, 6 av. de l=Armée d=Afrique, 13100 Aix-en-Provence, Telephone : 0618421316 Fax : 0442230613 Email : francoisxavier.coude@club-internet.fr

Neurofibromatosis type-1 (NF1) is a common genetic disorder associated with a variety of medical complications, cognitive impairments, and behavioral problems. One hundred and sixteen patients with NF1 (62 males, 54 females; mean age 12.4 years, SD 2.3) were studied in terms of complications and learning impairment (one or more grade repetitions or school exclusions). Seventy of 116 patients had significant learning impairment. Classical complications were present in 53 patients including the three most frequent complications in children, namely severe scoliosis (19), plexiform neurofibroma (16) and precocious puberty (14). There was no sex predominance except for plexidorm neurofibroma (11 male versus 5 female). Learning impairment predominated in first grades and was significantly sex dependent. Academic impairment is the most frequent complication in NF1. Because diagnosis is often a crucial problem in young children with six or more isolated café-au-lait spots, early developmental and bevahioral assessments might constitute a crucial diagnosis tool for these at risk children.

Brian M. D’Onofrio¹ , Eric Turkheimer¹, Robert E. Emery¹, Hermine H. Maes², Judy Silberg², and Lindon J. Eaves². The association between marital instability and offspring substance-use and emotional problems: A children of twins.

¹ Psychology Department, University of Virginia, ² Virginia Institute for Psychiatric and Behavior Genetics, Virginia Commonwealth University, ³ The analyses were supported by grants from the National Institute of Mental Health (MH67300), William T. Grant Foundation, and John Templeton Foundation. Data collection was supported by Grants GM-30250, AG-04954, AA-06781, MH-40828, and HL-48148 from the National Institutes of Health and a gift from RJR Nabisco.

Address : Department of Psychology University of Virginia, PO Box 400400 Charlottesville, VA 22904-4400 Telephone: 434-982-4750 Fax: 434-982-4766 email: bmd8q@virginia.edu

Although the association between parental marital instability and psychological problems in young adults has been well documented, the developmental mechanisms responsible for the statistical relations have remained unclear. The current study utilized the children of twins design to explore whether genetic or shared environmental factors confound the intergenerational associations related to marital instability in a sample of twins and their offspring from Virginia and the American Association of Retired Persons. A univariate twin analysis indicated that genetic factors contributed to variation in marital instability. Comparisons of offspring from monozygotic and dizygotic twins discordant for divorce, while also statistically controlling for parental characteristics, suggested that environmental factors specifically related to marital instability are associated with higher levels of alcohol problems and risk of smoking in the offspring. These findings are consistent with a quasi-causal theory of the effects of divorce on children. In contrast, selection factors, including genetic confounds, accounted for the increased risk of emotional problems and depression in offspring from divorced families. The study illustrates that passive gene-environment correlation must be considered when studying family risk factors because erroneous conclusions would have been reached if the analyses only relied on traditional, statistical approaches (e.g. analysis of covariance).

Cindy M. de Frias^{1, 2}, Kristina Annerbrink³, Lars Westberg³, Elias Eriksson³, Rolf Adolfsson⁴, and Lars-Göran Nilsson^{1, 5}. COMT gene polymorphism is associated with cognitive functioning in adulthood and old age.

¹Department of Psychology, Stockholm University, Stockholm, Sweden, ²Division of Geriatric Epidemiology, Karolinska Institute, Stockholm, Sweden, ³Department of Pharmacology, Göteborg University, Göteborg, Sweden, ⁴Department of Clinical Psychiatry, Division of Psychiatry, Umeå University, Umeå, Sweden, ⁵Center for Advanced Study, Norwegian Academy of Science and Letters, Oslo, Norway.

Address: Department of Psychology, Stockholm University, 106 91 Stockholm, Sweden Telephone: +46 8 163915 Fax: +46 8 159342 Email: cdefrias@psychology.su.se

Variation in cognitive performance is to a large extent explained by genes. In the prefrontal cortex, the catechol O-methyltransferase (COMT) gene is essential in the metabolic degradation of dopamine, a neurotransmitter implicated in cognitive functions. The present study examined the effect of a polymorphism in the COMT gene on individual differences and changes in memory and executive functions in adulthood and old age. Tests assessing episodic and semantic memory and executive functions were administered to 286 men (initially aged 35-85 years) from a random sample of the population (i.e., the Betula prospective cohort study; L-G. Nilsson, R. Adolfsson, L. Bäckman, C. M. de Frias, B. Molander, and L. Nyberg, in press, Aging, Neuropsychol, and Cogn.) at two occasions followed over a 5-year period. Carriers of the Met/Met genotype (with low enzyme activity) performed better on episodic memory, semantic memory, visuospatial ability, and verbal fluency as compared to carriers of the Val allele (with higher enzyme activity). Division of episodic memory into its recall and recognition components showed that the difference was specific to episodic recall, not recognition tasks; an effect that was observed across three age groups (middle-age, young-old, and old-old adults) and over a 5-year period. The COMT gene is a plausible candidate gene for cognitive functioning in adulthood and old age.

John C. DeFries¹ and Sally J. Wadsworth¹. Colorado Twin Study of Reading Disabilities²

¹Institute for Behavioral Genetics, University of Colorado, Boulder CO, USA, ²Supported by NICHD Center Grant HD-27802.

Address : Institute for Behavioral Genetics, University of Colorado, 447 UCB, Boulder CO 80309-0447 Telephone: 303 492 2839 Fax: 303 492 8063 Email: John.DeFries@Colorado.EDU

A major goal of the Colorado Learning Disabilities Research Center is to assess the genetic and environmental etiologies of reading deficits, Attention-Deficit/Hyperactivity Disorder (ADHD), and their comorbidity. To accomplish this goal, twin pairs are systematically ascertained from schools in Colorado and parental permission is sought to review the children=s school records. If either member of a pair manifests a school history of reading difficulties (e.g., low reading achievement test scores) or ADHD symptoms, both members of the pair are invited to complete an extensive test battery, including the Peabody Individual Achievement Test (PIAT). Employing discriminant weights estimated from an analysis of PIAT Reading Recognition, Reading Comprehension, and Spelling data, a discriminant function score is computed for each subject. Criteria for reading disability (RD) include being classified as affected by the discriminant score, having a verbal or performance IQ of at least 90, no evidence of neurological problems, and no uncorrected visual or auditory acuity deficits. As of November 30, 2003, 269 monozygotic and 216 same-sex dizygotic twin pairs have been ascertained in which at least one member of each pair meets these criteria. The probandwise concordance rates for RD in these twin pairs are 0.64 and 0.34, respectively (p = 1.05 H 10^-10). When the DeFries-Fulker basic regression model (J. C. DeFries and D.W. Fulker, 1985, Beh. Genet., 15, 467-473) was fitted to their discriminant function score data, h²_g = 0.57 (p = 1.67 H 10^-14), indicating that reading difficulties are due substantially to genetic influences. Results regarding the differential etiology of RD as a function of gender, comorbid RD and ADHD, bivariate linkage analyses for RD and ADHD symptoms, and association and candidate-gene analyses obtained from this and other ongoing studies will also be presented at the 2004 BGA symposium, AGenetics of Reading Disabilities.@

Ester M. Derks¹, J.J. Hudziak², C.E.M. Beijsterveldt¹, and D.I. Boomsma, D.I¹. Genetic analyses of Teacher Ratings on Aggression, Attention Problems and Anxiety in 7-, 10-, and 12-year-old children^3.

¹Department of Biological Psychology, Vrije Universiteit, Amsterdam, The Netherlands, ²Department of Psychiatry and Medicine (Division of Human Genetics), Center for Children, Youth and Families, and University of Vermont, College of Medicine, Burlington, USA, ³This work was supported by NWO Grant numbers 575-25-006, 575-25-012, and 904-57-94 (Boomsma, P.I.), and by NIMH Grant number MH58799 (Hudziak, P.I.)

Address: Vrije Universiteit Department of biological psychology van der Boechorststraat 1 1081 BT Amsterdam The Netherlands Telephone: 0031(0)20-4448743 Fax: 0031(0)20-4448832 Email: em.derks@psy.vu.nl

Parental ratings have shown large additive genetic, and moderate non-shared environmental influences on Attention Problems (AP), Aggression (AGG), and Anxiety (ANX). Shared environmental influences were present in AGG and ANX, and non-additive genetic influences were present in AP. (M.J.H. Rietveld, J.J. Hudziak, M. Bartels, C.E.M. Van Beijsterveldt, and D.I. Boomsma, 2004, Journal of Child Psychology and Psychiatry, 45, 577-588), and (J.J. Hudziak, C.E.M. van Beijsterveldt, M. Bartels, M.J.H. Rietveld, D.C. Rettew, E.M. Derks, and D.I. Boomsma, 2003, Behavior Genetics, 33, 575-589). The goal of this study is to determine how these findings might differ when using teacher reports. Objective : Perform univariate genetic analyses on teacher ratings of AP, AGG, and ANX in 7-, 10-, and 12-year-old boys and girls. Methods: Teachers completed the TRF at age 7 (N=1184 pairs), 10 (N=1055 pairs), and 12 (N=807 pairs). The current data do not yet show overlap between ages, therefore cross-sectional analyses were performed. Results: Individual differences in AP, AGG, and ANX in boys and girls were mainly explained by additive genetic factors. Non-shared environment contributed moderately to the variation in these syndromes. Non-additive genetic effects influenced variation in AGG, but only in 10-year-old girls. Influences of shared environment were significant on AGG, and AP in 12-year-old girls. Conclusions: Like parental ratings, teacher ratings of AP, AGG, and ANX in children are mainly explained by large genetic contributions and moderate non-shared environmental influences. Unlike parental ratings, no non-additive genetic influences were found on AP, and no shared environmental influences were found on ANX, although this latter finding may have been due to a lack of power. In addition, non-additive genetic influences were found on AGG in 10-year-old girls. In conclusion, estimates of genetic and environmental influences vary according to who provides the information (mother, father or teacher). However, regardless of informant, genetic influences are the most important contributors to individual differences in expression of psychopathology.

Danielle M. Dick¹, Howard J. Edenberg², Tatiana Foroud², Alison Goate¹, Laura Bierut¹, Bernice Porjesz³, & Henri Begleiter³. Identifying Genes Influencing Alcohol Dependence in the Collaborative Study on the Genetics of Alcoholism (COGA) Sample⁴.

¹ Washington University, St. Louis, MO USA, ² Indiana University School of Medicine, Indianapolis, IN USA, ³State University of New York, Brooklyn, NY USA, ⁴COGA is supported by the NIH Grant U10AA08403 from the National Institute on Alcohol Abuse and Alcoholism (NIAAA).

Address : Washington University School of Medicine, Department of Psychiatry, Box 8134, 660 South Euclid, St. Louis, MO 63110. Telephone: 314-362-3999 Fax: 314-362-4247 E-mail: dickd@psychiatry.wustl.edu

The Collaborative Study on the Genetics of Alcoholism is a multi-center project with the goal of identifying genes influencing alcohol dependence. Families were ascertained through a proband in an inpatient or outpatient treatment center at one of six sites across the United States. Individuals completed a polydiagnostic interview, the Semi-Structured Assessment for the Genetics of Alcoholism (SSAGA), personality questionnaires, and an electrophysiological protocol. Data on 2282 individuals from 262 multiplex alcoholic families are available for genetic analyses. Linkage analyses were conducted on this sample to identify chromosomal regions potentially containing genes influencing alcohol dependence and related phenotypes and endophenotypes. Some of the strongest evidence of linkage (lod=5.0) was obtained for the beta 2 band of EEG on chromosome 4p (Porjesz et al., 2002, PNAS, 99, 3729-3733). There was also evidence of linkage to alcohol dependence diagnoses (lod=2.8) to a region of chromosome 4 just distal to these findings, and a bivariate analysis incorporating electrophysiological data increased the lod to 4.8 (Williams et al., 1999, AJHG, 65, 1148-1160). These linkage peaks corresponded to clusters of GABA-A receptor genes and ADH genes, respectively. These genes were considered strong candidates for potential involvement in alcohol dependence. We followed up these linkage findings by conducting family-based association analyses, testing multiple SNPs in each of the 4 GABA-A receptor genes and 7 ADH genes in the regions. We found significant evidence of association with one of the GABA-A receptor genes, GABRA2, and with one of the ADH genes, ADH4, with alcohol dependence. We have also used this strategy to identify genes that appear to influence the risk for alcohol dependence in several other chromosomal regions. We are hopeful that this new stage of high throughput association testing in the COGA sample will rapidly lead to the identification of additional genes contributing to the risk for alcohol dependence and related phenotypes.

Danielle M. Dick¹, Shaun Purcell², Richard J. Viken³, Jaakko Kaprio⁴, Lea Pulkkinen⁵, & Richard J. Rose³. Identifying Environmental Influences on Adolescent Substance Use in the Finnish Twin Studies⁶

¹ Washington University, St. Louis, MO USA, ² Whitehead Institute, Boston, MA USA, ³ Indiana University, Bloomington, IN USA, ⁴University of Helsinki, Finland, ⁵University of Jyvaskyla, Finland, ⁶The Finnish Twin Studies are supported by AA12502 from the NIAAA and by the Academy of Finland.

Address : Washington University School of Medicine, Department of Psychiatry, Box 8134 660 South Euclid, St. Louis, MO 63110, USA Telephone: 314-362-3999 Fax: 314-362-4247, E-mail: dickd@psychiatry.wustl.edu

Genetically informative study designs provide a powerful method for studying environmental influences and the interplay of genetic and environmental factors. Advances in the analysis of twin data make it possible to test for complex gene-environment interactions in ways not previously possible. With data from two population-based Finnish twin studies, FinnTwin12 and FinnTwin16, we are currently testing the impact of a variety of environmental factors, including familial, peer, and school and community influences, on adolescent substance use and related behaviors. Previously, we have demonstrated that the importance of genetic and environmental factors on adolescent drinking frequency varies in urban and rural settings. These effects were magnified, with nearly five-fold differences in genetic influence across environments, by the incorporation of more detailed, continuous environmental information, such as the percentage of migration in and out of a neighborhood to index neighborhood stability, into the genetically informative design. Recently, we have extended these gene-environment interaction models for the use of ordinal data. We will present information about the power of gene-environment interaction models using ordinal moderators and outcomes. Furthermore, we have begun to apply these models to data from FinnTwin12 to better characterize the effects of family, peer, and school/community influences on adolescent substance use. Our analyses suggest that some environmental influences have main effects on adolescent outcome, while others act by moderating the importance of genetic and environmental effects.

Mara Dierssen, R. Benavides-Piccione, I. Ballesteros-Yáñez, M. Martínez de Lagrán, M.L. Arbonés, V. Fotaki, X. Estivill, J. DeFelipe and G. N. Elston. Brain size and circuit complexity: a study of the cerebral cortex in DYRK1A mouse.

Address: Genes and Disease Program ; Genomic Regulation Center , Passeig Marítim 37-49; 08003 Barcelona E-mail: mara.dierssen@crg.es

The dual-specificity tyrosine-regulated kinase DYRK1A gene maps to the chromosomal segment HSA21q22.2 in the Down syndrome critical region (DSCR), and is believed to be involved in some neurological deficits of DS. It has been demonstrated previously that the lack of one copy of the homologous Dyrk1A in mouse (Dyrk1A) leads to a reduction in brain size and behavioural deficits. However, it remains unclear the structural basis of these behavioural deficits. In the present work we analysed the microstructure of cortical circuitry in the Dyrk1A mouse and control littermates by intracellular injection of Lucifer Yellow in neurones in fixed cortical tissue. We found that pyramidal cells in Dyrk1A mice were considerably smaller, less branched and less spinous than those sampled from control littermates. These results suggest that Dyrk1A is involved in the determination of the size and complexity of pyramidal cells, and thus, in their capability to integrate information.

Thalia C. Eley¹. Phenotypic and genetic/environmental structure of anxiety sensitivity in adolescents².

¹Social, Genetic, Developmental Psychiatry Centre, Institute of Psychiatry, King=s College London, UK, ² The G1219 study was supported by the W T Grant Foundation and by a Medical Research Council training fellowship to the author.

Address : Social, Genetic and Developmental Psychiatry Centre, Box P080, Institute of Psychiatry, De'Crespigny Park, London, UK, SE5 8AF Telephone: +44 (0)20 7848 0063 Fax: +44 (0)20 7848 0866 Email: t.eley@iop.kcl.ac.uk

Anxiety sensitivity is a cognitive risk factor for anxiety in general and panic in particular. There is considerable debate over the structure of anxiety sensitivity and whether it represents one homogenous risk or a group of associated risks. Furthermore, little is know about its genetic and environmental structure, which may shed light on this debate. The current study used confirmatory factor analysis to test the structure of anxiety sensitivity in adolescents using both previously identified models and new models indicated from exploratory factor analysis of our own data. We then tested the genetic and environmental factor structure of these data. Methods. The Children=s Anxiety Sensitivity Index (W. Silverman, 1991, J. Clin. Child Psychol. 20,162-168) was completed by 1,152 twin and sibling pairs aged 12- to 20-years. Exploratory factor analyses and previous studies indicated a total of 6 models. The best-fitting model consisted of four correlated factors reflecting physiological concerns, social concerns, mental concerns, and control. The structure of the genetic and environmental influences on the variation of covariation between these four scales was then examined using 3 multivariate genetic models. The best-fitting model included one set of genetic, shared environment and non-shared environmental factors which influenced each of the four anxiety sensitivity scales, and four further sets which were each specific to one aspects of anxiety sensitivity. The heritability of each of the four scales ranged from 8% (social concerns) to 40% (physical concerns), but was almost all shared across the four scales. There was negligible contribution from the shared environment, with non-shared environment accounting for the remainder of the variance in each of the four factors. These results indicate that although there are four correlated scales within anxiety sensitivity in adolescents, from a genetic perspective, the measure is a unitary construct.

Manuel A.R. Ferreira, David L. Duffy, Nicholas G. Martin. Accounting for strong age-specific sex-limitation in QTL linkage analysis.

Queensland Institute of Medical Research, Brisbane, Australia. Collection of phenotypes and DNA samples was supported by Boehringer Ingelheim (via the former Sequana Therapeutics Inc.). The genome scans were performed by Sequana.

Address : QIMR, Brisbane 4029, Australia, fax 61-7-3362 0101, Telephone 61-7-3356 4581 Email: manuelF@qimr.edu.au

We have measured the levels of total serum immunoglobulin E (IgE) and other asthma related phenotypes in 934 sib-pairs from 803 Australian twin families ascertained via one asthmatic twin proband. A genome scan (~10cM) was performed with subsequent fine mapping (1-5cM) of candidate regions. Overall, the sib correlation for logIgE was 0.25, with no significant differences between same-sex (SS) and opposite-sex (OS) sib pairs, nor between young (both sibs =20) or older sib-pairs. In spite of this apparent covariance homogeneity, a strong age-specific sex-limitation was found: young OS sib-pairs had an IgE correlation (r=0.04, 95% c.i.: -0.12-0.20) significantly lower than young SS pairs (r=0.35, 95% c.i.: 0.21-0.47). This sex-specific effect was no longer observed at a later age, with old OS pairs having a logIgE correlation (r=0.23, 95% c.i.: 0.09-0.36) similar to older SS pairs (r=0.21, 95% c.i.: 0.08-0.33). Thus, young OS sib-pairs displayed a degree of atopic discordance larger than expected from their degree of relatedness. If unaccounted for, this can significantly reduce the power of linkage analysis under the hypothesis of linkage. A similar effect was observed for the skin-prick test data. Since ~25% of our sib-pairs consist of OS pairs with at least one young sib, this sample provides appropriate data to compare the power of linkage analysis under different genetic models, namely sex-specific age-of-onset, classical sex-limitation and parental imprinting. Fine mapping data from chromosome 11 will be used to test these models, namely at the known asthma gene FCER1B and near the candidate genes ELF5 and EHF.

