- Email: [email protected]
GABRA2 Association With Addiction-Related Endophenotypes Is Environmentally Influenced
S410
Abstracts of the XXIV World Congress of Psychiatric Genetics (WCPG) 2016
M61. SHARED GENETIC VARIANTS BETWEEN SCHIZOPHRENIA AND GENERAL COGNITIVE FUNCTION INDICATE COMMON MOLECULAR GENETIC MECHANISMS
Olav B Smeland1, Karolina Kauppi2,3,4, Yunpeng Wang1,2,3, W David Hill5,6, Gail Davies5,6, Oleksandr Frei1, Wen Li1, Jon A Eriksen1, Aree Witoelar1, Francesco Bettella1, Chun C Fan2,7, Wes Thompson8, Schizophrenia Working Group of the Psychiatric Genomics Consortium, neuroCHARGE Cognitive Working Group1,2,3,4,5,6,7,8,9,10,11, Chi-Hua Chen2,3, Srdjan Djurovic9,10, Ian J Deary5,6, Anders M Dale2,3,8,11, Ole A Andreassen1,8 1
NORMENT, KG Jebsen Centre for Psychosis Research, Institute of Clinical Medicine, University of Oslo and Division of Mental Health and Addiction, Oslo University Hospital, Oslo, Norway 2 Department of Radiology, University of California, San Diego, La Jolla, CA, United States of America 3 Multimodal Imaging Laboratory, University of California San Diego, La Jolla, CA, United States of America 4 Department of Radiation Sciences, Umeå University, SE-901 87 Umeå Sweden 5 Centre for Cognitive Ageing and Cognitive Epidemiology, University of Edinburgh, Edinburgh, UK 6 Department of Psychology, University of Edinburgh, Edinburgh, UK 7 Department of Cognitive Science, University of California, San Diego, La Jolla, CA, USA 8 Department of Psychiatry, University of California, San Diego, La Jolla, CA, USA 9 Department of Medical Genetics, Oslo University Hospital and University of Oslo, Oslo, Norway 10 NORMENT, KG Jebsen Centre for Psychosis Research, Department of Clinical Science, University of Bergen, Bergen, Norway 11 Department of Neurosciences, University of California San Diego, La Jolla, CA 92093, United States of America Background: Schizophrenia (SCZ) is a severe mental disorder characterized by widespread cognitive impairments including deficits in learning, memory, processing speed, attention and executive functioning. Although cognitive deficits are a strong predictor of functional outcome in SCZ, current treatment strategies largely fail to ameliorate these impairments. Thus, in order to develop more efficient treatment strategies in SCZ, a better understanding of the pathogenesis of these cognitive deficits is needed. Given that both SCZ and cognitive ability are substantially heritable, we here aimed to determine whether SCZ share genetic influences with general cognitive function (COG), a phenotype that captures the shared variation in performance across several cognitive domains.
Methods: We analyzed GWAS results in the form of summary statistics (p-values and z-scores) from SCZ (the Psychiatric Genomics Consortium; n = 82 315) and COG (CHARGE Consortium; n= 53 949). We applied a conditional false discovery rate (FDR) framework. By leveraging SNPassociations in a secondary trait (SCZ or COG), the conditional FDR approach increases power to detect loci in the primary trait (COG or SCZ), regardless of the directions of allelic effects of the risk loci. We then applied the conjunction FDR to identify shared loci between the phenotypes. The conjunction FDR is defined as the maximum of the conditional FDRs for both directions, and we used an overall FDR threshold of 0.05. Results: To visualize pleiotropic enrichment, we constructed conditional Q-Q plots which indicate substantial polygenetic overlap between SCZ and COG. For progressively stringent p-value thresholds for SCZ SNPs, we found approximately 150-fold enrichment for COG. For progressively stringent p-value thresholds for COG SNPs, we found approximately 100-fold enrichment for SCZ. We then used the conjunction FDR and identified fourteen independent loci shared between SCZ and COG. The majority of the shared loci show inverse associations in SCZ and COG, in line with the observed cognitive dysfunction in SCZ. Discussion: Our preliminary findings indicate shared molecular genetic mechanisms between SCZ and COG, which may provide important new insights into the pathogenesis of cognitive dysfunction in SCZ.