Deborah Finkel¹, Chandra A. Reynolds², and Nancy L. Pedersen³. Surprising lack of sex differences in cognitive decline⁴.

¹Department of Psychology, Indiana University Southeast, New Albany IN USA, ²Department of Psychology, University of California Riverside, Riverside CA USA, ³Karolinska Institutet, Stockholm Sweden and University of Southern California, Los Angeles CA USA, ⁴Supported by NIA AG10175 and a fellowship from the American Philosophical Society.

Address: Department of Psychology, Indiana University Southeast, 4201 Grant Line Rd., New Albany IN 47150 Telephone: 812-941-2668 Fax: 812-941-2591 E-mail:dfinkel@ius.edu

Gender differences in the etiology of normal cognitive functioning in the second half of the lifespan remain largely unexplored. We conducted a behavioral genetic investigation of genetic and environmental contributions to sex differences in level of cognitive performance and rate of decline in the SATSA data set. Behavioral genetic parameterizations of a latent growth curve model were fit to longitudinal data on 11 cognitive measures. 798 non-demented individuals had available cognitive data across 4 waves of measurement covering 13 years. Participants ranged in age from 44 to 88 at the first testing wave, and 60% were female. Results indicate sex differences in mean intercept on 5 cognitive measures. Men had higher mean intercepts on Information and two spatial measures. Women had higher mean intercepts on memory and speed measures. Regardless of differences in intercept, men and women demonstrated the same aging trajectories for most of the cognitive measures. Sex differences in rates of decline were found for only two cognitive tests. Faster quadratic decline in Information was found for women and faster linear decline in Card Rotations for men. Only Synonyms demonstrated sex differences in genetic and environmental contributions to mean intercept. Heritability was higher in men than women, although growth curve parameters indicated heritability estimates begin to converge in late adulthood. In both sexes heritability began to decrease at age 70 and nonshared environment began to increase at the same age. Despite differential longevity and susceptibility to disease, there are no consistent indications that men and women show different patterns of cognitive aging.

Gene S. Fisch¹. Developing a Murine Model for Williams Syndrome.

¹Biostatistics Unit, North Shore LIJ Research Institute, Manhasset, NY USA.

Address: 1129 Northern Blvd, Suite 302, Manhasset, NY 11030 USA, Telephone: 516-240-8300 Fax: 516-240-8344 E-mail: gfisch@nshs.edu

Williams Syndrome (WS) is associated with a ~1.5 Mb microdeletion on 7q11.23. WS is characterized by various clinical, behavioral and cognitive features: Craniofacial dysmorphologies, renal and vascular stenoses, growth deficiencies, dental anomalies, neurobehavioral and musculoskeletal abnormalities. Individuals with WS exhibit mild to severe mental retardation (MR), and an unusual cognitive-behavioral profile: strengths in verbal and selected expressive language skills; weaknesses in abstract/visual reasoning and short-term memory. Individuals with WS are unusually sociable, but may also exhibit anxiety. More than 20 genes have been mapped to this region. Physical mapping has been problematic, due to the high density of repetitive genetic sequences and close blocking of almost identical genetic sequences. Researchers have also sought to identify a murine homolog to the human WS genomic region, with mixed results. Genes associated with chromatin remodeling (WSTF), skeletal muscle specification (FZD9) have been found in humans and mice. Genes associated with brain development, such as PCLO, CYLN2 and CLIP-115, when deficient in mice, exhibit features of WS: growth deficiencies, brain abnormalities, hippocampal dysfunction, and deficits in motor coordination. TFII-I, a transcription factor widely expressed in mouse and human brain, is anatomically altered in humans with WS, and may contribute to MR. However, differences between human and mouse genomic regions have also been observed. The orthologous WS region in mouse is devoid of large, duplicated segments characteristic in humans. WBSCR15 in humans and Wbscr15 in mice show different cDNA, corresponding protein patterns, and tissue expression patterns. Constructing murine orthologues of contiguous gene disorders such as WS may be more problematic than those posed by single gene mutations, e.g., the fragile X mutation (FRAXA). The Dutch-Belgian fragile X consortium created an fmr1 knockout mouse, but most studies found only mild visual-spatial deficits. We found that fmr1 ko mice performed as well or better than controls.

Gene S. Fisch¹, Richard J. Simensen², Roger J. Schroer². Longitudinal Assessment of Cognitive-Behavioral Features of Children and Adolescents with either Autism or the Fragile X Mutation.

¹ Biostatistics Unit, North Shore / LIJ Research Institute, Manhasset, NY, ²Greenwood Genetics Center, Greenwood, SC.

Address : North Shore / LIJ Research Institute, Biostatistics Unit,

1129 Northern Blvd., Suite 302, Manhasset, NY 11030 Telephone: 516-240-8300 Fax: 516-240-8344 E-mail: gfisch@nshs.edu

Although there are differences in the two populations, individuals diagnosed with either autism or the fragile X mutation (FRAXA) exhibit many similar behavioral characteristics and maladaptive behavior patterns. We examined 36 children, ages 4-15 years, diagnosed with either FRAXA (N=18) or autism (N=18), to determine whether cross-sectional profiles or longitudinal patterns in cognitive-behavioral development existed between the two disorders. All children were tested, then retested two years later, using the same two standardized instruments. The Stanford-Binet (4^th Ed.)(SBFE) was used to test cognitive abilities (IQ); the Vineland Adaptive Behavior Scale (VABS) was used to assess adaptive behavior levels (DQ). We found significant decreases in IQ scores in younger children with either autism or FRAXA. Older children with autism exhibited stable test-retest IQ scores, whereas comparably aged children with FRAXA continue to show significant declines in IQ scores. Declines in specific SAS area scores were exhibited by the younger cohort of children diagnosed with either FRAXA or autism. Older children with FRAXA continue to show declines in all SAS area scores, however older children with autism show significant increases in all areas. Declines in adaptive behavior composite (DQ) scores were observed in both groups of children and adolescents, and in both the younger and older cohorts of individuals with FRAXA or autism. Declines in all adaptive behavior domains - Communication, Socialization, Daily Living Skills - were observed in both FRAXA and autism groups, and in both age cohorts.

Tom A. Fowler¹ and Anita Thapar¹. The co-action and interaction of genetic and environmental influences on ADHD.

Department of Psychological Medicine, University of Wales College of Medicine, Cardiff, UK

Address : Department of Psychological Medicine, University of Wales College of Medicine, Cardiff, UK, CF14 4XN, Telephone: (+44) 029 20742201 Fax: (+44) 029 2074**** Email: fowlerta@cardiff.ac.uk

Children=s ADHD symptoms have been consistently shown to be one of the most highly heritable childhood traits (A. Thapar, R. Harrington, K. Ross and P. McGuffin, 2000, J. Am. Acad. Child Adolesc. Psychiatry 39, 1528-36). A variety of factors have been identified as being associated with ADHD, including family conflict and prenatal influences (S.A. Burt, R.F. Krueger, M. McGue and W. Iacono,2003, Arch Gen Psychiatry 60, 505-13; K.M. Linnet, S. Dalsgaard, C. Obel, K. Wisborg, T.B. Henriksen, A. Rodriguez, A. Kotimaa, I. Moilanen, P.H. Thomsen, J. Olsen, and M.R. Jarvelin, 2003, Am J Psychiatry, 160(6), 1028-40). The additive influence of maternal smoking during pregnancy, pregnancy problems and family conflict on children=s symptoms of ADHD was examined in a sample of twins using the MACE model (S. Purcell, 2002, Twin Research 5(6), 554-571). The study also examined whether the contribution of genetic factors to the ADHD symptoms varied as a function of the level of the environmental factors. The sample consisted of the Greater Manchester subset of the population based UK twin register CASTANET (1452 twin pairs, 5-16 years of age). Small but significant direct effects for all the risk factors were. The relative contribution of genetic factors varied as a function of family conflict. These results suggest that even in a highly heritable disorder such as ADHD it is still profitable examine the influence of directly measured environmental variables.

Clyde Francks¹, Silvia Paracchini¹, Shelley D. Smith², Alex J. Richardson³, Tom S. Scerri¹, Lon R. Cardon¹, Angela J. Marlow¹, I. Laurence MacPhie¹, Janet Walter³, Bruce F. Pennington⁴, Simon E. Fisher¹, Richard K. Olson⁵, John C. DeFries⁵, John F. Stein³, and Anthony P. Monaco¹. Association mapping of the 6p23-21.3 QTL for reading disability⁶

¹Wellcome Trust Centre for Human Genetics, University of Oxford, Roosevelt Drive, Oxford, OX3 7BN, UK. ²Department of Pediatrics, University of Nebraska Medical Center, Omaha, Nebraska, 68198-5455, USA. ³Department of Physiology, University of Oxford, Parks Road, Oxford, OX1 3PT, UK. ⁴Department of Psychology, University of Denver, Denver, Colorado 80208, USA. ⁵Institute for Behavioral Genetics, University of Colorado, Boulder, Colorado 80309-0447, USA.

⁶Supported by the Wellcome Trust, NICHD (HD-11681 and HD-27802), the British Council and NSERC (Canada), and the Royal Society (UK).

Address : Wellcome Trust Centre for Human Genetics, University of Oxford, Roosevelt Drive, Oxford, OX3 7BN UK Telephone: +44(0)1865287509 Fax: +44(0)1865287650 Email: clyde.francks@well.ox.ac.uk

Several quantitative trait loci (QTLs) that influence developmental dyslexia (reading disability; RD) have been mapped to chromosome regions by linkage analysis. The most consistently replicated linkage is on chromosome 6p21.3-23. This region showed one of the two strongest linkage signals in our previous genome-wide screen of 195 sibling pairs from the UK, each with at least one RD proband. However, linkage studies have failed so far to refine the candidate interval on 6p to less than roughly 16 Mb. We previously used multivariate linkage analysis to show that the QTL on 6p21.3-23 appeared to influence trait variability that was shared between reading-related cognitive measures, but that was not shared with measures of general intelligence (IQ). We therefore reasoned that shared variance between reading-related measures and IQ could be considered as noise in further linkage and association analysis of this specific QTL. We have refined our linkage mapping of the QTL by removing the variance in reading-related measures that is shared with IQ. We then selected brain-expressed genes within the new candidate interval for association analysis with single nucleotide polymorphsisms, in families from the UK and Colorado. Results from these analyses will be presented.

Qiang Fu¹, Andrew C. Heath², Kathleen K. Bucholz², Michael J. Lyons³, Ming T. Tsuang⁴, William R. True¹, Seth A. Eisen⁵. A twin study of posttraumatic stress disorder in Vietnam era veterans.

¹Department of Community Health, St. Louis University School of Public Health, St. Louis, Missouri, USA, ²Department of Psychiatry, Washington University School of Medicine, St. Louis Missouri, USA, ³Department of Psychology, Boston University, Boston, Massachusetts, USA, ⁴Department of Psychiatry, University of California School of Medicine, San Diego, California, USA, ⁵Department of Internal Medicine, Washington University School of Medicine, St. Louis, Missouri, USA.

Address : Department of Community Health, St. Louis University School of Public Health, 3545 Lafayette Ave., Suite 300, St. Louis MO, 63104, USA

Telephone: (314) 977 8134 Fax: (314) 977 3234 Email: qjfu@slu.edu

Although conduct disorder (CD) and major depression (MD) have found to be associated with posttraumatic stress disorder (PTSD), little is known about how those disorders were correlated. This study was to examine whether PTSD shares common genetic and environmental risk factors for CD and MD. Data were analyzed from 6,735 middle aged male-male monozygotic and dizygotic twins from the Vietnam Era Twin Registry where both members of a pair responded to the diagnostic items in a teleTelephone interview conducted in 1992. Lifetime DSM3R diagnoses of CD, MD, and PTSD were assessed using the DIS3R. Cholesky models were fitted to the polychoric correlation and weight matrices calculated using PRELIS to estimate additive genetic, shared environmental and nonshared environmental effects common and specific to the CD, MD, and PTSD using Mx. A set of competing nested models were assessed by comparing the goodness-of-fit between the full model and the submodel using the chi-square test and AIC. We found that genetic, shared environmental and nonshared environmental risk factors contributed to 21% (95%CI: 11-30%), 9% (95%CI: 2-17%), and 70% (95%CI: 64-77%) of the total variance of PTSD, respectively. All genetic effects on PTSD overlapped with those on MD, while all shared environmental effects on PTSD were the same as those on CD. Approximate 4% and 11% of nonshared environmental risk factors for PTSD were explained by those for CD and MD, respectively. Our data suggest that CD and MD were associated with different liabilities to PTSD.

Jan Fullerton¹, Saffron Brady¹, Matthew Cubin¹, Hemant Tiwari³, Chenxi Wang³, Amarjit Bomhra¹, Stuart Davidson¹, Sue Miller¹, Christopher Fairburn², Guy Goodwin², Michael C Neale⁴, Simon Fiddy¹, Richard Mott¹, David B Allison³, and Jonathan Flint¹. Molecular approaches to identifying candidate genes in depression⁵.

¹The Wellcome Trust Centre for Human Genetics, University of Oxford, UK, ²Department of Psychiatry, University of Oxford, UK, ³Department of Biostatistics, Section of Statistical Genetics, University of Alabama at Birmingham, Birmingham, UK, ⁴Virginia Institute for Psychiatric and Behavioural Genetics, Richmond, USA, ⁵Funded by The Wellcome Trust.

Address: The Wellcome Trust Centre for Human Genetics, University of Oxford, Roosevelt Drive, Oxford OX3 7BN, United Kingdom Telephone: +44 (0)1865 287 520 Fax: +44 (0)1865 287 501 Email: janf@well.ox.ac.uk

Depression is an extremely common and genetically complex trait, whose presentation is not only affected by environmental factors, but also shows gender effects and is genetically correlated with the personality trait neuroticism. We have conducted a genome wide scan for quantitative trait loci (QTL) which influence variation in the personality trait neuroticism, using 87 extreme discordant and 190 extreme concordant sibling pairs, selected from 34,000 sibships in the south west of England who completed a personality questionnaire. We found evidence for linkage to chromosomes 1q (-logP₁₀ = 3.95), 4q (3.84), 7p (3.90), 11q (3.70), 12q (4.74) and 13q (3.81), three of these loci appear to be female specific. One locus, on chromosome 1 is syntenic with a QTL detected for rodent emotionality, an animal model of neuroticism. Fine mapping of the rodent QTL using heterogeneous stock mice (derived from an eight way inbred strain cross) has revealed a series of attractive functional candidate genes, which we have tested for association in our extreme selected human sample. We show that a combined approach using multiple mapping strategies is an effective method for identifying molecular variants contributing to complex genetic traits.

Amber L. Gahagan¹, Irwin D. Waldman¹. Age-of-onset as a moderator of the genetic and environmental influences on Conduct Disorder

¹ Emory University, Atlanta, Georgia USA.

Address : Department of Psychology 532 N. Kilgo Circle Emory University, Atlanta, GA 30322 Telephone: 404 727 2741 Fax: 404 727 1284 Email: agahaga@emory.edu

Behavior genetic studies of antisocial behavior have suggested that both genetic and environmental influences underlie the development of antisocial behavior. Meta-analyses of these twin and adoption studies have also indicated that there is substantial heterogeneity in the estimates of genetic and environmental influences across studies. Recent developmental theories posit that heterogeneity in the etiology of early- as compared with late-onset antisocial behavior may account for these between study differences. In this study, we examine the effects of the age-of-onset of symptoms on the genetic and environmental influences on Conduct Disorder (CD) in a child and adolescent twin sample. Gender differences will also be taken into account, given the well documented gender differences in the prevalence and severity of antisocial behavior/CD. The participants were twin pairs from the Georgia Twin Registry, a registry of all multiple births recorded in the state birth records of Georgia from 1973 to 1991. In the present study, questionnaire data including DSM B IV symptoms of CD as well as age-of-onset of symptoms were available for ~500 twin pairs. We will first report the results of univariate behavior genetic analyses of CD. Second, we will repeat these analyses separately for boys and girls to test whether the genetic and environmental influences underlying CD differ by sex. Next, we will examine genetic and environmental influences on CD as a function of the age-of-onset of symptoms. More specifically, univariate behavior genetic analyses of CD will be conducted separately for children with early and late onset of symptoms in order to test whether the influences on CD differ as a function of the age at which symptoms first appear. We will also conduct analyses using age-of-onset as a continuous moderator of genetic and environmental influences on CD. Finally, we will repeat these analyses separately for boys and girls to see whether any age-of-onset differences vary by gender.

Javier Gayán¹ and The US-Venezuela Collaborative Research Project². Genetic and environmental factors influence Huntington's disease age of onset³.

¹Wellcome Trust Centre for Human Genetics, University of Oxford, U.K, ²Columbia University, 1051 Riverside Dr., New York, NY, U.S.A, ³Supported by NINDS, NIH, WMKF, and HDF.

Address : Wellcome Trust Centre for Human Genetics, University of Oxford, Roosevelt Drive, Oxford, OX3 7BN, U.K. Telephone: +44 1865 287 718 Fax: +44 1865 287 697 Email: gayan@well.ox.ac.uk

Huntington=s disease (HD) is an autosomal dominant neurodegenerative disease caused by a triplet (CAG) expansion mutation. The length of the triplet repeat is the most important factor in determining age of onset of HD, although substantial variability remains after controlling for repeat length. Analysis of the 83 kindreds that comprise the Venezuelan HD kindreds demonstrates that residual variability in age of onset has both genetic and environmental components. We created a residual age of onset phenotype from a regression analysis of the log of age of onset on repeat length. Familial correlations (correlation"standard error) were estimated for sibling (0.40"0.09), parent-offspring (0.10"0.11), avuncular (0.07"0.11) and cousin (0.15"0.10) pairs, suggesting a familial origin for the residual variance in onset. Using a variance components approach with all available familial relationships, we estimated the components of additive genetic (0.37), shared environment (0.22), and non-shared environment (0.41) variances, confirming that about 40% of the variance remaining in onset age is attributable to genes other than the HD gene and 60% is environmental.

Javier Gayán¹, Erik G. Willcutt², Simon E. Fisher¹, Clyde Francks¹, Lon R. Cardon¹, Richard K. Olson², Bruce F. Pennington³, Shelley D. Smith⁴, Anthony P. Monaco¹, and John C. DeFries. Bivariate Linkage Scan for Reading Disability and Attention-Deficit/Hyperactivity Disorder⁵

¹Wellcome Trust Centre for Human Genetics, University of Oxford, UK, ²Institute for Behavioral Genetics, University of Colorado, Boulder, USA, ³Department of Psychology, University of Denver, USA, ⁴Munroe Meyer Institute, University of Nebraska Medical Center, Omaha, USA, ⁵Supported in part by NICHD (HD-11681 and HD-27802) and by the Wellcome Trust.

Address : Wellcome Trust Centre for Human Genetics, University of Oxford, Roosevelt Drive, Oxford, OX3 7BN, U.K. Telephone: +44 1865 287 718 Fax: +44 1865 287 697 Email: gayan@well.ox.ac.uk

There is a growing interest in the study of the genetic origins of comorbidity, a direct consequence of the recent findings of genetic loci that are seemingly linked to more than one disorder. Here, we employ a bivariate linkage test for selected samples that could help identify these pleiotropic loci. This linkage method was employed to carry out the first bivariate genomewide analysis for reading disability (RD) and attention-deficit/hyperactivity disorder (ADHD), in a selected sample of 182 sibling pairs. Several potential pleiotropic loci were localized, such as a novel locus at chromosome 14q32 (multipoint LOD=2.50; singlepoint LOD=3.85). Other less significant loci are also suggested as possibly being pleiotropic for RD and ADHD, as well as other loci possibly having an unique effect on either phenotype.

Heather L. Gelhorn, Michael C. Stallings, Susan E. Young, Robin P. Corley, Soo Hyun Rhee, John K. Hewitt. Genetic and Environmental Architecture of Aggressive and Non-Aggressive Conduct Disorder.

Institute for Behavioral Genetics, University of Colorado, Boulder, CO USA. Supported by NIH grants DA11015, DA12845, MH01865, MH43899.

Address : Institute for Behavioral Genetics Campus Box 447 University of Colorado, Boulder, CO 80309-0447 Telephone: (303) 735-2428 Fax: (303) 492-8063 Email: gelhorn@colorado.edu

Studies of adolescents from twin and adoption studies have consistently implicated genetic factors in the etiology of antisocial behavior (ASB). Further, developmental theories of antisocial behavior posit that aggressive antisocial behavior is more heavily influenced by genetic factors and more stable over time, while non-aggressive behavior is transient and is more likely influenced by the environment. Previously published studies on aggressive versus non-aggressive antisocial behavior have reported substantial heritability of both domains with higher heritability estimates for the aggressive domain. Behavioral genetic studies comparing aggressive and non-aggressive behavior problems have commonly used the Childhood Behavior Checklist (CBCL), which includes items that are relatively normative and frequently endorsed in community samples. It is unknown whether or not the results of these studies will generalize to clinically significant forms of ASB such as conduct disorder (CD). The current study uses 1100 community-based adolescent twin pairs (531 MZ and 569 DZ pairs) participating in the Center for Antisocial Drug Dependence at the University of Colorado. The antisocial behavior of the twins was assessed using the Diagnostic Interview Schedule for Children; a diagnostic tool based on the clinical DSM-IV CD criteria. Analyses of our sample resulted in moderate heritability estimates for both the aggressive and non-aggressive domains. The aim of the current investigation is to determine the extent to which the genetic and environmental influences on aggressive and non-aggressive domains are shared. Identifying the nature and extent of covariation between the domains may aid in validating or refuting developmental theories, and be useful for developing phenotypes for future linkage and association studies.

Aude Gérard¹, Michèle Carlier¹, Nicole Philip², Laurent Villard³ and Brigitte Chabrol⁴. Do patients bearing oligophrenin 1 gene mutations present relative strength in visuo-spatial cognition? Preliminary results.

¹Laboratory PsyCLÉ (EA327, University of Provence, France, ²Department of Medical génetics, Timone Children=s Hospital, Marseille, France, ³INSERM U491-IFR 125 Faculty of Medecine La Timone, Marseille, France, ⁴Department of Paediatric Neurology, Timone Children=s Hospital, Marseille, France.

Address : PsyCLÉ, UFR de Psychologie Sciences de l=Education, 29 avenue Robert Schuman, 13624 Aix en Provence, France Telephone : (33) (0) 442933999 Fax : (33) (0) 442389170 E- mail : aude.gerard2@wanadoo.fr

Oligophrenin 1 gene mutations, located on the X chromosome (on Xq12), causes cerebellar hypoplasia, epilepsy and mental retardation. This gene was recently cloned (P. Billuart et al., 1998, Nature, 392, 923-926). Five families bearing such mutations have been described (N. Philip et al., 2003, J Med Genet, 40, 441-446; C. Bergmann et al., 2003, Brain, 126, 1537-1544; V. des Portes et al., Amer J Med Genet, 2004, 124A, 364-371). The cerebellar hypoplasia observed in these patients may be one of the cause of the mental retardation described in the publications since this cerebellum is known to be involved in higher cognitive processes. Psychological assessment of the members of two families followed in consultation at the Timone Children >s Hospital, were undertaken (the battery includes Wechsler intelligence scale, Kaufman ABC, part of the NEPSY and laterality tasks). The first family included four patients : three brothers and their heterozygous mother. In the second family, there is only one boy carrying a de novo mutation. Large variabilities in the intellectual levels of the boys carrying the mutation were observed. However, in each case, the Performance IQ was better than the Verbal IQ, with large inter subtest variations in the patient of the second family. A relative advantage in the visuo-spatial tests seems emerge. As expected, the heterozygous mother of the first family had an intellectual level higher than that of her children. A third family recently diagnosed will be assessed to confirm or not the relative advantage in visuo-spatial cognition.

Nathan A. Gillespie¹, David E. Evans¹, G. Zhu¹, Margie M. Wright¹, and Nicholas G. Martin¹. A genome wide scan for adolescent personality: Psychoticism, Extraversion, Neuroticism and Lie.

¹Genetic Epidemiology Laboratory, Queensland Institute of Medical Research, Brisbane, Queensland, Australia.

Address : Genetic Epidemiology Laboratory, Queensland Institute of Medical Research, Post Office, Royal Brisbane Hospital, Brisbane QLD 4029, Australia Telephone: +61 7 3362 0226 Fax: +61 7 3362 0101 Email:nathanG@qimr.edu.au

This paper will present the results from a genome-wide scan of adolescent personality as measured by the Junior Eysenck Personality Questionnaire. Genotyping data were available from 439 families comprising adolescent twins, their siblings and parents. Twins and siblings completed the JEPQ as close as possible to the twins= 12^th, 14^th and 16^th birthdays. Variance components linkage analyses were run in Merlin. These were followed by multivariate QTL analyses in Mx using the IBD probability pi-hat method to take advantage of the time series data. Based on the combined male and female sample, suggestive linkage peaks for Psychoticism were observed at 1q42, for Extraversion at 7q33, and for Lie at 4q34. The largest peaks for Neuroticism were linked to 4q21. When data were analyzed separately by sex, areas of suggestive linkage were found between 4q13 and 4q21 for female Neuroticism only. We were unable to replicate the results on the second and third occasions, which was probably caused by the smaller number of subjects with complete personality and genotyping data. In order to improve power and to locate more precisely QTLs for personality, we are increasing the current sample by genotyping an additional 200 families with markers at 5cM intervals.

Ian R. Gizer¹, Irwin D. Waldman¹, Ann Abramowitz¹, Craig Stever², and David C. Rowe². The Utility of Mother, Father, and Teacher Ratings in Association Analyses of Candidate Genes for Childhood ADHD.

¹Department of Psychology, Emory University, Atlanta GA USA, ²Department of Family and Consumer Resources and Genetics, University of Arizona, Tucson AZ USA.

Address : Emory University, Dept. of Psychology, 532 North Kilgo Circle, Atlanta GA 30322 Telephone: 404-727-7434 Email: igizer@emory.edu

Twin studies suggest that genetic influences contribute to the development of the inattentive and hyperactive-impulsive symptoms of DSM-IV attention deficit/hyperactivity disorder (ADHD). Thus, researchers have tested for association and linkage between ADHD and candidate genes hypothesized to be etiologically relevant to the disorder based on their function. Importantly, such tests require careful assessment of the phenotype if they are to represent valid tests of association and linkage. In molecular genetic studies of childhood disorders, such as ADHD, phenotypes are typically operationalized using symptom ratings from multiple informants - e.g., mothers, fathers, and teachers - and combined to produce a composite symptom score. Nonetheless, the methods by which these symptom ratings are typically combined have never been empirically evaluated within the context of molecular genetic studies. In the present study, we evaluated the evidence for association between ADHD and the dopamine transporter (DAT1) and dopamine D4 receptor (DRD4) genes for mother, father, and teacher ADHD symptom ratings uniquely and in combination. A clinic-referred sample of 101 children aged 3-17 was genotyped for the 40-bp repeat at the VNTR in the 3= UTR of DAT1 and the 48-bp repeat in exon 3 of DRD4. We used ordinal regression analyses to test for association between each gene and each informant=s symptom ratings uniquely and in combination. Our results suggested that among single informants, teacher ratings were the most informative for establishing association between DAT1 and DRD4 and ADHD. Importantly, regression analyses yielded stronger evidence for association when mother and teacher ratings were evaluated in combination relative to teacher ratings alone. In addition, father ratings failed to demonstrate an incremental contribution over and above mother or teacher ratings. Thus, our findings suggest using a regression-based approach that incorporates symptom ratings from mothers and teachers to test for association between candidate genes and ADHD symptoms.

Detre A. Godinez¹, Michael C. Stallings¹, Jeff Lessem¹, John K. Hewitt^1. Investigating Age at Onset Twin Analyses with Simulated Data².

¹Institute for Behavioral Genetics, University of Colorado, Boulder, CO USA, ²Supported by NIDA grants DA-05131, DA-11015 and DA-12845

Address : Institute for Behavioral Genetics, Campus Box 447 University of Colorado, Boulder, CO 80309-0447 USA Telephone: (303) 735-3076 Fax: (303) 492-8063 Email:godinez@Colorado.edu

Due to the analytical complications encountered with censored data, most twin studies have not examined etiological influences on age-at-onset phenotypes (e.g., age at initiation of alcohol use). Censored data occurs when some of the participants have not passed through the age of risk; therefore, their age-at-onset is unknown or censored. Survival analyses are typically used with age-at-onset data, but until recently the survival method was not adapted for the twin design. The current study uses simulated twin data to investigate the accuracy of genetic and environmental estimates on age-at-onset data. The first set of analyses excluded censored data, utilizing only twin pairs in which both twins had reported an age-at-onset. The second set of analyses utilized all the data including censored data. To account for censored data, a raw data maximum likelihood analysis of survival data proposed by Pickles et al. (1994, Behav. Genet. 24, 457-468) was used. In a third set of analyses systematic telescoping was simulated. Telescoping is a common phenomenon where, as the distance between current age (i.e., age at assessment) and onset age increases, a subject is more likely to report a biased (i.e., telescoped, or more recent) onset age. Results of analysis-1 showed that excluding censored data resulted in underestimates of both heritability and shared environment, while the results of analysis-2 provided estimates very close to those simulated. Preliminary analyses showed that varying the extent of censored data affects parameter estimates, and the particular effects may be model dependent (i.e., depend on the extent of genetic and environmental influences). The third set of analyses indicated that ignoring telescoping leads to overestimates of shared environment.

Robin P. Goino and Judy L. Silberg5

Genetic influences and developmental trends in comorbid asthma, anxiety, and depression.

oDepartment of Psychiatry, Virginia Institute for Psychiatric & Behavioral Genetics, Virginia Commonwealth University, Richmond, VA USA, 5Department of Human Genetics, Virginia Institute for Psychiatric & Behavioral Genetics, Virginia Commonwealth University, Richmond, VA USA.

Address : Virginia Institute for Psychiatric & Behavioral Genetics, P.O. Box 980126, Richmond, VA 23298-0126 Telephone: 1 804 828 8591 Fax: 1 804 828 1471 Email: rgoin@vcu.edu

Asthma has been described as the most widespread chronic illness of childhood (D.A. Mrazek, 2003, Child Adolesc Psychiatric Clin N Amer 12, 459-471), and several studies document links between asthma and both anxiety and depression (G. Vila, C. Nollet-Clemencon, J. de Blic, M.C. Mouren-Simeoni, and P. Scheinmann, 2000, J Affect Disord 58, 223-231). We sought to examine the rates of asthma, depression, and anxiety among a population-based sample of twins from the Virginia Twin Study of Adolescent and Behavioral Development (n = 2560; mean age = 11.96, SD = 2.58, range = 8 to 18). We were further interested in (a) assessing the relationships between asthma, depression, and anxiety; (b) the proportions of genetic and environmental contributions to these phenotypes; and (c) the developmental sequence of comorbidity. Mothers of 264 twins (10.4%) indicated that their children had asthma or attacks of wheezing occasionally (not as much as once per week) or at least once a week; most reported asthmatic problems only occasionally (90%). MZ twin correlations were .51 and .43 for boys and girls, respectively. DZ twin correlations were .17 for boys, .03 for girls, and .04 for opposite-sex pairs. Children=s scores on a self-report measure of mood and feelings (depression) ranged from 33 to 99 (indicating greater pathology), with a mean of 44.11 (SD = 9.57). The correlation between asthma and depression was modest, r = .04, p = .07. Child anxiety per maternal ratings on the Rutter=s scale ranged from 5 to 15 (indicating greater anxiety), with a mean of 6.76 (SD = 1.77). There was no significant relationship between asthma and anxiety, r = .01, p = .49. Additional analyses on the (a) proportions of variance attributed to genes and environment for asthma, depression, and anxiety and (b) comorbid emergence of these conditions are underway.

Julia D. Grant¹, Andrew C. Heath¹, and Kathleen K. Bucholz¹. Do genetic and/or environmental influences on abuse and dependence overlap? Explorations using cannabis and alcohol².

¹Department of Psychiatry, Washington University School of Medicine, St. Louis, MO USA, ²Supported by NIH Grants DA14632, DA14363, AA07728, AA11998.

Address: Dept. of Psychiatry, Washington Univ. School of Medicine, 40 N. Kingshighway, Ste. 3, St. Louis, MO 63108 Telephone: (314) 286-2255 FAX: (314) 286-2243 Email: grantj@msnotes.wustl.edu

Although DSM abuse and dependence are assessed using different criteria, it is possible that common genetic (and/or environmental) factors underlie both diagnoses. Data from 8,169 men in the Vietnam Era Twin Registry (VETS), a national registry of male-male twin pairs in which both individuals were in the military during the Vietnam-era, were used to explore the overlap between abuse and dependence (MZ pairs=1923, DZ pairs=1529, pairs with unknown zygosity=85, and 1,095 singletons; mean age = 42.0 years at the time of Telephone diagnostic interview in 1992). Two outcomes were assessed independently: cannabis and alcohol. A four-level variable was used to indicate lifetime number of DSM-III-R dependence symptoms (0, 1, 2, and 3+). Abuse was set to missing for all individuals with 3+ dependence symptoms; individuals with 0-2 dependence symptoms were classified according to whether or not they met DSM-III-R abuse criteria. Bivariate analyses indicated that alcohol was best described by a model with additive genetic and nonshared environmental influences (h²=.48 for dependence symptoms; h²=.47 for abuse diagnosis). Both the genetic correlation (0.96, 95% CI=0.94-0.99) and the nonshared environmental correlation (0.93, 95% CI 0.91-0.95) were significant and substantial. Cannabis abuse/dependence was best described by a model containing additive genetic, shared environmental, and nonshared environmental influences (h²=.31 and c²=.24 for dependence symptoms; h²=.34 and c²=.13 for abuse diagnosis). Both the genetic correlation (1.00, 95% CI = 0.93-1.00) and the nonshared environmental correlation (0.93, 95% CI = 0.90-0.95) were significant and substantial. The best fitting cannabis model was one with no shared environmental influences specific to abuse, suggesting that the shared environmental influences on cannabis dependence symptoms and cannabis abuse diagnosis are completely overlapping. These results suggest that, for both alcohol and cannabis, although the assessments for abuse and dependence are distinct, common underlying factors influence the two outcomes.

Alice M. Gregory¹ and Thalia C. Eley¹. Examining the origins of the association between childhood anxiety and sleep problems using a cognitive-genetic-environmental approach².

¹Social, Genetic, Developmental Psychiatry Centre, Institute of Psychiatry, King=s College London, UK, ²Supported by a Career Development Award from the UK Medical Research Council to the second author. We thank Jennifer Lau, Maria Napolitano and the families participating in the TEDS-ECHO study.

Address : Social, Genetic and Developmental Psychiatry Centre, Box P080, Institute of Psychiatry, De'Crespigny Park, London, UK, SE5 8AF Telephone: +44 (0)20 7848 0038 Fax: +44 (0)20 7848 0866 Email: a.gregory@iop.kcl.ac.uk

Twin studies suggest that the association between anxiety and sleep problems in the pre-school years is largely due to shared environmental influences. The etiology of this association has not been examined in school-aged children. Cognitive psychology suggests that similar processes are involved in both anxiety and sleep problems, however the etiologies of these cognitive processes are unknown. This study combines genetic and cognitive research to examine the association between sleep problems and anxiety. Parents of 232 pairs of 8-year-old twins completed the Child Sleep Habits Questionnaire. Children completed the Sleep Self Report and Self-report Childhood Anxiety Related Emotional Disorders questionnaires. A range of information processing factors were examined including attributional style, anxiety sensitivity and interpretation of ambiguous stimuli. Expectations of others were examined using the Children=s Expectations of Social Behaviours Questionnaire. There were moderate correlations between anxiety and both self and parent reported sleep problems (r(461) = .34, p <.01; r(439) = .14, p<.01 respectively). Model fitting analyses using the statistical package Mx showed that the associations between anxiety and sleep were mainly mediated by genetic influences (e.g. genetic influence explained 89% and 90% of the associations between child-report anxiety and parent- and child-reported sleep problems respectively). Correlations between cognitive processes and anxiety and sleep problems were examined. Only expectations of peers correlated significantly with anxiety and both self-, and parent reported sleep problems (r(462) = .19, p<.002, r(460) = .24, p<.001, r(438) = .10, p<.05 respectively). AExpectations of peers@ were largely mediated by environmental influences. This is the first study to combine behavioural genetic and cognitive research in examining the origin of the association between childhood sleep problems and anxiety. These results suggest that genetic influences may increase with age and environmental factors are particularly important in the etiology of certain cognitive factors associated with sleep problems and anxiety.

Elena L. Grigorenko¹, Damaris Ngorosho², Christina Romano¹, Adam Naples¹, Luke Turechek¹, Anna Kochetkova¹, & Joseph Chang¹. Molecular-genetic studies of reading disability in candidate genes³

¹Yale University, New Haven, Connecticut, USA, ²MAKWAMI, Bagamoyo, Tanzania, ³Supported by NIH (Fogarty Foundation) and James S. MacDonnell Foundation.

Address : Elena L. Grigorenko, Child Study Center &, Center for the Psychology of Abilities, Competencies, and Expertise (PACE), Yale University

340 Edwards Street, P.O. Box 208358, New Haven, CT 06520-8358 USA Telephone: 203-432-4660 Fax: 203-432-8317 E-mail: elena.grigorenko@yale.edu http://www.yale.edu/pace

Reading disability has been linked to a number of regions in the genome, including 15q. Recently, a gene, EKN1, with unknown function in the linked region (15q21.1), was identified via a translocation breakpoint in Finnish individuals with reading disability. Subsequently, the gene has been investigated as a susceptibility locus for reading disability by association studies in Finnish- and English-speaking individuals with reading disabilities. Although the gene has been supported as a risk locus for reading disability, it appears that the magnitude of the gene=s effect is rather small. It is possible that other genes in the region contribute to the signals reported on 15q. We have refined out previously reported linkage to 15q in a sample of Kiswahili-speaking sibpairs using multivariate linkage analyses. We then choose a set of candidate genes for association analysis. These analyses were conducted in samples of sibpairs from Tanzania and Russia. The presentation will summarize relevant findings.

Guang Guo and Jianmin Wang. The Mixed or Multilevel Models For Longitudinal Sibling Data.

¹ Carolina Population Center, University of North Carolina, Chapel Hill, ²Gynecologic Oncology Group, Statistical and Data Center, Roswell Park Cancer Institute, Buffalo, New York 14263.