Disclosure Nothing to Disclose. http://dx.doi.org/10.1016/j.euroneuro.2016.09.452
M62. GABRA2 ASSOCIATION WITH ADDICTION-RELATED ENDOPHENOTYPES IS ENVIRONMENTALLY INFLUENCED
Margit Burmeister, Elisa Trucco, Shweta Ramdas, Stephen Parker, Mary Heitzeg, Sandra Villafuerte, Robert Zucker University of Michigan Background: The GABRA2 gene contains more than 250 SNPs that in Caucasians form two major common “Yin-Yang” haplotypes - having either all or none of the non-ancestral (NA) alleles. Since 2004, association of the non-ancestral, slightly less common haplotype with alcohol use disorder (AUD) and with increased beta waves in EEG has been reported and replicated many times. Methods: In a longitudinal sample of 4400 low SES families at high risk for AUD, we investigated the influence of these GABRA2 haplotypes on addiction related endophenotypes, and started to investigate potential functional implications. We found that the NA haplotype is associated with impulsivity and with increased activation of the insula in response to reward expectation, both of which partially mediate the
Abstracts of the XXIV World Congress of Psychiatric Genetics (WCPG) 2016 association of the NA alleles with alcoholism. Impulsivity was also associated with reward expectation, independent of genetics. Results: Moreover, the effect of parental monitoring on externalizing behavior trajectories across childhood and adolescence, i.e., consistently low, developmentally limited, and rising trajectories, was moderated by the NA alleles of GABRA2. Subjects with the NA alleles were more strongly influenced – both positively and negatively–by the extent of parental monitoring, while subjects with the ancestral alleles were not significantly influenced by the monitoring. In addition to parents, peers are also known to influence addiction-related behaviors such as rule breaking during adolescence. Association with delinquent peers increased rule breaking, particularly in those with NA alleles of GABRA2, while association with peers who displayed positive behaviors (such as religious activity and scholastic competence) decreased externalizing behaviors. Subjects with the ancestral alleles of GABRA2 were less affected by peers. Discussion: Our results illustrate a more complex influence of genotypes on risk for common traits such as those associated with addiction. Our data suggest that the previously identified “risk” alleles impacts the strength of both adaptive and maladaptive environmental influences on risky behaviors such as rule breaking, and hence might be thought of as plasticity factors. None of the alleles of the GABRA2 haplotypes change an amino acid, and they have no major influence on level of expression (eQTL) in several large brain expression samples. Allelic imbalance is more sensitive to slight expression differences. Preliminary allelic imbalance RNASeq analyses of brain mRNA suggest that the NA alleles may increase expression. Our result is in contrast to a recent report of decreased expression of the risk alleles in iPSCs. Epigenetic modifications may be one pathway of how environmental influences on behavior are being molecularly mediated, and just as our results demonstrate plasticity at the behavioral level, plasticity may exist on the molecular level, in that expression of GABRA2 haplotypes may be influenced molecularly by different environments.