Address :

CB#? 3210, University of North Carolina, Chapel Hill, NC 27599

Most estimates of heritability are cross-sectional, failing to use the stronger within-person longitudinal designs to examine age-dependent patterns. Age-specific estimates are often necessary because results from one age do not carry over to other ages if there are age-specific influences on the trait or behavior under consideration. The estimated genetic influence at one age does not necessarily imply the stability of gene expression over a life time. Nor does it imply the stability of environmental influences over a life time either. Only longitudinal data allow us to determine whether genetic and/or environmental contributions remain constant over a life span. Guo and Wang (2002) demonstrated how the mixed or multi-level model can be adapted for the analysis of genetically related individuals. In the paper, we show how the mixed model can be used to cope with longitudinal sibling data. Given sibling data such as MZ and DZ twins, that is, repeated measures of the phenotype at different ages, the mixed model would estimate heritability as a function of not only race/ethnicity, gender, Y, but also age and cohort. Or the mixed model would yield trajectories of heritability estimates given longitudinal data. We illustrate the statistical model using data from the Add Health. We evaluate the changing level of genetic influence on overweight and obesity during adolescence and early youth and the possible changing level of genetic influences on obesity over the time period of 1994 to 2001. The Add Health respondents were aged 12 to 18 at Wave I in 1994 and they were 19 to 25 at Wave III in 2001. The data allow us to estimate age-specific heritability from 12 through 25. Since the respondents were followed only from 1994 to 2001 and the incidence of obesity was on the rise during the period, we must separate age from calendar period effect. In addition, both age and period effects may vary by demographic, social, and other environmental factors. References Guo, Guang and Jianmin Wang. 2002. The Mixed or Multilevel Model for Behavior Genetic Analysis. Behavior Genetics 32: 37-49.

Brett C. Haberstick.¹, David S. Timberlake ¹, Jeff M. Lessem ¹, Christian Hopfer ², Andrew Smolen ¹, Marissa Ehringer. ¹, and John K. Hewitt. ¹. Genetic and Candidate Gene Analysis of Nicotine Dependence in the Add Health Sample³.

¹Institute for Behavioral Genetics, University of Colorado at BoulderUSA, ²Department of Psychiatry, University of Colorado Health Sciences Center, Denver, ³This research uses data from Add Health, a program project designed by J. Richard Udry, Peter S. Bearman, and Kathleen Mullan Harris, and funded by a grant P01-HD31921 from the National Institute of Child Health and Human Development, with cooperative funding from 17 other agencies. B.C. Haberstick was supported by grant HD07289.

Address : Institute for Behavioral Genetics, University of Colorado at Boulder, Campus Box 447, Boulder, CO, Tel: 303-735-5388, Fax: 303-492-8063, Email: Brett.Haberstick@Colorado.Edu

About 1 million middle and high school students in the USA begin smoking before the age of 18. Of these, it has been estimated that greater than 75% will still be smoking in adulthood (Colby, M.S., Tiffany, S.T., Shiffman, S., and Niaura, R.S. (2000), Drug and Alcohol Dependence, 59: S83-S95). As such, the years of adolescence and young adulthood can be considered a risk period for the onset of daily smoking and the transition into nicotine dependence. While amongst adults, nicotine dependence affects between 20-40% of regular smokers and roughly 50% of daily smokers, similar but slightly smaller estimates have been determined for adolescents (Kandel, D.B. and Chen, K. (2000) Nicotine and Tobacco Research, 2: 263-274). In the current study, we used the Fagerstrom Test for Nicotine Dependence (FTND) and its Heavy Smoking Index (HSI) to explore the genetic and environmental contributions to nicotine dependence. Our study population comprised current smokers assessed during the third wave of data collection in the Add Health sample. For the HSI, monozygotic twins were correlated more strongly than dizygotic twins whose correlation was roughly equal to that of full-siblings, both of which were greater than half-siblings. For the overall FTND the correlations less clearly related to the degree of genetic relationship. Univariate modeling of these correlations confirmed the strong heritable influence (51%) and no shared environmental effects on the HSI, but could not differentiate between an AE and CE model for the FTND. Using the scores on the HSI, we further examined the genetic influences on nicotine dependence by testing for positive associations with a variety of nicotine dependence, indexed by the HSI, related polymorphisms. Preliminary evidence suggests a positive association between nicotine dependence and the promoter polymorphism in the MAOA gene.

Narelle K. Hansell^1,2, M.J. Wright¹, G.M. Geffen², L.B. Geffen², N.G. Martin¹. Linkage Analyses of ERP Slow Wave Measures of Working Memory³.

¹Genetic Epidemiology Laboratory, Queensland Institute of Medical Research, Brisbane, Queensland, Australia, ²Cognitive Psychophysiology Laboratory, Departments of Psychology & Psychiatry, University of Queensland, Brisbane, Queensland, Australia, ³Phenotype collection was supported by Australian Research Council Grants (A79600334, A79906588, A79801419, DP0212016) and the Human Frontier Science Program (RG0154/1998-B).

Address : Genetic Epidemiology Laboratory, Queensland Institute of Medical Research, Post Office, Royal Brisbane Hospital, 4029 Australia. Telephone: 61-7-3362 0299 Fax: 61-7-3362 0101 Email: narelleH@qimr.edu.au

Electrophysiological brain activity was recorded from 549 twin pairs aged 16 years (M = 16.2, SD = 0.3, range = 15.4-18.1) as they participated in a delayed-response working-memory task. An event-related potential (ERP) slow wave (SW) was elicited during the task delay period, in which participants were required to remember the location of a target stimulus. This component is heritable (h² ranges 0.37 to 0.51 at prefrontal and parietal sites B brain regions associated with increased activity during working memory), but has not previously been examined in linkage analyses. Preliminary univariate analyses show some inconsistencies. For example, significant linkage was found on Chromosome 10 for SW recorded at the midline parietal site (LOD = 4.6), suggestive linkage (LOD = 2.8) was found at the left-hemisphere parietal site, but at the right-hemisphere parietal site, although these phenotypes are highly correlated (0.83-0.92), the LOD score at the same marker was a low 1.2. Examples of low but consistent LOD scores were also found at other chromosomal regions (e.g. on Chromosome 1, LOD scores ranging 0.97 to 1.26 peaked at the same marker for SW recorded at midline and left- and right-hemisphere parietal sites). We now plan to examine these phenotypes using multivariate linkage analyses, as these may be able to resolve some of the above issues. This appears to be a promising approach, as the first multivariate genome-wide scan for QTLs linked to a complex trait (developmental dyslexia B A.J. Marlow et al., 2003, Am. J. Hum Genet. 72, 561-570) showed advantages in power and in clarification of the pattern of QTL influence compared to analysing correlated measures separately. The results from our multivariate linkage analyses of SW phenotypes will be reported.

Nicole Harlaar¹ and Robert Plomin¹. Reading abilities and disabilities in a UK population sample of 5500 7-year-old twin pairs²

¹Social, Genetic and Developmental Psychiatry (SGDP) Centre, Institute of Psychiatry, King=s College London, UK, ²Supported by the UK Medical Research Council (Grant G9424799).

Address : SGDP Centre, PO 83, Institute of Psychiatry, De Crespigny Park, London, SE5 8AF, UK Telephone: +44 (0)207-848-0600 Fax: +44 (0)207-848-0895 Email: n.harlaar@iop.kcl.ac.uk

Although the persistence and deleterious consequences of early reading difficulties are well documented, there has been surprisingly little systematic research into the development of individual differences among young readers. In this study we examine the etiology of early reading difficulties in the context of normal reading development. A representative population sample of 5500 same- and opposite-sex twin pairs were assessed at age 7 by UK National Curriculum teacher ratings of reading achievement. These teacher assessments correlate substantially both phenotypically (r = .68) and genetically (r_g = .74) with word and non-word recognition, two key indicators of reading skill. Quantitative genetic analyses show that reading difficulties at age 7, as defined by scores below the 10th percentile on a composite index of reading ability, are substantially heritable (h²_g = .69) and that the genetic risk for reading difficulties is continuous, rather than discrete. Furthermore, genetic influences on reading correlate substantially with genetic influences on more general cognitive abilities (r_g = .59). These findings suggest that susceptibility genes contributing to reading disabilities will operate throughout the distribution of reading ability and that many of these genes will have non-specific effects, influencing other cognitive abilities as well as reading. These hypotheses are being tested in a second seam of our ongoing research, consisting of a case-control association study using pooled DNA from large case (N = 435) and control (N = 1000) samples that seeks to identify non-synonymous single-nucleotide polymorphisms (nsSNPs) associated with early reading and cognitive abilities.

Christie Hartman¹, Erik Willcutt¹, and Bruce Pennington². The relation between ADHD and sluggish cognitive tempo: genetic and environmental contributions³.

¹Institute for Behavioral Genetics, University of Colorado at Boulder, USA, ²Department of Psychology, University of Denver, CO, USA, ³Supported by grants HD-27802, MH-16880, MH-62120, MH-63941, MH-38820, MH-04024.

Address : Institute for Behavioral Genetics Campus Box 447 University of Colorado, Boulder CO 80309-0447 USA Telephone: 303-492-2817 Fax: 303-492-8063 Email: christie.hartman@colorado.edu

Sluggish cognitive tempo (SCT) is an internally consistent construct that is highly correlated with the inattention dimension of DSM-IV ADHD and only weakly related to the hyperactivity/impulsivity dimension. The present study examined whether inattention, hyperactivity/impulsivity, and SCT share a common genetic influence or whether genetic influences specific to one or more of these 3 variables make a significant contribution. The sample included 8-18 year-old twins recruited for participation in the Colorado Learning Disabilities Research Center twin study, an ongoing study of learning disabilities and ADHD. Preliminary univariate analyses of total ADHD symptoms showed no evidence for shared environmental influences and significant evidence for non-additive genetic effects, as seen in prior ADHD studies. Further preliminary multivariate analyses suggested that a common pathways model provided the best fit to the data to explain the covariation between inattention, hyperactivity/impulsivity, and SCT. These results provide evidence that inattention, hyperactivity/impulsivity, and SCT share similar genetic etiology.

Marianna E. Hayiou-Thomas¹, Yulia Kovas¹, Nicole Harlaar¹, Dorothy V. M. Bishop², Philip Dale³, Robert Plomin¹. Common aetiology for diverse linguistic skills in 4 1/2 year old twins⁴.

¹Social, Genetic, and Developmental Psychiatry Center, Institute of Psychiatry, Kings College London, UK.

Address: SGDP Centre, Institute of Psychiatry, Box P083, De Crespigny Park, London SE5 8AF, UK Telephone: +44-20-7848 0074 Fax: +44-20-7848 0092 Email: m.hayiouthomas@iop.kcl.ac.uk

A major issue in child language development concerns whether and to what degree different linguistic skills are independent of each other. The current study used data from more than 1000 4 1/2-year-old UK twins who were tested at home on a battery of diverse language measures, which tapped lexical, grammatical and phonological ability. Multivariate genetic analysis was used to examine the aetiology B in terms of the relative contributions of genetic and environmental factors B of the inter-relationships of these linguistic skills. Phenotypic analysis suggested two latent factors: articulation (2 measures) and general language (the remaining 7), and a genetic model incorporating these factors provided a good fit to the data. Almost all genetic and shared environmental influences on the 7 individual measures acted through the latent factors; with two exceptions, there were no genetic or shared environment effects that were specific to the individual measures. For the articulation latent factor, almost all the genetic and shared environmental influence of the 2 individual measures acted through the latent factor. Even for the two latent factors, there was also evidence of substantial aetiological overlap, with a genetic correlation of .60 and shared environment correlation of 1.00. We conclude that to a large extent, the same genetic and environmental factors underlie the development of different linguistic skills, with some genetic differentiation remaining, especially for articulation.

Norman D. Henderson. Some Biased Reflections on Genetic Interactions

Department of Psychology, Oberlin College, Oberlin, OH 44074

Adddress : Department of Psychology, Oberlin College, Oberlin, OH 44074 USA E-mail : nhenders@oberlin.edu Telephone: 1-440-775-7695

The issue of non-additive genetic influences, arising from Gene H Environment interactions, Gene H Age interactions and Gene H Gene interactions, tends to provoke a wide range of views from behavioral genetic researchers and their critics. Some proclaim that genetic interactions seriously weaken or nullify most BG generalizations, whereas others argue that the interactions are usually too small to worry about. While some researchers look for genetic interactions for the clues they can provide about evolutionary or developmental processes, others regard many of these non-additive effects as measurement artifacts. In this address, I present some of my own biased opinions, both about genetic interactions and about the biases of others.

Susanne Henningsson¹, Anna Håkansson¹, Lars Westberg¹, Michael Landén², Fariba Baghaei³, Roland Rosmond³, Göran Holm³, Elias Eriksson^1,2. Association between interleukin-6 gene polymorphism and personality traits in women.

¹Department of Pharmacology, Göteborg University, Göteborg, Sweden, ²Institute of Clinical Neuroscience, Section of Psychiatry, Göteborg University, Göteborg, Sweden, ³Institute of Heart and Lung Disease, Göteborg University, Göteborg, Sweden. Supported by grants from the Swedish Medical Research Council, Lundberg=s Foundation, Wallenberg=s Foundation, Thuring=s Foundation and Lundbeck=s Foundation

Address:

Dept. of Pharmacology, Box 431, 405 30 Göteborg, Sweden

Telephone: +46 31 773 31 70 Fax: +46 31 82 10 85 Email: susanne.henningsson@pharm.gu.se

Interleukin-6 (IL-6) has been implicated in the regulation of mood and behaviour; hence we hypothesized that the IL-6 gene might be associated with interindividual variations in personality traits. 41-year-old-women (n=197) were assessed using Karolinska Scales of Personality (KSP) and genotyped for the C to G transition at nucleotide -174 of the IL-6 gene promoter (-174 C/G). The possible associations between the polymorphism and four different factors (Aextraversion@, Anon-conformity@, Aneuroticism@ and Apsychoticism@) were analysed. We found significant association between the KSP factor non-conformity and the IL-6 polymorphism (P=0.004). In addition, two of its subscales, indirect aggression (P=0.003) and verbal aggression (P=0.04), were found to be significantly associated with the gene. The results suggest that the studied IL-6 polymorphism may contribute to specific components of personality.

Christian Hopfer¹, David Timberlake² , Brett Haberstick², Jeff Lessem², Marissa Ehringer², Andrew Smolen², John Hewitt². Genetic Influences on Quantity of Alcohol Consumed by Adolescents and Young Adults.

¹Department of Psychiatry, University of Colorado Health Sciences Center, Denver, Colorado, ²Institute for Behavioral Genetics, Boulder, CO USA. Supported by: P01-HD31921, DA000357, HD07289, AA07464, DA11015, EY 012562, and DA015522.

Address : Dr. Christian J. Hopfer, 4200 E. Ninth Ave, Box C-268-35 University of Colorado Health Sciences Center, Denver, CO, 80262; Email: Christian.Hopfer@uchsc.edu

Objective: To examine the relative influence of genetic and environmental factors on drinking patterns in a nationally representative, genetically informative, dataset of adolescents and young adults and to test four candidate genes for association with drinking behaviors. Method: The average quantity of alcohol used per drinking episode during the past year was analyzed in a genetically informative subsample of the Add Health longitudinal sample of adolescents and young adults. Total Sample Size: 4,432. Structural equation modeling was used to estimate the genetic and environmental influences and common genetic influence across three waves of data collection. Within-and-between family association tests were used to test the effect of four candidate genes: DRD2, DAT1, DRD4, and 5HTT. Results: Uniivariate ACE model fitting (a² = additive genetic, c² = shared environmental, e² = non-shared environmental): Wave 1) a² -.52 e²-.48, Wave 2) a²-.28 e²-.72, Wave 3) a²-.30 e²-.70. Genetic correlations between Waves 1 and 2 were .85, Waves 1 and 3 were .34. The DAT1 440 allele was associated with average amount of alcohol consumed per drinking episode at Wave 1 (p = .007). DRD2 was associated at Wave 3 (p = . 007). DRD4 and 5HTT were not associated. The DRD2 and DAT1 polymorphisms accounted for 2.0% and 3.1% of the variation respectively. Conclusion: This study demonstrates a genetic influence on adolescent and young adult drinking behaviors and some common genetic influences across waves. Polymorphisms in genes of the dopaminergic system influence variation in this trait.

Marc Jamon and Michel Pratte. Neural adhesion cell disorders.

Functional genomics, pathology and behavior CNRS, Marseille

Plasticité et Physiopathologie de la Motricité, UMR 6196 CNRS / Université de la Méditerranée 31, Chemin Joseph Aiguier 13402 Marseille Cedex 20, France E-mail : jamon@dpm.cnrs-mrs.fr

Cell adhesion molecules (CAM) are transmembrane proteins that play a role in many functions involved in the development of neural circuitry, including neurite growth and fasciculation, axon guidance and synaptogenesis. The L1 family of CAM (L1CAMs) includes L1, CHL1, NrCAM, NgCAM, and neurofascin in vertebrates. The gene encoding the prototype of the family, L1 (on chromosomeXq28), is the target for mutations in the human mental retardation syndrome known as CRASH (acronym for corpus callosum agenesis, retardation, aphasia, spastic paraplegia, hydrocephalus) or L1 disease. Recently was generated a mutant mouse with a deletion for the gene coding for another member of the family, CHL1, the Close Homologue of L1. The deficiency in this molecule might be expected to produce L1 deficiency-related deficits, due to the homology and to a wide overlapping expression of the two molecules. In the present study, the CHL1-deficient mice displayed signs of decreased stress and a modification of exploratory behaviour. The mice also showed motor impairments on the Rotarod, but they were able to move as fast as controls in the alleys of a T-maze. The observed changes were assumed to be related to a deficit in attention or in novelty processing. The results are discussed in relation with motor and cognitive deficits in the human, and possibly schizophrenia, caused by mutations of the distal part of the chromosome 3 which contains the CHL1 ortholog.

Wendy Johnson, Matt McGue, and William G. Iacono¹. Genetic and environmental influences on academic achievement trajectories during adolescence².

¹Department of Psychology, University of Minnesota, Twin Cities, Minneapolis, MN USA, ²The Minnesota Twin Family Study is supported by National Institute on Drug Abuse Grant #DA05147 and National Institute on Alcohol Abuse and Alcoholism Grant #AA09367.

Address : Department of Psychology, University of Minnesota, 75 East River Road, Minneapolis, MN 55455 Telephone: 952-473-1673 Fax: 952-473-1998 Email: john4350@tc.umn.edu

Using the population-based Minnesota Twin Family Study, we investigated the main and interactive effects of child ability, motivation, depression, disruptive behavior, and family social risk at age 11 on academic achievement trajectories through high school. Hierarchical linear modeling showed main effects on initial level of achievement for all variables, though ability mitigated the effects of family risk. Only family risk showed a main effect on change in achievement over time, tending to impact it negatively. Girls showed higher overall achievement than boys, and the gap widened over time. Influences on achievement level were largely genetic (60% in girls, 50% in boys), and mediated primarily by genetic influences on ability, motivation, and disruptive behavior. Shared environmental influences on achievement level (25% in girls, 30% in boys) were mediated similarly. Influences on change were also largely genetic (60% in girls, 70% in boys) with the major contribution from genetic influences on family risk. This suggests that genetic and environmental influences tend to reinforce each other in the determination of academic trajectories. Interventions intended to improve academic achievement should reflect this.

Jaakko Kaprio¹, Urho M. Kujala² and Richard J Rose³. The development of health-related behaviors from adolescence to young adulthood - a longitudinal twin study⁴.

¹ Department of Public Health, University of Helsinki, Helsinki, Finland and Department of Mental Health and Alcohol Studies, National Public Health Institute, Helsinki, Finland, ² Department of Health Sciences, University of Jyväskylä, Jyväskylä, Finland, ³ Department of Psychology, Indiana University, Bloomington, Indiana, USA, ⁴ Supported by the Academy of Finland and AA-12502

Address: Dept. of Public Health, University of Helsinki, PO Box 41, FIN-00014 Helsinki, Finland, Telephone: +358 9 191 27 595 Fax: +358 9 191 27 600 Email: jaakko.kaprio@helsinki.fi

Physically active lifestyle is associated with many positive health habits, while alcohol abuse and smoking are associated with many negative outcomes. We studied twins to examine which associations between physical activity, other health habits and outcomes can be accounted for by childhood family environment or genetic factors. We studied how persistent physical activity in late adolescence (based on questionnaire studies at the ages of 16, 17 and 18.5 years in the FinnTwin16 study ) predicted questionnaire-reported health habits and outcomes at the mean age of 24 years. The final study group of 4240 individuals included 1870 twin pairs. Those who reported in all the three baseline questionnaires that their frequency of leisure physical activity was more than 3 times a week were classified as persistent exercisers and those who exercised less than 3 times a month were called persistently inactive, and others were occasional exercises. Follow-up outcomes included measures of physical activity, self-estimated health, daily smoking, use of illicit drugs, use of alcohol and relative weight. Pairwise analyses were done within pairs, comparing a physical active twin with the inactive co-twin. In analyses of both individuals and discordant co-twins persistent physical activity in adolescence was a significant predictor of high physical activity and very good self-estimated health at follow-up (p=0.003). Persistent physical inactivity compared to persistent activity predicted increased risk of daily smoking (p<0.001 for individual based analyses and p=0.017 for pairwise analyses) as well as increased risk of using drugs (p<0.001 and p=0.007, respectively). The associations between baseline physical activity and follow-up indices of overweight or use of alcohol were not found to be statistically significant in pairwise analyses. Persistent physical activity at late adolescence appears to be causally related to adult physical activity, self-estimated health, low smoking and low illicit drug use.

Ariel Knafo¹, Alessandra C. Iervolino¹ and Robert Plomin¹. Masculine Girls and Feminine Boys: Genetic and Environmental Contributions to Atypical Gender Development in Early Childhood².

¹Social, Genetic and Developmental Psychiatry Research Centre, Institute of Psychiatry, King=s College London, United Kingdom, ²Supported by a program grant (G9424799) from the UK Medical Research Council programme.

Address : Social, Genetic, and Developmental Psychiatry Research Centre, Box PO80, Institute of Psychiatry, De Crespigny Park, London SE5 8AF, UK. Telephone: +44 (0) 20 7848 0963. Fax: +44 (0) 20 7848 0895. E-mail: arknafo@hotmail.com.

In this first prospective genetic study of atypical gender role development, parents of 5799 same-sex pairs of twins from the Twins= Early Development Study (TEDS) reported on their twin children=s level of masculinity and femininity at 3 and 4 years. Boys were selected as gender-atypical if they were highly feminine (top 5%, 10%, or 15%) relative to other boys. Girls were selected as gender-atypical if they were highly masculine relative to other girls. Gender-atypical boys and girls were each divided into two groups: fully gender-atypical children (e.g., feminine boys also low on masculinity) and partially gender-atypical children (e.g., feminine boys who are not low on masculinity). DF extremes analysis yielded moderate group heritability and substantial shared environment effects for atypical gender role behavior at 5%, 10% and 15% cut-offs. However, for fully gender-atypical girls, group heritability accounted for most of the variance, and shared environment had no effect. The results are discussed in light of past studies and potential implications for atypical gender development.

Karestan C. Koenen^1,2,3. Sex differences in the relationship between IQ and antisocial behavior in young children.

¹Boston University School of Medicine, ²Boston University School of Public Health, ³National Center for PTSD, VA Boston Healthcare System. Supported by NIH GRANT 1K08MH70627

Address : Women=s Health Sciences Division (116B-3), National Center for PTSD, VA Boston Healthcare System, 150 South Huntington Street, Boston MA 02139

Telephone: 617 232 9500 x5913 Fax: 617 278 4515 Email: Karestan.Koenen@bmc.org

Early onset antisocial behavior is a strong risk factor for poor mental health, criminality, unemployment, and a host of other adjustment problems in adult life ( T.E. Moffitt, A. Caspi, H. Harrington, and B.J. Milne, 2002, Development & Psychopathology, 14, 179-207). Understanding the etiology of antisocial behavior in young children is necessary to inform prevention efforts, and therefore, remains an important public health goal. Low IQ is a consistent risk factor for antisocial behavior across the life course ( J.T. Nigg and C.L. Huang-Pollock, 2003, in B.B. Lahey, T.E. Moffitt, and A. Caspi, eds., Causes of Conduct Disorder and Juvenile Delinquency, Guilford: New York). Antisocial behavior in childhood is also more prevalent among males then females and evidence suggests compromised intelligence has greater effects on the development of antisocial behavior in males (T.E. Moffitt, A. Caspi, M. Rutter, and P.A. Silva, 2001, Sex differences in antisocial behavior: conduct disorder, delinquency, and violence in the Dunedin Longitudinal Study, Cambridge University Press: Cambridge, UK). This paper uses data from the Environmental Risk (E-Risk) Longitudinal Twin Study (N = 1116 pairs) to examine the sex differences etiology of the association between low IQ (age 5) and the development of antisocial behavior (age 7) in young children. IQ had a significant effect on the development of antisocial behavior (r = -.17, p< .001). This effect was significantly stronger in males (r = -.25, p< .001) than females (r = -.10, p< .05). Bivariate twin model-fitting revealed that the etiology of the covariance between IQ and antisocial behavior also differed in males and females. Common genetic influences explained 100% of the covariance between IQ and antisocial behavior in males. Alternatively, non-shared environmental influences explained 100% of the covariance in females. Findings are discussed in relation to sex differences in the etiology of antisocial behavior.

Yulia Kovas¹, Nicole Harlaar¹, Stephen A. Petrill², and Robert Plomin¹. Mathematics, Reading, and IQ: A Multivariate genetic Analysis in 7-year-old Twins ³

¹Social, Genetic and Developmental Psychiatry (SGDP) Centre, Institute of Psychiatry, King=s College London, UK, ² Department of Biobehavioral Health, The Pennsylvania State University, USA, ³Supported by the UK Medical Research Council (Grant G9424799)

Address : SGDP Centre, PO 83, Institute of Psychiatry, De Crespigny Park, London, SE5 8AF, UK Telephone: +44 (0)207-848-5403 Fax: +44 (0)207-848-0895 Email: y.kovas@iop.kcl.ac.uk

Mathematics performance in the early school years has recently been shown to be highly heritable as assessed by UK teachers following National Curriculum criteria. To what extent does this genetic influence on mathematics overlap with genetic influence on reading and on general cognitive ability (g or IQ)? For 2,825 pairs of 7-year-old same-sex twins, we used teacher assessments of mathematics performance, teacher assessments of reading performance and the Test of Word Reading Efficiency (administered by telephone), and a telephone-administered cognitive battery (adapted form Wechsler Intelligence Scale for Children) to assess g. Strong correlations were found between reading and mathematics (.70) and moderate correlations between mathematics and g (.43) and between reading and g (.47). Multivariate genetic analysis indicated substantial genetic overlap between mathematics, reading and g - genetic correlations ranged between .62 and .76. Nonetheless, fitting a trivariate Cholesky decomposition model to the data revealed the existence of both common and independent genetic influences for mathematics. That is, significant mathematics-specific genetic variance was found, as well as significant genetic variance specific to mathematics and reading but independent of g, and significant genetic variance in mathematics in common with reading and g. Shared and nonshared environmental influences only contributed modestly to the specific and common variance for all measures. We conclude that the same genes largely affect mathematics, reading, and g, although some specific genetic and, to a lesser extent environmental influences, also exist.

Tomas S Kubarych, Michael C Neale. Item analysis of structured clinical interview data.

Department of Psychiatry, Virginia Commonwealth University

Address : Box 980126 MCV Richmond VA 23298-0126 USA

Most research in psychiatry is based on structured clinical interviews, in which detailed Aprobe@ items are omitted unless subjects respond positively to Astem@ items. Item-level data obtained from such an instrument contain missing data which, if analyzed by methods that use listwise deletion, restrict analysis to cases with at least one positive screen, which may bias the results. Inclusion of subjects who do not receive the probes, however, introduces a class of subjects with zero variance on the probes.

Data from relatives yield a proxy form of information that allows estimation of the relationship between skipped items and the probes. Using data collected from adult female twins in Virginia, we examined criteria for generalized anxiety disorder using a causal, contingent, common pathway model. Direct effects between stems and probes are first estimated. Multivariate analyses using similar technology estimated the relationship between the stem and multiple probe questions.

Janine A Lamb¹, Elena Bonora¹, Gaby Barnby¹, Nuala Sykes¹, Elena Maestrini², Francesca Blasi², Elena Bacchelli², Kim S Beyer³, Sabine M Klauck³, Annemarie Poustka³, Anthony J Bailey⁴, Anthony P Monaco¹, International Molecular Genetic Study of Autism Consortium (IMGSAC)⁵.

Towards identification of autism susceptibility variants in the IMGSAC sample.

¹Wellcome Trust Centre for Human Genetics, University of Oxford, Oxford, UK, ² University of Bologna, Department of Biology, Bologna, Italy, ³Department of Molecular Genome Analysis, Deutsches Krebsforschungszentrum, Heidelberg, Germany, ⁴ Section of Child and Adolescent Psychiatry, Park Hospital for Children, Oxford, UK.

⁵ http://www.well.ox.ac.uk/~maestrin/iat.html

Autism is a severe neurodevelopmental disorder, likely to arise on the basis of a complex genetic predisposition. In order to identify susceptibility genes, the IMGSAC has performed a genome screen for linkage in affected sibling pairs (ASP) and identified putative susceptibility loci on chromosomes 2,7,9,16 and 17. These findings are supported in part by cytogenetic abnormalities in autistic individuals and linkage findings from independent genome screen studies. However, the peaks of linkage remain broad, often encompassing hundreds of genes. In the absence of convincing association data, and in order to improve localisation of linkage signals, researchers have made efforts to reduce clinical and genetic heterogeneity. Analysis of the IMGSAC data set according to ASP sex has provided evidence for sex-limited effects and a possible parent-of-origin specific effect on chromosome 7. Functional candidate genes mapping to the regions of linkage on chromosomes 2, 7 and 16 are being systematically screened to identify variants that may contribute to autism aetiology. All coding regions and putative regulatory sequences have been screened by DHPLC and sequencing in affected individuals from families contributing to the linkage. The DNA variants identified have been tested for association with autism by case-control and/or TDT studies in the whole IMGSAC family sample. The finding of rare missense variants in the Reelin gene on chromosome 7q22.1 and in cAMP-GEFII on chromosome 2q32 suggests that these two genes may have a role in autism aetiology in a small proportion of cases, and requires further investigation. However, the clear involvement of any gene(s) in the large proportion of IMGSAC families remains to be established. Targeted high resolution association mapping using a high density of single nucleotide polymorphisms across these regions will enable advances towards the identification of autism susceptibility genes.

Kate Langley¹, Frances Rice¹, Darko Turic¹, Marianne Van den Bree¹, Michael, J. Owen¹, Michael C. O’Donovan,¹ Peter, A. Holmans¹, Anita Thapar¹. Possible interaction between maternal smoking during pregnancy and DRD4 genotype influencing severity of ADHD diagnosis².

¹Dept. of Psychological Medicine, University of Wales College of Medicine, Cardiff, United Kingdom. Supported by Wellcome Trust, SPARKS and ACTION RESEARCH. We thank all the families, clinicians and field workers for their help

Address : Department of Psychological Medicine, University of Wales College of Medicine, 4^th Floor, Heath Park, Cardiff, UK, CF14 4XN Telephone: 0044 (0)2920 742 934 Fax: 0044 (0)2920 747 839 Email: langleyk@cardiff.ac.uk

The 7-repeat allele of a VNTR in the DRD4 gene has been found to be, significantly associated with Attention Deficit Hyperactivity Disorder (ADHD) in a recent meta-analysis (Faraone et al., 2001 American Journal of Psychiatry.158(7):1052-7. ). Influence of environmental factors including maternal smoking during pregnancy (Linnet et al., 2003 American Journal of Psychiatry;160(6):1028-40) has also been recognised. We investigated possible interplay between DRD4 genotype and smoking during pregnancy and their association with diagnosis of ICD-10 Hyperkinetic Disorder (HKD) which has been proposed as a more severe, and possibly more neurobiological manifestation of the disorder (Taylor et al, 1991, The Epidemiology of Childhood Hyperactivity, Oxford University Press, Oxford,UK).111 children aged 6-16 diagnosed with ICD-10, DSM-IV or DSM-III-R ADHD participated in this study. All probands were genotyped for the DRD4 variant. Information regarding smoking during pregnancy was obtained retrospectively from mothers. Main and interaction effects of DRD4 genotype and smoking status were analysed using logistic regression and chi-square analysis. Stepwise logistic regression analysis revealed that, although there were no main effects of DRD4 or smoking status on ICD-10HKD diagnosis, there was a significant statistical interaction between DRD4 genotype x smoking (p=0.049, OR=2.288). Further analysis revealed that an association between smoking and increased ICD-10HKD was only present in children with the DRD4 7-repeat allele (7*-present, n=41, )²=5.306, p=0.023, 7*-absent n=68, )²=0.402, p=0.354). This analysis reveals a significant interaction between child=s DRD4 genotype and maternal smoking during pregnancy which influences severity of ADHD diagnosis. When the 7-repeat susceptibility allele is present, it appears to moderate association between maternal smoking during pregnancy and ADHD, resulting in the more severe, narrower diagnosis of the disorder. This finding may have significant implications regarding advice given to pregnant smokers, whilst highlighting the importance of analysing genetic and environmental factors together when investigating influences for ADHD.

Henrik Larsson¹, Jan-Olov Larsson², and Paul Lichtenstein¹. Differences in DSM-IV subtypes of ADHD: Genetic and Environmental Influences.

¹ Department of Medical Epidemiology and Biostatistics, Karolinska Institutet, Stockholm, Sweden, ² Department of Child & Adolescent Psychiatry, The Karolinska Hospital, Stockholm, Sweden.

Address : Box 281, SE-171 77 Stockholm, Sweden, Telephone: +46-8-524 87416 FAX: +46-8-31 49 75, E-mail: Henrik.Larsson@meb.ki.se

Previous behavior genetic studies of Attention-Deficit Hyperactive Disorder (ADHD) have not considered the developmental changes in symptoms for individuals from childhood to adolescence. Therefore, the aim of this study was to elucidate the etiology of the development among and within the three subtypes of ADHD in DSM-IV, primarily inattentive, primarily hyperactive-impulsive, and combined subtypes.

We used a population-based Twin study of Child and Adolescent Development (TCHAD) which comprises 1500 Swedish twins, first contacted when they were 8-9 years old and followed up when they were 13-14 and 16-17 years old. Multivariate genetic analyses were applied to study the importance of persistent, time-specific, and subtype-specific genetic and environmental effects for parental-rated symptoms of Inattention and Hyperactivity-Impulsivity.

The results suggested a genetic factor influencing persistent symptoms of the combined type at all three time-points (age 8-9, 13-14 and 16-17). Persistent subtype-specific genetic and environmental factors were also important. Moreover, there was a genetic factor specific to each time-point.

The findings support the DSM-IV conceptualization of the subtypes for ADHD. Genetic influences explain in principle all variation for persistence in the combined subtype, whereas both genes and environments are responsible for the persistence of the Hyperactive-Impulsive and Inattentive subtypes. The genetic time-specific variation in the expression of symptoms suggests that Aan immaturity factor@ and new effects during and/or after puberty are also important.

Jennifer Y.F. Lau¹ and T. C. Eley¹. Adolescent depression: interactions, correlations and cognitions².

¹ SGDP Centre, Institute of Psychiatry, King=s College London, London, UK, ² The G1219 study was supported by the W T Grant Foundation and by a Medical Research Council training fellowship to the second author. The first author is supported by a Medical Research Council doctoral studentship.

Address : SGDP Centre, Box PO 83, Institute of Psychiatry, De Crespigny Park, London SE5 8AF Telephone: (00-44) 207 848 5414 Email: j.lau@iop.kcl.ac.uk

Emerging evidence suggests that genetic and environmental factors interact (G x E) and correlate (rGE) to produce vulnerability for adolescent depression (for a review see J.Y. Lau and T.C. Eley, in press, Cur Psychiatry Rep, 6). Consistent with these processes of interplay, a negative attributional style is thought to interact with social stressors to precipitate symptoms (L.Y. Abramson, M. E. Seligman, & J.D. Teasdale, 1978, J Abnorm Psychol, 87, 49-74) and to be influenced by parenting styles (J. Garber and C. Flynn, 2001, Cognit Ther Res, 25, 353-376), which may reflect both genetic and environmental risks associated with parental genotype. Thus attributional style may be involved in these risk mechanisms of interplay. Self-reported data on attributional style, depression symptoms and maternal negative control were collected from over 1300 adolescent twin and sibling pairs, and used to examine two genetic models. The first tested for the moderation of genetic risk to depression symptoms by negative maternal control (G x E), whilst also specifying a genetic influence on this environmental variable (rGE) (S. Purcell, 2003, Twin Res, 5. 554-571). The second decomposed the phenotypic covariation between attributional style and both depression and maternal negative control into genetic and environmental influences. Initial results show that genetic effects on depression are moderated by negative maternal control (G x E), and that these genetic influences were also those implicated in the aetiology of this parenting stressor (rGE). The second set of findings demonstrated roughly equal proportions of genetic and environmental contributions to the association between attributional style and depression. Similarly the covariation between attributional style and maternal negative control also reflected genes and environmental factors. These results reinforce the importance of gene-environment interactions and correlations on adolescent depression. Moreover, there is indirect support that a negative attributional style may be involved in these risk mechanisms.

Jeffrey M. Lessem¹, Brett C. Haberstick¹, Christian Hopfer², Andrew Smolen¹, David Timberlake¹, John K. Hewitt^1. Relation between early use of marijuana and later illicit drug use in the Add Health sample.

¹Institute for Behavioral Genetics, University of Colorado, Boulder USA, ²Div. of Subst. Dependence, Univ. of Co. Health Sciences Ctr., Denver. Supported by P01-HD31921 (PI: JR Udry), EY-12562 (PI: JK Hewitt).

Address: Institute for Behavioral Genetics, 447 UCB, Boulder, CO 80309-0447 USA Telephone: +1 (303) 492-2843 Email: Jeff.Lessem@Colorado.EDU

In the Add Health study a school based design was used to assess health related information on adolescents in grades 7-12. Interviews were conducted twice, about one year apart, between 1994 and 1996, creating Waves I and II. Wave III of data collection was undertaken six years later, and was released in 2003. Marijuana use before the age of 17 was found to be a significant predictor of illicit drug use at Wave III, when the mean age of the sample is 22 years old. The variance in marijuana use before age 17 is due to genetic, shared environmental, and non-shared environmental factors, while the variance in use of illicit drugs at Wave III is due to genetic and non-shared environmental factors.

Kate J Lifford¹, K. Langley¹, D. Turic¹, M. B. Van den Bree¹, J. Williams¹ , M. J. Owen¹ & M. C. O=Donovan¹ A. Thapar¹. Assessing the influence of GRIN2A on reading ability in attention-deficit/hyperactivity disorder².

¹Department of Psychological Medicine, University of Wales College of Medicine, Cardiff, UK, ²Data collection funded by Wellcome Trust. Kate Lifford sponsored by Biostatistics/Bioinformatics Unit (funded by Higher Education Funding Council for Wales).