Disclosure Nothing to Disclose. http://dx.doi.org/10.1016/j.euroneuro.2016.09.453
M63. GENOME WIDE ASSOCIATION RESULTS OF ALCOHOLIC USE DISORDER PATIENTS AND HEALTHY CONTROLS
Swapnil Awasthi1, Eva Friedel1, HansUlrich Wittchen2, Andreas Heinz1, Henrik Walter1, Stephan Ripke3 1
Charite Universitatsmedizien 2 Universität Dresden, Dresden, Germany 3 MGH
S411
Background: Recent genome wide association studies (GWAS) of alcoholic use disorder (AUD) have estimated a 40-60% variance explained by common genetic variants. Since GWAS require large sample sizes this study aims at combining genetics and imaging data to help finding genetics risk variants in smaller datasets (referred to as imaging genetics data). In this presentation we will focus on the genetic results of this study Methods: We had genotypic data of 115 cases of AUD according to DSM-IV-TR and 286 healthy controls. We used Ricopili pipeline for quality control (QC), principal component analysis (PCA), imputation, association analysis and polygenic risk scoring. Standard thresholds during QC were applied: call rate for individuals 98%, call rate for SNPs 98%, missing differences between cases and controls 2%, Hardy Weinberg Equilibrium for controls 10e-06, HWE for case 10e-10. We also checked for sex inconsistencies between reported sex and empirical sex as well as for related individuals between all the individuals by using identity by descent on a set of LD-pruned SNPs. After removing the related and overlapping individuals, PCA results were checked for hidden population stratification. Pre-phasing was done using Shapeit and imputation using Impute2 (using 1000 Genomes phase 1 as reference). We clumped GWAS summary statistics from the biggest currently available AUD dataset (n=16,087) and scored the individuals of our study. We also inspected for genetic correlation between Schizophrenia by using the largest available data sets from the Psychiatric Genomic Consortium (PGC). Results: All cases passed QC but four controls were excluded due to call rate. Out of 603,132 SNPs, 42% (253,361) were excluded, most of which were invariant (39%). PCA revealed slight population stratification without strong outliers; overlap testing could reveal one related and one completely overlapping case, of which one each was excluded from downstream analysis. After imputation 8,696,851 SNPs, 115 cases and 280 controls were chosen for the analysis. As expected, we did not find any genome wide significant SNPs. The QQ plots showed only slight inflation from expectation (lambda = 1.054). Polygenic risk score with four alcoholism phenotypes (two Discovery sample for alcohol dependent and ordinal trait and two SAGE sample with alcohol dependent and ordinal trait) (Gelernter J, et al, 2014) could not show significant separation of cases and controls in our study for any of the p-value thresholds. We also found no significant correlation with schizophrenia as the training dataset. Discussion: QC, imputation of this study was straightforward and as expected could not identify new genome wide significant SNPs. The genetic data is now ready to use to understand more about genetic aetiology of AUD using imaging genetics. Polygenic risk scoring with the biggest data available for AUD GWAS showed no signal, which might be due to low power of the training data (Gelernter J, et al, 2014). Newer and much more powerful PGC results will soon give more insight. As expected this study also showed no genetic association with schizophrenia.
Disclosure Nothing to Disclose. http://dx.doi.org/10.1016/j.euroneuro.2016.09.454
Abstracts of the XXIV World Congress of Psychiatric Genetics (WCPG) 2016
M61. SHARED GENETIC VARIANTS BETWEEN SCHIZOPHRENIA AND GENERAL COGNITIVE FUNCTION INDICATE COMMON MOLECULAR GENETIC MECHANISMS
Olav B Smeland1, Karolina Kauppi2,3,4, Yunpeng Wang1,2,3, W David Hill5,6, Gail Davies5,6, Oleksandr Frei1, Wen Li1, Jon A Eriksen1, Aree Witoelar1, Francesco Bettella1, Chun C Fan2,7, Wes Thompson8, Schizophrenia Working Group of the Psychiatric Genomics Consortium, neuroCHARGE Cognitive Working Group1,2,3,4,5,6,7,8,9,10,11, Chi-Hua Chen2,3, Srdjan Djurovic9,10, Ian J Deary5,6, Anders M Dale2,3,8,11, Ole A Andreassen1,8 1
NORMENT, KG Jebsen Centre for Psychosis Research, Institute of Clinical Medicine, University of Oslo and Division of Mental Health and Addiction, Oslo University Hospital, Oslo, Norway 2 Department of Radiology, University of California, San Diego, La Jolla, CA, United States of America 3 Multimodal Imaging Laboratory, University of California San Diego, La Jolla, CA, United States of America 4 Department of Radiation Sciences, Umeå University, SE-901 87 Umeå Sweden 5 Centre for Cognitive Ageing and Cognitive Epidemiology, University of Edinburgh, Edinburgh, UK 6 Department of Psychology, University of Edinburgh, Edinburgh, UK 7 Department of Cognitive Science, University of California, San Diego, La Jolla, CA, USA 8 Department of Psychiatry, University of California, San Diego, La Jolla, CA, USA 9 Department of Medical Genetics, Oslo University Hospital and University of Oslo, Oslo, Norway 10 NORMENT, KG Jebsen Centre for Psychosis Research, Department of Clinical Science, University of Bergen, Bergen, Norway 11 Department of Neurosciences, University of California San Diego, La Jolla, CA 92093, United States of America Background: Schizophrenia (SCZ) is a severe mental disorder characterized by widespread cognitive impairments including deficits in learning, memory, processing speed, attention and executive functioning. Although cognitive deficits are a strong predictor of functional outcome in SCZ, current treatment strategies largely fail to ameliorate these impairments. Thus, in order to develop more efficient treatment strategies in SCZ, a better understanding of the pathogenesis of these cognitive deficits is needed. Given that both SCZ and cognitive ability are substantially heritable, we here aimed to determine whether SCZ share genetic influences with general cognitive function (COG), a phenotype that captures the shared variation in performance across several cognitive domains.