Address: Department of Psychological Medicine, University of Wales College of Medicine, Heath Park, Cardiff, CF14 4XN Telephone: 029 20742934 Fax: 029 20747839 Email: liffordkj@cardiff.ac.uk

One of the regions of linkage suggested in a genome-wide scan for attention-deficit/hyperactivity disorder (ADHD) and reading disability is on chromosome 16p (S. K. Loo et al, 2003, Molecular Psychiatry, 1-9). GRIN2A, a gene found in this region, is a candidate gene in which a polymorphism on exon 5 has been found to be associated with ADHD (D. Turic et al, 2004, Molecular Psychiatry, 9, 169-173). This study included 161 ADHD children, a subset of the sample in which the association with GRIN2A has been found. Reading ability was measured in this sample using mean score of the three sub-tests of the Weschler Objective Reading Dimension (Weschler, 1993, The Psychological Corporation, UK). Multiple regression analysis showed that association of the allelic variation of the GRIN2A polymorphism with reading score was not significant (F_{(1, 129)} = 0.182, p=0.670). The genetic variation in GRIN2A therefore does not appear to account for phenotypic variation in reading ability in children with ADHD.

Michelle Luciano¹, Mark Wainwright¹, Margaret J. Wright¹, Gina M. Geffen², David L. Duffy¹, Nicholas G. Martin¹. Using multivariate linkage methods to locate QTLs for word recognition and IQ³.

¹ Genetic Epidemiology, Queensland Institute of Medical Research, Brisbane, Queensland, Australia, ² Cognitive Psychophysiology Laboratory, University of Queensland, Brisbane, Queensland Australia, ³ Collection of phenotypes and DNA sample was supported by grants from the Australian Research Council and the Human Frontier Science Program. The genome scans were supported by the Australian NHMRC=s Program in Medical Genomics [NHMRC-219178] and a grant to Dr Jeff Trent from the Center for Inherited Disease Research at Johns Hopkins University. CIDR is fully funded through a federal contract from the National Institutes of Health to The Johns Hopkins University, Contract Number N01-HG-65403.

Address : Genetic Epidemiology, Queensland Institute of Medical Research, PO Box Royal Brisbane Hospital, 4029, Herston Australia Telephone: +61 7 3362 0218, Fax: +61 7 3363 0101, Email: michelLu@qimr.edu.au

In a sample of 500 adolescent twin pairs we have previously shown that the covariation between IQ subtest scores and tests of word recognition is primarily mediated by genes. While most of the genetic variance in IQ (29-57%) and whole-word reading (30-40%) measures is explained by this common genetic factor, unique genetic influences are also significant (M. Wainwright et al, in press, Behav. Genet.). Our aim now is to find QTLs influencing the variation in these measures. DNA has been collected from twins, their siblings and available parents (>80%) and a genome scan of 400-761 autosomal markers has been completed for 342 families (includes between 2 and 5 siblings) with IQ and reading phenotypes. Here we report our first results for multivariate genome-wide linkage of IQ (Multidimensional Aptitude Battery: verbal and performance scales) and word-recognition (Schonell Graded Word Reading Test, Cambridge Contextual Reading Test) measures. In a genome-wide linkage analysis of reading and language-related phenotypes, multivariate methods were shown to confer greater statistical power compared to the univariate case (A.J. Marlow et al., 2003, Am. J. Hum. Genet.,72, 561-570). Our findings also find support for advantages of multivariate over univariate linkage methods, generally confirming gains in power plus the moderation of apparent biases to a single phenotype. Suggestive linkage on the long arm of chromosome 2 was indicated. We intend to broaden our range of reading measures to include tests which tap diverse components of reading processing (orthographic skill & phonological decoding) and are predictive of reading disability subtypes.

Stacy K. Lynch¹, Eric Turkheimer¹, Robert E. Emery¹, Brian M. D=Onofrio¹, Jane Mendle¹, Wendy S. Slutske², Andrew C. Heath³, & Nicholas G. Martin⁴. A Genetically Informed Study of the Association Between Harsh Punishment and Offspring Behavioral Problems

¹Department of Psychology, University of Virginia, Charlottesville VA USA. ²Department of Psychology, University of Missouri B Columbia, Columbia MO USA, ³School of Medicine, Washington University, St. Louis MO USA, ⁴Queensland Institute of Medical Research, Brisbane, Australia. Supported by William T. Grant Foundation, Grant # 2054

Address : Dept. of Psychology, Box 400400 University Of Virginia

Charlottesville, VA 22903 USA Telephone: (434) 962-2836

E-mail: akl8t@virginia.edu

Although harsh punishment styles have been associated with a variety of negative outcomes in children, the definition of harsh punishment remains inconsistent. Moreover, studies of the putative causal effects of harsh parenting on children have been hindered by uncontrolled third variables and environmental contexts. The present study uses a children of twins design to analyze the consequences of various degrees and definitions of punishment, ranging from mild non-physical punishment to abusive physical punishment. Forms of punishment associated with high levels of offspring pathology are further examined within a hierarchical linear modeling (HLM) framework, which facilitates the discriminate of any specific effects of punishment style from genetic or environmental variables that may confound the relationship. Results indicate that corporal punishment in general does not have significant associations with negative childhood outcomes, but other more harsh forms of physical punishment do appear to have specific causal effects.

Faïza Ben Mabrouk, Edouard Pearlstein, Jean-Françios Pflieger and Laurent Vinay. Serotoninergique shaping of the locomotor pattern in the in vitro neonatal rat spinal cord.

ADDRESS : CNRS, Marseille, Plasticité et Physiopathologie de la Motricité, UMR 6196 CNRS / Université de la Méditerranée 31, Chemin Joseph Aiguier 13402 Marseille Cedex 20, France Email : faiza@dpm.cnrs-mrs.fr

The serotoninergic projections from raphe nuclei arrive in the lumbar enlargement of the spinal cord during the late fetal period in the rat, a time window during which the locomotor pattern switches from left/right (L/R) and flexor/extensor (F/E) synchrony to alternation. The goal of the present study was to investigate the role played by serotonin (5-HT) in shaping the locomotor pattern, i.e. L/R and F/E alternations. Fictive locomotion was induced by bath application of N-methyl-D,L-aspartate (NMA) in the in vitro neonatal rat spinal cord preparation. By means of cross-correlation analysis we show that 5-HT, when added to NMA, strengthens L/R and F/E (recorded from the 3^rd and 5^th lumbar ventral roots, respectively) alternations. This effect was partly reproduced by activation of 5-HT_2A/2C receptors[-]. In a last series of experiments, we tested the contribution of endogenous 5-HT to NMA-induced fictive locomotion. Reducing the functional importance of endogenous 5-HT, either by inhibiting its synthesis with daily injections of p-chloro-phenylalanine (PCPA), starting on the day of birth, or by application of ketanserin (á 5-HT₂ receptor antagonist) or SB269970 (á 5-HT₇ receptor antagonist), disorganized the NMA-induced locomotor pattern. This pattern was restored in PCPA-treated animals by adding 5-HT to the bath. These results demonstrate that 5-HT is critical for shaping the locomotor pattern in the neonatal rat. This action may be exerted through a potentiation of reciprocal inhibitory connections.

Hiroko Maekawa¹, Juko Ando², and Yutaka Ono³. The structure of eating behavior and attitudes among Japanese adolescent women.

¹Department of Education, Graduate school of human relations, Keio University, Minato-ku Tokyo Japan, ² Faculty of letters, Keio University, Minato-ku Tokyo Japan, ³ Health Center, Keio University, Yokohama Kanagawa Japan.

Address : 5-19-9-304, Nishikamata, Ota-ku, Tokyo 144-0051 Japan

Telephone: 81 3 3734 5125 Email: hm-hiro@pop21.odn.ne.jp

Recent research has indicated that most of Japanese adolescent women think themselves fat, and they want to be thinner, although, in reality they are thinner than they were 20 years ago. It has been known that the weight and shape concerns are the risk factors of eating disorders. Many studies about disordered eating behavior have been conducted, but the genetic and environmental structure of eating behaviors and attitudes is unclear. To reveal the structure of eating behaviors and attitudes, 416 adolescent female twins (MZ:316 pairs, DZ:100 pairs) from the Keio Twin Project were examined with the Eating behavior Eating Disorder Inventory (EDI; Garner, Olmsted, & Polivy, 1983) which assessed Drive for Thinness, Bulimia, and Body Dissatisfaction. The pairs whose BMI is over 25 were excluded to examine the structure of eating behavior and attitudes among women who don=t need to lose their weight. Results obtained from Cholesky decomposition indicated that a substantial amount of the covavriance between Drive for Thinness, Bulimia, and Body Dissatisfaction was accounted for by an additive genetic factor. Specifically, additive genetic factor accounted for approximately 64% of the covariance between Drive for Thinness and Bulimia, 60% of the covariance between Drive for Thinness and Body Dissatisfaction. Non-shared environmental factor accounted for 36% and 40%, respectively. In contrast, the covariance between Bulimia and Body Dissatisfaction was accounted for 97% by non-shared environmental factor. The results showed that additive genetic factor made the largest contribution to the covariation among eating behaviors and attitudes.

Hermine H. Maes, Michael C. Neale, Judy L. Silberg, Debra L. Foley, Lindon J. Eaves¹. Genetic and environmental transmission of retrospective conduct disorder².

¹Virginia Institute for Psychiatric &amp; Behavioral Genetics, Department of Human Genetics and Psychiatry, Virginia Commonwealth University, Richmond VA 23298, USA, ²Supported by NIH grants HL60688, MH45268, MH55557 and MH57761.

Address : VIPBG, MCV/VCU, P.O. Box 980003, Richmond VA 23298-0003 Telephone: 804 828 8145 Fax: 804 828 8801 Email: hmaes@vipbg.vcu.edu URL: http://www.vipbg.vcu.edu

With the recent attention on gene by environment interactions including parental treatment, it is important to disentangle genetic from cultural transmission of antisocial behavior. The classical twin design augmented with parental data on the same phenotype allows such a partitioning of variance . The goals of this study were to quantify the genetic and non-genetic sources of inter-generational transmission for conduct disorder and to extend the twin-parent design to include biological and non-biological parents. Data on conduct disorder before age 18 (retrospective conduct disorder) from the Young Adult Follow Up (YAFU) of the Virginia Twin Study of Adolescent Behavioral Development (VTSABD) were analyzed using Mx. Given a quarter of the parental informants in the VTSABD are non-biological, we extended the twin-parent model to include not-related caregivers. Familial resemblance for retrospective conduct disorder was explained by additive genetic (h²~35%) and shared environmental factors (c²~20-30%) in biological families. Including the non-biological parents resulted in a moderate increase in heritability and decrease in the shared environmental contribution. While cultural transmission parameters could be dropped without loss of goodness-of-fit, assortative mating was significant. Sex differences were more pronounced for non-parental shared environment. We conclude that genetic factors accounted for the majority of the parent-offspring transmission of retrospective conduct disorder.

Sergei B. Malykh¹, Elena D. Gindina¹, Viktoria V. Nadyseva¹. The developmental changes in sources of individual differences in adolescent temperament.

¹Laboratory of Developmental Behavior Genetics, Psychological Institute of Russian Academy of Education, Moscow, Russia

Address: Psychological Institute of Russian Academy of Education, Mokhovaya street, 9 AV@, Moscow, 125009 Russia Telephone: +7(095)202-9363 Fax +7(095)134-2119 Email: malykh@pirao.ru

The study was performed as a part of a Moscow longitudinal twin project and aimed to explore genetic and environmental contributions to temperament of Russian adolescents in pubertal and post-pubertal periods. 53 pairs of monozygotic (MZ) and 55 pairs of same-sex dizygotic (DZ) Russian twins completed the children version of the Structure of Temperament Questionnaire (C-STQ; Rusalov, V.M., Oprosnik struktury temperamenta. [A Temperament Structure Questionnaire.] Methodological manual, Moscow: Smysl, 1992. In Russian.) at the age of 12-14 years and four years later. The correlation and model fitting analysis showed significant differences in the degree of genetic determination of temperament dimensions in two age periods. At the age of 12-14 years, significant genetic influences were revealed only on 3 of the C-STQ dimensions - Social Tempo, Emotionality and Social Emotionality. The best fitting model for these traits was simple genetic model. The heritability estimates for Social Tempo, Emotionality and Social Emotionality were 60%, 66% and 66% - respectively; the remaining variance was explained by non-shared environment effects. Individual differences in the other C-STQ dimensions were due to environmental factors. At the age of 16-18 years, evidence for moderate to high genetic influence was seen for all C-STQ dimensions. Under the best fitting simple genetic model, the genetic influences accounted for 64% of the variance in Object-related Ergonicity, 42% - in Social Ergonicity, 27% - in Object-related Plasticity, 40% - in Social Plasticity, 32% - in Object-related Tempo, 47% - in Social Tempo, 43% - Object-related Emotionality and 37% - in Social Emotionality. Non-shared environmental effects explained the rest of the total variance in these dimensions. A substantial increase in genetic influences and a declining influence of shared environmental factors during the transition from pubertal to post-pubertal period was related to decline of temperament resemblance among DZ twins.

Carol A Manning¹, Lillian J Currie¹, Madeline Harrison¹, G. Frederick Wooten¹, and Eric N. Turkheimer². Intellectual Impairment Differences in Parkinson=s Patients With And Without Affected Mothers².

¹Department of Neurology, University of Virginia, Charlottesville, VA USA, ²Department of Psychology, University of Virginia, Charlottesville, VA USA, ³Supported by the American Parkinson's Disease Association Advanced Center for Parkinson's Research B University of Virginia and Morris K. Udall Parkinson's Disease Research Centers of Excellence (NINDS-P50-NS39788).

Address : Department of Neurology, Box 800394, University of Virginia, Charlottesville, VA 22908 Phone: (434) 982-1012 Fax: (434) 982-1996 Email: cm4r@virginia.edu

Genetic factors play a role in the expression of Parkinson=s Disease (PD) but the mechanism of inheritance remains unclear (Currie LJ. Harrison MB. Trugman JM. Bennett JP. Swerdlow RH. Manning CA. Wooten GF in Journal of Neurology, Neurosurgery & Psychiatry. 71(1):130-1, 2001 Jul; Marder K. Levy G. Louis ED. Mejia-Santana H. Cote L. Andrews H. Harris J. Waters C. Ford B. Frucht S. Fahn S. Ottman R. in Annals of Neurology. 54(4):507-13, 2003 Oct). In addition, intellectual impairment is being recognized as an important clinical feature in some individuals with PD. Although intellectual impairment has been associated with increased age of onset, little is known about the relationship between inheritance of PD and cognitive symptoms. We examined intellectual impairment in our database of 629 carefully characterized individuals with PD. Patients were organized into three groups of probands: 1) PD patients with an affected mother (N=45), 2) PD patients with an affected father (N= 33) and 3) PD patients without a first-degree relative with the disease (N = 551). The average age at disease onset, number of years with the disease, and overall disease severity as measured by the Unified Parkinson=s Disease Rating Scale (UPDRS) were similar for all groups. Logistic regression revealed that probands with an affected mother were significantly more likely to have intellectual impairment, as measured by the UPDRS, than were probands without an affected mother (p <0.05). However, the groups did not differ in severity of motor symptoms, degree of depression or presence of thought disorder. These results suggest that disease phenotype is related to mode of inheritance and that maternal inheritance of PD is associated with an increase in cognitive symptoms.

Cecilia Marino¹, L Vanzin¹, R Giorda¹, A Frigerio¹, M L Lorusso¹, M Nobile¹, M Molteni¹, M Battaglia². An assessment of transmission disequilibrium between quantitative measures of childhood problem behaviors and DRD2/TaqI, DRD2/-141C Ins/Del and DRD4/48bp-repeat polymorphisms.

¹ Eugenio Medea Scientific Institute, Department of Child Psychiatry, Bosisio Parini, Italy, ² San Raffaele Scientific Institute, Milan, Italy.

Address: Eugenio Medea Scientific Institute, Department of Child Psychiatry, Via Don Luigi Monza 20, Bosisio Parini, Lecco, Italy. Tel ++39-031-877-381 cmarino@bp.lnf.it

In a pilot study of 120 children with reading disabilities we assessed the presence of linkage disequilibrium between the DRD2/TaqI and the DRD4/48bp-repeat polymorphisms and quantitative measures of behavioral problems derived from parental rated Child Behavior Checklist (CBCL) (Achenbach, T.M. Manual for the CBCL/4-18 and 1991 Profile, 1991. Burlington, VT: University of Vermont Department of Psychiatry).

Analyses included measures of between-individuals association, and a test of the presence of linkage disequilibrium by a logistic regression-based extension of the Transmission Disequilibrium Test (logistic regression QTDT) (Waldman, I.D., Robinson, B.F., Rowe, D.C.,1999, Ann. Hum. Genet. 63: 329-340). In between-individuals association analyses the AWithdrawn@ scale of CBCL was related to the number of risk alleles of DRD4 and the ASocial Problem@ scale was associated with the number of risk alleles of DRD2 (the 7-repeat and the A1 allele respectively were considered the risk alleles). Logistic regression QTDT yielded statistically significant evidence in favor of linkage disequilibrium between DRD2 and ASocial Problems@ (Wald χ²= 4.13, df=1, p= .042, Odds Ratio= 1.97). No linkage disequilibrium was found for AWithdrawn@ and DRD4 (Wald χ²= 2.64, df=1, p= .1, Odds Ratio= 1.44). Although preliminary, these findings are in harmony with previous data that suggest a role of the same polymorphism in influencing individual sensitivity to reward and response to social clues and reinforcements in man and animal (Blum, K., Braverman, E.R., Holder, J.M., Lubar, J.F., Monastra, V.J., Miller, D., Lubar, J.O., Chen, T.J., Comings, D.E., 2000, J. Psychoactive Drugs. 32 Suppl:i-iv, 1-112). To further investigate the imporance of the DRD2 gene, quantitative haplotype analyses were run with haplotype B141C Ins/Del-DRD2/TaqI and dimensions of the CBCL (Horvath S., Xu X., Lake S.L., Silverman E.K., Weiss S.T., Laird N.M., 2004, Genet Epidemiol. 26, 61-9).

Nicholas G. Martin, Gu Zhu, David L. Duffy. A genome scan for eye color using a full two-locus model

Queensland Institute of Medical Research, Brisbane, Australia. Collection of phenotypes and DNA samples was supported by grants from the Queensland Cancer Fund, the Australian National Health and Medical Research Council, and the U.S. National Cancer Institute (CA88363). The genome scans were supported by the Australian NHMRC=s Program in Medical Genomics [NHMRC-219178] and a grant to Dr Jeff Trent from the Center for Inherited Disease Research at Johns Hopkins University. CIDR is fully funded through a federal contract from the National Institutes of Health to The Johns Hopkins University, Contract Number N01-HG-65403.

Address : QIMR, Brisbane 4029, Australia, fax 61-7-3362 0101, Telephone 61-7-3362 0278, Email: nickm@qimr.edu.au

We have rated eye colour on a three point scale (1 = blue/grey, 2 = hazel/green, 3 = brown) in 502 twin families and carried out a 5-10cM genome scan (400-757 markers). We analysed eye colour as a threshold trait and performed multipoint sib pair linkage analysis using variance components analysis in Mx. A lod of 19.2 was found at the marker D15S1002, less than 1cM from OCA2, which has been previously implicated in eye colour variation. We estimate that 74% of variance in eye colour liability is due to this QTL and a further 18% due to polygenic effects. However, a large shoulder on this peak suggests that other loci affecting eye colour may be telomeric of OCA2 and inflating the QTL estimate. No other peaks reached genome-wide significance, although lods >2 were seen on 5p and 14q and lods>1 were additionally seen on chromosomes 2, 3, 6, 7, 8, 9, 17 and 18. Most of these secondary peaks were reduced or eliminated when we repeated the scan as a two locus analysis with the 15q linkage included, although this does not necessarily exclude them as false positives. We also estimated the interaction between the 15q QTL and the other marker locus but there was only minor evidence for additive x additive epistasis. Elaborating the analysis to the full two locus model including non-additive main effects and interactions did not strengthen the evidence for epistasis. We conclude that most variation in eye colour in Europeans is due to polymorphism in OCA2 but that there may be modifiers at several other loci.

Motoko Matsuura¹, Masumi Sugawa², Atsushi Sakai³, Taketoshi Takuma⁴, Sachiko Amou⁵. Genetic and environmental influences on career attitudes and vocational preparation in Japanese junior high school and high school students.

¹Development of Human Developmental Science, Graduate School of Human Culture, Ochanomizu University Japan, ²Department of Psychology, Faculty of Letters and Education, Ochanomizu University Japan, ³Department of Psychology and Education, Faculty of Education Human Science, University of Yamanashi Japan, ⁴Graduate School of clinical psychology, Tokyo International University, ⁵Aoyama Institute of Education Japan.

Address : 3-26-12-305, Futaba, Shinagawa-ku,Tokyo,142-0043, Japan Telephone:03 5978 5279 Fax:03 5978 5270 E-mail: g0370324@edu.cc.ocha.ac.jp

This study examined the genetic and environmental contributions on career attitudes and vocational preparation . The Japanese twin sample of junior high school and high school students (120 pairs of MZ and 99 pairs of DZ ) completed the questionnaires about their career attitudes and the level of their vocational preparation. Univariate genetic analyses were indicated that CE models yielded the best fit in both of career attitudes and vocational preparation. This study showed that eating attitudes and behaviors were influenced by gender and age differences.

John J. McArdle ¹ and Fumiaki Hamagami ¹. The addition of biometric components to longitudinal dynamic structural models of growth and survival².

¹ Department of Psychology, University of Virginia, Charlottesville VA USA, ² Supported by NIH Research Grant #AG-07407 from the National Institute on Aging.

Address : P.O. Box 400400, University of Virginia, Charlottesville VA 22904 USA Telephone: (434) 924-0656; Fax: (434) 982-4766; Email: jjm@virginia.edu

There are now a number of studies in human aging research where longitudinal data have been collected on adults who are twins or members of the same families (e.g., SATSA). The analyses of these longitudinal data initially emphasized the description of patterns of growth and change over time, including the estimation of biometric sources of variation. In more recent analyses, there has been an emphasis on age as a focal basis of change, an effort to understand the lead-lag relationships across different phenotypes and genotypes, and the importance of accounting for survival (mortality) of individuals and families. In this presentation we examine some ways in which all of these components can be estimated in a joint analysis of the same longitudinal data. The structural models of growth and change are based on the recent approach for longitudinal twin data presented in McArdle & Hamagami (Behavior Genetics, 33 (3): 137-1592003). The survival-frailty models for twin data are initially based on the recent work of Yashin, Iachine, & Harris (Behavior Genetics, 29 (1):11-19, 1999). We next consider several recent proposals for combining latent growth and survival models, including the useful overview of Lin, McCulloch, & Mayne (Statistics in Medicine, 21: 2639-2382, 2002) and the recent contributions of Guo & Carlin (American Statistician, 58 (1): 1-9, 2004). We highlight the latter because standard software for maximum likelihood estimation is presented (i.e., NLMIXED, winBugs). To combine the most useful features of all longitudinal models we: (1) add biometric components to the growth/survival components, (2) examine several alternative models for their inter-relationships, (3) illustrate the accuracy and reliability of this approach using standard software, and (4) examine a few simpler but practical alternatives. Our results highlight both the potential benefits and the potential limitations of integrated structural analyses based of all these popular longitudinal considerations.

Matt McGue¹, Margaret Keyes¹, Irene Elkins¹ & William G Iacono¹. Shared Environmental Effects B Do They Endure?

¹Department of Psychology, University of Minnesota. Supported by NIAAA grant # AA11886

Address : 75 East River Road, Minneapolis, Minnesota 55455 USA

In 1987, Plomin and Daniels showed that siblings growing up in the same home bore little psychological similarity to one another if they were not genetically related. This provocative observation has stimulated much research, which is generally supportive of the general conclusion that shared environmental influences are minimal or non-existent. Two behavioral domains, cognitive abilities and antisocial behavior, appear to run counter to the general conclusion regarding the relative unimportance of shared environmental effects. Nonetheless, it remains unclear whether shared environmental effects on these behaviors endure much beyond the period of common rearing. Using data from the ongoing follow-up of a large cohort of adoptive sibling pairs we will explore evidence for the durability of shared environmental effects. We will interpret study findings in the context of evolving notions of the nature of family socialization.

Sarah Medland^1,2, Danuta Loesch³, Bogdan Mdzewski⁴, Gu Zhu¹, Nicholas G Martin¹. A multivariate genome scan for finger ridge count.

¹Queensland Institute of Medical Research, Australia, ²University of Queensland, Australia, ³La Trobe University Australia, ⁴ Bogdan Mdzewski died on the 30/11/2001. Collection of phenotypes and DNA samples was supported by grants from the Australian National Health and Medical Research Council. The genome scans were supported by the Australian NHMRC=s Program in Medical Genomics [NHMRC-219178] and a grant to Dr Jeff Trent from the Center for Inherited Disease Research at Johns Hopkins University. CIDR is fully funded through a federal contract from the National Institutes of Health to The Johns Hopkins University, Contract Number N01-HG-65403.

MAIL: QIMR, Brisbane 4029, Australia, fax 61-7-3362 0101, phone 61-7-3845 3551, email: sarahMe@qimr.edu.au

Previous genetic studies of finger ridge count have typically found that ridge count is influenced both by a common genetic factor (that influences the magnitude of ridge count on all ten fingers) and by a series of independent factors or developmental fields that take into account the correlation in ridge count which is highest between adjacent fingers. We performed a 5-10cM multivariate genome scan (400-761 markers) on the ridge count of individual fingers in a sample of 486 twin families. In the multivariate genome scan we estimated 5 QTL effects (one for each of the five digits, with the QTL estimates of corresponding fingers on the left and right hands constrained to be equal). The results of these analyses were compared to those of a univariate QTL analysis for Total Finger Ridge Count (a summary measure in which the ridge counts from all fingers are weighted equally). We found that the univariate analysis performed better than the multivariate analysis when the QTL effect loaded equally across all fingers (as seen on 1q43). However, the multivariate analysis conferred greater statistical power when the estimates of the QTL effects behaved in a manner consistent with a developmental field (ie as seen on 3q25 and 7p15). Our findings provide evidence for the suggestion that both common genetic factors and developmental fields may influence the development of dermatoglyphic characteristics.

Jane Mendle¹, Eric Turkheimer¹, Robert E. Emery¹, Brian M. D=Onofrio¹, Stacy K. Lynch¹, Wendy S. Slutske², Andrew C. Heath³, & Nicholas G. Martin⁴. Association of Early Menarche with Adolescent Depression⁴.

¹Department of Psychology, University of Virginia, Charlottesville VA USA, ²Department of Psychology, University of Missouri B Columbia, Columbia MO USA, ³School of Medicine, Washington University, St. Louis MO USA, ⁴Queensland Institute of Medical Research Australia, ⁵Supported by William T. Grant Foundation, Grant # 2054.

Address : Dept. of Psychology, Box 400400, University Of Virginia Charlottesville, VA 22903 Telephone: (434) 982-5573 E-mail: jm4ky@virginia.edu

Early onset of menarche is associated with increased risk for depression, eating disorders, and substance abuse during adolescence. It has been posited that the emotional and psychological stress related to non-normative pubertal timing plays a causal role in the etiology of these disorders. Since cognitive and physical development do not necessarily occur in synchrony, precocious physical maturation may increase the likelihood that girls will be forced to confront new environments, stressors, and social expectations before they are emotionally prepared to do so. It may also be that those girls who exhibit obvious signs of physical maturation may find it difficult to maintain friendships with same-sex peers who have not developed at a similar rate.Given that menarche and pubertal timing are strongly influenced by familial factors, the association between early menarche and adolescent psychopathology may be muddied by genetic and environmental confounds. The primary purpose of this study is to clarify whether early menarche wields a pure, causal effect on the development of depression during adolescence. Using data from the Australian National Health and Medical Research Council Twin Registry, we will conduct a cotwin control analysis of age of menarche and development of depressive symptomatology utilizing a sample of twins comparatively discordant for menarcheal age. If early pubertal timing has a causal influence on depressive etiology, the earlier maturing twin will be more likely to develop depression than the later maturing one. However, if some confound associated with familial propensity to develop depression and pubertal timing exists, there will be no difference in the emergence of adolescent depression between the twins. Additional analyses will model the inclusion of potential mediators of depression into hierarchical linear models.

Christel M. Middeldorp^1,2, Danielle C. Cath², A. Leo Beem¹, Dorret I. Boomsma¹. Genetic epidemiology of depression in a selected population of Dutch twins and their siblings.

¹Department of Biological Psychology, Vrije Universiteit Amsterdam, Amsterdam, The Netherlands, ²Department of Psychiatry, GGZ Buitenamstel / Vrije Universiteit Amsterdam, Amsterdam, The Netherlands, ³Supported by ZonMW Grant 940-37-024.

Address: Vrije Universiteit Amsterdam, Dep Biological Psychology, Van der Boechorststraat 1, 1081 BT, Amsterdam, The Netherlands Telephone: +31-20-4448787 Fax: +31-20-4448832 Email: cm.middeldorp@psy.vu.nl

A multivariate genetic analysis on the anxiety, depression, neuroticism and somatic-anxiety data collected in twins and their siblings of the Netherlands Twin Register (NTR) revealed that genetic covariances for these traits could be fully attributed to a common genetic factor (D.I. Boomsma, A.L. Beem, M. van den Berg, C.V. Dolan, J.R. Koopmans, J.M. Vink, E.J.C. de Geus and P.E. Slagboom, 2000, Twin Research 3, 323-334). The value of this common genetic factor can be estimated for each individual using the individual scores on the traits and the factor loadings on the common genetic factor. Based on these genetic factor scores (GFS), available for 7770 twins and their siblings from 3323 families, a selection was performed to obtain a sub-sample that is most informative for linkage studies. Of the selected subjects, 1256 twins and their siblings from 478 families also participated in a Telephone interview during which the Composite International Diagnostic Interview (CIDI) (World Health Organization, 1997, CIDI-Auto, Version 2.1: Administrator=s guide. World Health Organization, Sydney, Australia) was administered. We will present the genetic epidemiological analysis of the liability for major depressive disorder in an extended twin design. Because the sample was selected based on the GFS, bivariate genetic analyses will be carried out for major depression, our variable of interest and the GFS, the variable used for the selection (R.J.A. Little and D.B. Rubin, 1987, Statistical analysis with missing data. Wiley, New York, USA).

Michael Miller. Linkage Conditional on Measured Genotypes Under Conditions of Low Statistical Power

Division of Epidemiology, University of Minnesota, USA. Email: mbmiller@taxa.epi.umn.edu

Almasy and Blangero (Behavior Genetics, March 2004) demonstrated that when there is a single QTL in a chromosomal region and a significant LOD-score peak (LOD > 3.0) is observed in that region, use of the QTL as a covariate in the linkage analysis usually reduces the peak to a non-significant level (LOD < 0.588) at the QTL. Thus, one may test a candidate gene as a covariate in quantitative linkage analysis to see if it accounts for an observed peak. The simulation undertaken by Almasy and Blangero was highly powered with a QTL accounting for 28% of the trait variance, a mean LOD-score of 5.57 and 87% of replicates achieving LOD > 3.0. It is more realistic to simulate data with a QTL that accounts for much less variance (maybe 10%) and produces much lower statistical power. When we analyze sibling-pair data simulated under such a model, we find that using the QTL as a covariate will rarely reduce a statistically significant peak (LOD > 3.0) to nonsignificance (LOD < 0.588). We also show through simulation that this finding is due to the fact that much of the contribution to the LOD-score peak comes from lucky associations of extra-QTL genetic effects and random environmental influences with identity-by-descent patterns. Such effects remain when the QTL is used as a covariate. Our findings imply that the method recommended by Almasy and Blangero works differently, and is less definitive, when power is low to detect the effect of the QTL. Alternative methods and interpretations are recommended.

Maria Napolitano, Thalia C. Eley¹. Parental disciplinary styles and associations with anxiety in adolescent and young adults².

¹ MRC, Social, Genetic and Developmental Psychiatry Research Centre, Institute of Psychiatry, London UK, ² Supported by an MRC career development award to the second author.

Address: MRC, SGDP, PO80, Institute of Psychiatry, De Crespigny Park, London, SE5 8AF Telephone: 44 (0)207 848 0979 Email: m.napolitano@iop.kcl.ac.uk

Minimal attention has been paid to the association between aspects of anxiety and parenting and, to our knowledge, this has not been examined within a behavioural genetic design. This study considers the importance of parental disciplinary styles as risk factors for adolescent anxiety symptoms. Perceptions of maternal and paternal punitive and constructive discipline (M. Hetherington & Clingempeel, 1992, Monographs of the society for research in Child Development, 57, 2-3) were assessed from over 1,300 pairs of adolescents and young adults twins and siblings (age range 13-21 years). In addition the young people completed 6 anxiety scales: panic, separation anxiety, social anxiety, physical injury anxiety, ocd, and general anxiety. (S. Spence, 1990, Multivariate Behavioral Research, 25, 175-180). The results indicated a significant interaction between age and sex on the anxiety scale; for females, increasing age predicted higher anxiety scores. Univariate genetic analyses indicated that all the anxiety scales were highly heritable (e.g. panic h² = .48), with indications of non-additive genetic influence. In contrast, the parental discipline variables were less heritable and also influenced by the shared environment (e.g. paternal constructive discipline h² = .28 and c² = .21). All anxiety and parental discipline variables varied as a function of sex, and most by age. Associations between anxiety and perceptions of parental discipline were higher for adolescent females (aged 11-15) than adolescent males or young adults (aged 16-21). In the adolescent females, associations between perceived maternal punitive discipline and general anxiety, panic, social anxiety and OCD were highest (range r = .20-.26). Bivariate genetic analyses on the adolescent females indicated that the associations between perceptions of maternal punitive discipline and anxiety sub-types were largely genetically influenced. These results suggest that links between anxiety vary as a function of age and sex during adolescence and young adulthood.

Michael C. Neale¹. Extensions of models for genetic linkage: mixture distributions and epistasis.

¹ Virginia Institute for Psychiatric & Behavioral Genetics, Department of Psychiatry, Virginia Commonwealth University, Richmond VA 23298 USA, ²Supported by NIH grants RR-08123, MH-01458 and MH-65322.

Address: VIPBG, MCV/VCU, P.O. Box 980126, Richmond VA 23298-0126 USA Telephone: 804/828-3369 Fax: 804/828-1471 Email: neale@hsc.vcu.edu Web: http://www.vipbg.vcu.edu

Epistatic interactions are thought to be important in the etiology of complex traits, especially for characteristics subject to selection pressure. The detection of epistasis in linkage studies is relatively straightforward in principle, but suffers from lack of statistical power. Methods for detecting epistasis in linkage studies using Mx are presented, and the statistical properties of likelihood-ratio, bootstrap and permutation tests for it are described.

Jenae M. Neiderhiser¹. Identifying Sources of Nongenetic Influences on Change in Parent-Child Relationships from Adolescence to Young Adulthood².

¹Center for Family Research, Department of Psychiatry and Behavioral Sciences, George Washington University, Washington, DC USA, ²Supported by NIH grant MH59014.

Address : 2300 K St., N.W., 3^rd Floor Warwick Building, Washington, DC 20037 Telephone: (202) 994-2212 Fax: (202) 994-4812 Email: cfrjmn@gwumc.edu

One of the first studies to examine genetic and environmental influences on stability and change in parent-child relationships from adolescence to young adulthood found that there was only a moderate degree of stability (.28 to .39). In most cases, this stability was due to shared environmental influences. The bulk of the change in parent-child relationships through to young adulthood was also due to environmental factors, although in this case, nonshared environmental. The current study will attempt to identify the sources of these nongenetic influences on stability and change in parent-child relationships. Subjects from the Nonshared Environment in Adolescent Development (NEAD) project will be used for the current report. NEAD included 720 families consisting of two parents married at least 5 years in nondivorced and step families with two siblings (< 4 year age difference). Average child age at Time 1 was 13.6 (+3) years and 15.0 (+2) years at Time 2. There were five types of sibling pairs: MZ and DZ twins, full, half and step siblings. Measures included in these analyses index parent-child relationships as rated by mothers, fathers, adolescent self-reports and observer ratings for each parent=s behavior towards each adolescent. A follow-up of the NEAD sample is just being completed. The now young adult participants range in age from 23 to 31 years. Measures for this study include young adult and parent reports of current adult child-parent relationship. Potential sources of environmental influences that will be examined will include parental mental health, adolescent peer and sibling relationships and current status of romantic relationship and children in the home that differ for the twin and sibling pairs.

Ragnhild B. Nes, E. Røysamb, K. Tambs & J. R. Harris. Sleep problems and well-being: Genetic and environmental contributions. Division of Epidemiology, The Norwegian Institute of Public Health, Oslo, Norway

Address: The Norwegian Institute of Public Health, Division for Epidemiology, PO Box 4404 Nydalen, 0403 Oslo Norway Telephone: 47-22042200 Fax: 47-23408247 Email: ragnhild.bang.nes@fhi.no

Sleep problems are common, affecting health and well-being, risk of accidents, and morbidity. This study estimates the genetic and environmental contributions to subjectively reported sleep problems, and investigates the relationship between sleep disturbance and subjective well-being (SWB) in a cohort of Norwegian twins born 1967 to 1979. Self report questionnaire data on sleep problems, health and well-being collected in 1998 from 8,045 twins were analysed using structural equation modelling to investigate the relative magnitudes of genetic and environmental influences on phenotypic variance. Furthermore, sex-specific effects were explored. Preliminary analyses indicate substantial genetic contributions to the observed variance in sleep problems. Polychoric correlations were considerably higher in MZ than in DZ twins for both males (.51 versus .04) and females (.42 versus .15). A correlation of .15 was observed in the DZ unlike sex group. A model specifying additive and non-additive genetic effects and individual environmental effects fit the data best. The phenotypic correlation between sleep disturbance and SWB was estimated to .40. Cross-twin cross-trait correlation correlations for each of the five zygosity by sex groups suggest that genetic factors may contribute differentially to the association between sleep disturbance and SWB in males and females. Further analyses of the relationship between reported sleep problems and SWB will be presented, and results pertaining to sex differences discussed. Conclusions: Genetic factors contribute substantially to the variation in reported sleep problems. Moreover, cross-twin cross-trait correlations indicate sex-specific genetic effects for sleep problems and SWB

Lars-Göran Nilsson^1,7, Rolf Adolfsson², Lars Bäckman³, Marc Cruts⁴, Lars Nyberg⁵, Brent J. Small⁶, and Christine Van Broeckoven⁴. Impairment of Episodic Memory and Word Fluency in Carriers of the APOEå4 Allele.

¹Department of Psychology, Stockholm University, Sweden, ²Department of Clinical Sciences, Psychiatry, Umeå University, Sweden, ³Aging Research Center, Karolinska Institute, Stockholm, Sweden, ⁴Department of Molecular Genetics, Flanders Interuniversity Institute for Biotechnology, University of Antwerp, Belgium, ⁵Department of Psychology, Umeå University, Sweden, ⁶School of Aging Studies, University of South Florida, Tampa, USA, ⁷Center for Advanced Study, Norwegian Academy of of Science and Letters, Oslo, Norway.