Methods: We analyzed GWAS results in the form of summary statistics (p-values and z-scores) from SCZ (the Psychiatric Genomics Consortium; n = 82 315) and COG (CHARGE Consortium; n= 53 949). We applied a conditional false discovery rate (FDR) framework. By leveraging SNPassociations in a secondary trait (SCZ or COG), the conditional FDR approach increases power to detect loci in the primary trait (COG or SCZ), regardless of the directions of allelic effects of the risk loci. We then applied the conjunction FDR to identify shared loci between the phenotypes. The conjunction FDR is defined as the maximum of the conditional FDRs for both directions, and we used an overall FDR threshold of 0.05. Results: To visualize pleiotropic enrichment, we constructed conditional Q-Q plots which indicate substantial polygenetic overlap between SCZ and COG. For progressively stringent p-value thresholds for SCZ SNPs, we found approximately 150-fold enrichment for COG. For progressively stringent p-value thresholds for COG SNPs, we found approximately 100-fold enrichment for SCZ. We then used the conjunction FDR and identified fourteen independent loci shared between SCZ and COG. The majority of the shared loci show inverse associations in SCZ and COG, in line with the observed cognitive dysfunction in SCZ. Discussion: Our preliminary findings indicate shared molecular genetic mechanisms between SCZ and COG, which may provide important new insights into the pathogenesis of cognitive dysfunction in SCZ.
Disclosure Nothing to Disclose. http://dx.doi.org/10.1016/j.euroneuro.2016.09.452
M62. GABRA2 ASSOCIATION WITH ADDICTION-RELATED ENDOPHENOTYPES IS ENVIRONMENTALLY INFLUENCED
Margit Burmeister, Elisa Trucco, Shweta Ramdas, Stephen Parker, Mary Heitzeg, Sandra Villafuerte, Robert Zucker University of Michigan Background: The GABRA2 gene contains more than 250 SNPs that in Caucasians form two major common “Yin-Yang” haplotypes - having either all or none of the non-ancestral (NA) alleles. Since 2004, association of the non-ancestral, slightly less common haplotype with alcohol use disorder (AUD) and with increased beta waves in EEG has been reported and replicated many times. Methods: In a longitudinal sample of 4400 low SES families at high risk for AUD, we investigated the influence of these GABRA2 haplotypes on addiction related endophenotypes, and started to investigate potential functional implications. We found that the NA haplotype is associated with impulsivity and with increased activation of the insula in response to reward expectation, both of which partially mediate the
Abstracts of the XXIV World Congress of Psychiatric Genetics (WCPG) 2016 association of the NA alleles with alcoholism. Impulsivity was also associated with reward expectation, independent of genetics. Results: Moreover, the effect of parental monitoring on externalizing behavior trajectories across childhood and adolescence, i.e., consistently low, developmentally limited, and rising trajectories, was moderated by the NA alleles of GABRA2. Subjects with the NA alleles were more strongly influenced – both positively and negatively–by the extent of parental monitoring, while subjects with the ancestral alleles were not significantly influenced by the monitoring. In addition to parents, peers are also known to influence addiction-related behaviors such as rule breaking during adolescence. Association with delinquent peers increased rule breaking, particularly in those with NA alleles of GABRA2, while association with peers who displayed positive behaviors (such as religious activity and scholastic competence) decreased externalizing behaviors. Subjects with the ancestral alleles of GABRA2 were less affected by peers. Discussion: Our results illustrate a more complex influence of genotypes on risk for common traits such as those associated with addiction. Our data suggest that the previously identified “risk” alleles impacts the strength of both adaptive and maladaptive environmental influences on risky behaviors such as rule breaking, and hence might be thought of as plasticity factors. None of the alleles of the GABRA2 haplotypes change an amino acid, and they have no major influence on level of expression (eQTL) in several large brain expression samples. Allelic imbalance is more sensitive to slight expression differences. Preliminary allelic imbalance RNASeq analyses of brain mRNA suggest that the NA alleles may increase expression. Our result is in contrast to a recent report of decreased expression of the risk alleles in iPSCs. Epigenetic modifications may be one pathway of how environmental influences on behavior are being molecularly mediated, and just as our results demonstrate plasticity at the behavioral level, plasticity may exist on the molecular level, in that expression of GABRA2 haplotypes may be influenced molecularly by different environments.