Address : Department of Psychology, Stockholm University, 10691 Stockholm, Sweden Telephone: +46 8 163940 Fax: +46 8 159342 Email: lgn@psychology.su.se

We examined the effect of the å4 allele of ApolipoproteinE (APOE) in non-demented individuals in the age range of 35-90 years of age. In a prospective cohort study we demonstrated a more pronounced å4-related deficit for participants in the age of 65 years and older in tasks assessing episodic recall. Somewhat smaller å4-related deficits for these persons were found in episodic recognition and verbal fluency. No deficits were found in tasks assessing vocabulary, primary memory and priming. We also demonstrated that carriers of the å4 allele in the age of 50-60 years performed at a higher level than non-carriers of this allele in episodic recall tasks, but not in episodic recognition tasks, nor in tasks assessing semantic memory, primary memory and priming. Furthermore, we found a dose effect, such that carriers of two å4 alleles fail more profoundly in acquiring and recollecting episodic information than carriers of one å4 allele, who in turn fail more than carriers of non-å4 alleles. The pattern of findings observed for older å4 carriers suggests that these individuals have particular difficulty when the executive task demands are high. Several factors (e.g., smaller hippocampal volumes, less effective neural repair mechanisms, preclinical dementia) may account for these findings. On the basis of the data obtained we argue that analyses of the effect of specific genes in cognition should be accompanied by assessment of performance at a specific level, with due attention to the age of the individual.

Syuichi Ooki. Genetic and environmental influences on some habitual behaviors in childhood.

Department of Health Science, Ishikawa Prefectural Nursing University, Kahoku, Ishikawa, Japan.

Address : Department of Health Science, Ishikawa Prefectural Nursing University, Tsu7-1 Nakanuma, Takamatsu, Kahoku, Ishikawa, 929-1212, Japan

Telephone: 81 76 281 8377 Fax: 81 76 281 8377 Email: sooki@kj8.so-net.ne.jp

The author conducted this study to estimate the genetic and environmental contribution to seven habitual behaviors in childhood, namely sleep talking, half-sleeping, night terrors, nocturnal enuresis, nail biting, stuttering, and tics. The subjects were 2029 pairs of twins, consisting of two groups. The first group of subjects consisted of 937 twin pairs of whose mothers belong to the several maternal associations for the parents of multiples. The second group of subjects consisted of 1092 twin pairs of the applicants to the secondary education school attached to the faculty of education of the University of Tokyo. The twins' mothers had completed a Addressed or handed questionnaire. With regard to the above mentioned seven traits, they selected one answer from the choices 'often', 'sometimes' 'never' and 'unknown' for each twin. Genetic analysis was performed as follows. First the answers were summarized in the form of two patterns of 2 x 2 contingency table; 'often' and 'sometimes' or >sometimes= and >never= were included in one category. First, proband-wise concordance rates were calculated according to zygosity and sex combination. Then, tetrachoric correlations of the contingency table were calculated using the program package PRELIS2. Furthermore, covariance structure analysis was performed for several genetic models using the program package LISREL8. The results were as follows. Univariate genetic analysis showed that all traits were under genetic control of varying level. Sleep talking, half-sleeping and night terrors were under strong or moderate genetic control. As to nocturnal enuresis, the genetic effect was moderate and shared environmental factors played an important role. Moreover several traits tended to occur together, and sleep talking, half-sleeping and night terrors shared common genetic and environmental factors in addition to specific genetic and environmental factors.

Maarten W. Peeters¹, Martine A. Thomis¹, Hermine H.M. Maes², Ruth J.F. Loos³, Albrecht L. Claessens¹, Robert Vlietinck^4,5, Gaston P. Beunen¹. Causes of stability in explosive strength during adolescence.

¹ Faculty of Physical Education and Physiotherapy, Katholieke Universiteit Leuven, Belgium, 5 Virginia Institute for Psychiatric and Behavioral Genetics, Department of Human Genetics, Virginia Commonwealth University, Richmond, VA, USA, ³ Pennington Biomedical Research Center, Human Genomics Lab, Baton Rouge, LA, USA, ⁴ Centre for Human Genetics, Faculty of Medicine, Katholieke Universiteit Leuven, Belgium, ⁵ Division of Genetics and Molecular Cell Biology, Maastricht University, The Netherlands.

Address : Faculteit Lichamelijke Opvoeding en Kinesitherapie Tervuursevest 101 B-3001 Leuven (Belgium) Telephone: +32 16329086 Fax: +32 16329197 Email: Maarten.Peeters@flok.kuleuven.ac.be

Tracking or stability of vertical jump (VTJ), a measure of explosive strength, has been shown to be low to moderately high during adolescence (R.M.Malina, RQES, 1996, 67 (suppl), 48-57) though some temporary instability is observed due to differences in tempo and timing of the adolescent growth spurt. Previous twin studies have reported heritability estimates ranging between .45 and .93. The aim of the present study was to link these two findings and try to determine whether the stability of the trait is mainly caused by genes and/or environment. Subjects are from the Leuven Longitudinal Twin Study (LLTS) (n=105 pairs, equally divided over five zygosity groups). VTJ data was aligned on age at peak height velocity (APHV). Simplex models (D.I. Boomsma and P.C.M. Molenaar, Behav. Genet., 1987, 17, 111-123) were fitted to the data in a five group analysis. Non Scalar (NS), Specific Scalar (SS), General Scalar (GS) and models without sex-differences were tested. Results show that after aligning the data on APHV, the annual autocorrelations show a clear simplex structure over a 4 year interval. The most parsimonious model (AIC) was the GS-AE model. Heritability estimates ranged between 72.2% (95%CI: 58.0-82.7%) and 82.3% (95%CI: 70.8-92.3%) which is in agreement with values found in the literature. Additive genetic >innovation= variance was 81.3% (95%CI: 70.1-88.4%) for the first measurement occasion, and trough the transmission paths explained more than 50% of the total variance at all subsequent observations, with genetic innovation after the first observation never explaining more than 9.1%. Unique environmental innovation at the first measurement occasion, which was 18.7% (95%CI=11.6-29.9%) on the other hand explained less than 4% of the variation at subsequent time points. It thus can be concluded that the observed stability of explosive strength during adolescence is mainly caused by a stable genetic influence in boys and girls.

Nicole Philip, Sabine Sigaudy, Anne Moncla. Microdeletion syndromes as a model for identifying genes involved in behavioral development in humans. The example of 22q11.3 microdeletion and Smith-Magenis syndromes.

Department of Medical Genetics, Timone Children=s Hospital, Marseille, France

Address: Département de génétique médicale, Hôpital d=Enfants de la Timone, 13385, Marseille Cedex 5, France Telephone : 04 91 38 66 30. Email : nicole.philip@ap-hm.fr

Microdeletions and microduplications, not visible by routine chromosome analysis are a major cause of congenital malformations and mental retardation in human. Most of these rerarrangements were described during the last 10 years, as the result of improved cytogenetic techniques. In most cases, the deletion ranges from 2 to 3Mb and include up to 30 different genes. A microdeletion syndrome is characterized by specific but rather subtle clinical manifestations that make the diagnosis easy to trained clinicians. The presence of characteristic behavioral traits is another hallmark of many of these syndromes. We will focus our presentation on two examples and present the experience of the department of medical genetics of Marseille. Microdeletion of chromosome 22q11.3 is a rather frequent anomaly, affecting one child out of 4000. The main clinical manifestations are a subtle facial dysmorphism, congenital heart defects in 2/3, hypocalcemia, velopharyngeal insufficiency and learning disabilities. Psychiatric problems, especially schizophrenia occur in approximately 15% of patients. According to the wide clinical variability among patients, it is hard to define a single pattern of behavior in 22q11 microdeletion syndromes. Conversely, patients with a Smith-Magenis syndrome due to a 17p12.2 microdeletion display a peculiar behavioral phenotype which is characteristic enough to be a major diagnostic clue. Self-injurious behavior and prominent sleep disturbances are seen in the majority of patients, but appear to be age-related.

Alison Pike¹ and Thalia C. Eley². Relations among Parenting and Extra-familial Relationships: Nature or Nurture?

¹Department of Psychology, University of Sussex, UK, ²Social, Genetic, Developmental Psychiatry Centre, Institute of Psychiatry, King=s College London, UK, ³ The Surrey Adolescent Twin Study was supported by a grant from the University of Surrey Pro-Vice Chancellor=s Fund to the first author. The G1219 study was supported by the W T Grant Foundation and by a Medical Research Council training fellowship to the second author.

Address : Psychology Department, University of Sussex, Brighton, BN1 9QH, UK Phone: 44 1273 877288 Fax: 44 1273 678058 Email: alisonp@sussex.ac.uk

By the time children enter into adolescence, time spent with peers is more than double that spent with parents (R. W. Larson, M. H. Richards, G. Moneta, G. Holmbeck, and E. Duckett, 1996, Dev. Psych. 32, 744-754). However, both cross-sectional and longitudinal associations between parenting and peer group characteristics, as well as friendship quality, suggest that parents continue to exert influence (e.g., M. Gold, and D. S. Yanof, 1985, J of Pers. And Soc. Psych. 49, 654-659). The current study examined associations between parenting and both peer group characteristics and friendship quality within two twin samples. The Surrey Adolescent Twin Study consists of 200 twin pairs aged 12-14 years, and includes parent and child reports of the parent-child relationships, and adolescent reports of peer group characteristics and friendship quality. G1219 consists of over 1100 twin and sibling pairs aged 12-20 years, and includes adolescent reports of the relationship constructs. Phenotypic analysis revealed that parenting was moderately associated with peer group characteristics and friendship quality for the younger twins (aged 12 B 15), but correlations among the constructs was negligible for the older adolescents in G1219. Univariate analyses indicated that both the parent and peer relationship measures are moderately genetically influenced, and that nonshared environmental factors account for more of the variance than do shared environmental factors when adolescent reports are utilized. Replicating previous research, parent reports of their own parenting indicated similar treatment of their children (shared environment). Bivariate model-fitting analyses from G1219 indicate that the covariance between parenting and peer relationships is largely due to common genetic influence. If replicated using the Surrey Adolescent Twin Study, such findings suggest that individual (personality?) traits of children may explain much of the consistency seen across relationships, and that parenting per se may not play a vital role in shaping children=s peer relations.

Robert Plomin¹. Pooled DNA on chips: Genotyping large samples of cases and controls for thousands of SNPs simultaneously.

¹Social Genetic and Developmental Psychiatry Centre, Institute of Psychiatry, London, UK. Supported by UK Medical Research Council Program Grant G9424799.

Address :

Social, Genetic and Developmental Psychiatry Centre, Institute of Psychiatry, King=s College London, Box P080, London, UK SE5 8AF

Telephone: 44 207 848-0893 Fax: 44 207 848-0895 E-mail: rplomin@iop.kcl.ac.uk

Heritability is ubiquitous for behavioral dimensions and disorders, which means that DNA differences exist that cause these differences among individuals. However, progress has been slow towards reliably identifying these DNA differences. A major reason for this slow progress is inadequate power to detect small effect sizes. Power to detect small effects sizes (<1%) requires allelic association rather than linkage. Rather than examining just a few genes, allelic association approaches need to be systematic as linkage studies are. One way to conduct systematic genome scans with association is to use microarrays (chips) that simultaneously genotype thousands of single-nucleotide polymorphisms (SNPs). However, the use of chips seems incompatible with large samples because each chip costs several hundred dollars and can be used only once. The resolution is to use pooled DNA on chips. Since 1996, we have been using pooled DNA to genotype groups of cases and controls in which a small amount of DNA from each individual in a group is pooled. For example, 500 SNPs can be screened for the cost of genotyping one SNP for 500 cases and 500 controls. We are now able to use pooled DNA on chips (Affymetrix). We have shown that genotyping pooled DNA on chips is highly reliable (comparing replicate pools), valid (comparing pooled results to individual genotyping results) and sensitive to small allelic frequency differences between cases and controls (using spiked pools that systematically vary allelic frequency differences). We are currently using pooled DNA on chips to conduct allelic association genome screens for 10,000 SNPs in genes (GeneChip7 Mapping 10K Array Xba 131) for 500 cases and 1000 controls selected from a community sample of 15,000 UK twins. This approach is being applied to several disorders including reading disability, language disability, mathematics disability, mild mental retardation, autistic spectrum disorder, and psychopathy.

Michael F. Pogue-Geile¹, Raquel E. Gur², Ruben C. Gur², Laura Almasy³, Joel Wood⁴, Judy Thompson¹, Sarah Tarbox¹, and Vishwajit Nimgaonkar⁴. Executive functioning and genetic liability to schizophrenia: Preliminary results from a large-pedigree, multiplex study⁵.

¹Department of Psychology, University of Pittsburgh USA, ²Department of Psychiatry, University of Pennsylvania USA, ³Southwest Foundation for Biomedical Research, ⁴Department of Psychiatry, University of Pittsburgh, ⁵Supported by NIH MH63480

Address : Department of Psychology, University of Pittsburgh, Pittsburgh, PA 15260 USA, Telephone: 412.624.8818; Email: mfpg@pitt.edu

The overall importance of genetic variation on the etiology of schizophrenia has been clearly documented for some time. However, reliably identifying specific genetic polymorphisms that contribute to this overall risk has been very difficult, presumably because there are many contributing loci, each with small effect. One strategy that may aid in such situations is to develop phenotypes that are more sensitive to genetic liabilities than is the clinical diagnosis itself (i.e., endophenotypes, Gottesman & Shields, 1972). The current study employs this approach using neuropsychological phenotypes within the context of an ongoing linkage study of large pedigrees, multiply affected with schizophrenia. Here we report preliminary results on the familial association between the Trails Making Test (TMT), a neuropsychological test of executive function, and schizophrenia in a subsample of this ongoing study. The TMT is a timed Aconnect the dot@ test that involves visual search and working memory. It has two subtests, TMT-A and TMT-B, with TMT-B having an additional memory load. If the TMT is sensitive to genetic liability to schizophrenia, then performance should decrease as relatives= genetic relatedness to schizophrenia probands increases. Large families were ascertained that had at least two relatives diagnosed with schizophrenia or schizo-affective disorder. One hundred twelve members from nine families were assessed with the TMT and diagnosed using the Diagnostic Interview for Genetic Studies (DIGS). Eighteen probands were diagnosed with schizophrenia, leaving 94 of their non-schizophrenia relatives for analysis. Each relative was assigned a kinship coefficient based on their degree of genetic relatedness to a schizophrenia proband (e.g., .50 for a sibling, .25 for a nephew, etc.). Using multiple regression with age as a covariate, there were significant linear associations between the time to complete TMT-A and TMT-B and genetic relatedness to schizophrenia probands (b=.34, p=.001 and b=.27, p=.005). As the genetic relationship to a schizophrenia proband increased, the time to complete TMT-A and BB also increased, in a linear fashion. Such results suggest that the TMT may be sensitive to genetic liability to schizophrenia and thus may be useful in linkage studies.

Danielle Posthuma¹, Michelle Luciano², Eco JC de Geus¹, Margie J Wright², Nick G Martin², Dorret I Boomsma¹. Full genome scans for cognition in Dutch and Australian samples.

¹Vrije Universiteit, Amsterdam, The Netherlands; ²Queensland Institute of Medical Research, Brisbane, Australia. Supported by GenomEUtwin, European Union Contract No. QLG2-CT-2002-01254, the Human Frontiers of Science Program (grant number rg0154/1998-B), the Australian Research Council and the Australian NHMRC=s Program in Medical Genomics [NHMRC-219178]. Genotyping was carried out by the Center for Medical Genetics in Marshfield.

Address: Biological Psychology, Vrije Universiteit, Van der Boechorststraat 1,

1081 BT Amsterdam, The Netherlands, fax 31-20-4448832, Telephone 31-20-4448814, Email: danielle@psy.vu.nl

Psychometric IQ was assessed in Dutch and Australian extended twin families. The Dutch sample consisted of 793 subjects from 317 families, aged between 17 and 68 years. The Australian sample consisted of 1339 subjects from 603 families, aged between 15 and 22 years. Heritability estimates of psychometric IQ ranged between .80 and .90 in both samples. On subsamples of the Dutch and Australian samples genotypic marker data has been collected. Preliminary independent scans in the Australian sample (762 individuals from 345 families) and Dutch samples (102 individuals from 37 families) suggested linkage for at least one similar region (lodscores Australian sample > 3, Dutch sample > 1.5) . Simultaneous scans will be conducted to determine whether these linkage peaks are actual replications in both samples.

Carol A. Prescott¹, Jonathan S. Kuhn¹ and Nancy L. Pedersen² . French Paradox Redux: Are the cardio-protective effects of moderate alcohol consumption direct or indirect? ³

¹ Virginia Institute for Psychiatric and Behavioral Genetics, Virginia Commonwealth University, Richmond VA USA, ² Department of Medical Epidemiology, Karolinska Institutet, Stockholm, Sweden, ³ Supported by the Swedish Scientific Council, the Swedish Ministry of Higher Education, and a grant from the U.S. National Institutes of Health (AG08724).

Address : P.O. Box 980126, Virginia Commonwealth University, Richmond VA 23298 USA Telephone: (804) 828-5968 Fax: (804) 828-1471 Email: cprescott@vcu.edu

Moderate drinking has been associated with decreased risk for atherosclerotic cardiovascular disease (CVD). This effect has been called the AFrench Paradox@ because of the lower CVD rates in some regions of France despite a diet high in saturated fat. Several physiological mechanisms for the protective effects of alcohol have been proposed, but these do not appear to account completely for the effects observed. An alternative explanation is that the association is indirect: moderate drinkers, particularly wine drinkers, engage in health and lifestyle behaviors which are responsible for the reduced risk for CVD. We addressed these alternatives using a 30-year follow-up study of participants from the Swedish Twin Registry born between 1882 and 1958. Over 29,000 twins who provided data on alcohol consumption and other health and lifestyle factors in the 1960=s and 1970=s were matched to national hospitalization and cause of death records to identify morbidity and mortality due to CVD. The sample includes more than 9,000 deceased individuals with known cause of death (including >4,500 due to CVD), and over 8,900 individuals with hospital admissions associated with CVD. We examined the association between level of alcohol consumption and risk for CVD using a cotwin-control design (comparing CVD outcomes in pairs discordant for level of alcohol consumption) and a discordant pair design (comparing alcohol consumption in pairs discordant for CVD outcomes).

J. Drew Prosser¹ and Y.-K. Kim¹. Kin recognition in Drosophila paulistorum: relatedness vs familiarity².

¹Department of Genetics, University of Georgia, Athens GA USA, ²Partly supported by CURO Summer Fellowship at UGA

Address : Department of Genetics, University of Georgia, Athens, GA 30602

Telephone: 706 542 1448, Fax: 706 543910, Email:yongkyu@uga.edu

Kin recognition serves as an inbreeding avoidance mechanism, which is a key to maintaining the overall fitness of a species (W. D. Hamilton, 1964, J. Theor. Biol. 7, 1-52; P. Bateson, 1983, in P. Bateson, ed., Mate Choice, pp. 257-277, Cambridge University, Cambridge). Kim (in press) has demonstrated that Drosophila paulistorum, when given a choice, prefer to mate with genetically non-related individuals rather than with siblings. Further, when siblings or non-siblings were raised together, they had reduced sexual activities and consequently avoided mating. In an attempt to further understand kin recognition in D. paulistorum, we have investigated 1) whether they discriminate between different degrees of relatedness; and 2) whether they use genetically-determined cues such as recognition alleles. In female choice situations, one female was placed in mating chambers with a male sibling and a male half-sibling (HS tests); one female with a male sibling and a male cousin (C tests); and one female with a male sibling and a male non-sibling (NS tests), respectively, without prior experience of either male during development. Current data showed that there were no significant differences in mate choice between siblings and individuals of different degrees of relatedness (p=0.1573 for HS; p=0.3870 for C; p=0.2059 for NS). These results suggest that 1) Drosophila kin recognition is based solely on familiarity acquired during developmental experience rather than genetic relatedness; 2) they treat the A<