Disclosure Nothing to Disclose. http://dx.doi.org/10.1016/j.euroneuro.2016.09.453
M63. GENOME WIDE ASSOCIATION RESULTS OF ALCOHOLIC USE DISORDER PATIENTS AND HEALTHY CONTROLS
Swapnil Awasthi1, Eva Friedel1, HansUlrich Wittchen2, Andreas Heinz1, Henrik Walter1, Stephan Ripke3 1
Charite Universitatsmedizien 2 Universität Dresden, Dresden, Germany 3 MGH
S411
Background: Recent genome wide association studies (GWAS) of alcoholic use disorder (AUD) have estimated a 40-60% variance explained by common genetic variants. Since GWAS require large sample sizes this study aims at combining genetics and imaging data to help finding genetics risk variants in smaller datasets (referred to as imaging genetics data). In this presentation we will focus on the genetic results of this study Methods: We had genotypic data of 115 cases of AUD according to DSM-IV-TR and 286 healthy controls. We used Ricopili pipeline for quality control (QC), principal component analysis (PCA), imputation, association analysis and polygenic risk scoring. Standard thresholds during QC were applied: call rate for individuals 98%, call rate for SNPs 98%, missing differences between cases and controls 2%, Hardy Weinberg Equilibrium for controls 10e-06, HWE for case 10e-10. We also checked for sex inconsistencies between reported sex and empirical sex as well as for related individuals between all the individuals by using identity by descent on a set of LD-pruned SNPs. After removing the related and overlapping individuals, PCA results were checked for hidden population stratification. Pre-phasing was done using Shapeit and imputation using Impute2 (using 1000 Genomes phase 1 as reference). We clumped GWAS summary statistics from the biggest currently available AUD dataset (n=16,087) and scored the individuals of our study. We also inspected for genetic correlation between Schizophrenia by using the largest available data sets from the Psychiatric Genomic Consortium (PGC). Results: All cases passed QC but four controls were excluded due to call rate. Out of 603,132 SNPs, 42% (253,361) were excluded, most of which were invariant (39%). PCA revealed slight population stratification without strong outliers; overlap testing could reveal one related and one completely overlapping case, of which one each was excluded from downstream analysis. After imputation 8,696,851 SNPs, 115 cases and 280 controls were chosen for the analysis. As expected, we did not find any genome wide significant SNPs. The QQ plots showed only slight inflation from expectation (lambda = 1.054). Polygenic risk score with four alcoholism phenotypes (two Discovery sample for alcohol dependent and ordinal trait and two SAGE sample with alcohol dependent and ordinal trait) (Gelernter J, et al, 2014) could not show significant separation of cases and controls in our study for any of the p-value thresholds. We also found no significant correlation with schizophrenia as the training dataset. Discussion: QC, imputation of this study was straightforward and as expected could not identify new genome wide significant SNPs. The genetic data is now ready to use to understand more about genetic aetiology of AUD using imaging genetics. Polygenic risk scoring with the biggest data available for AUD GWAS showed no signal, which might be due to low power of the training data (Gelernter J, et al, 2014). Newer and much more powerful PGC results will soon give more insight. As expected this study also showed no genetic association with schizophrenia.
Disclosure Nothing to Disclose. http://dx.doi.org/10.1016/j.euroneuro.2016.09.454