Genome-wide Significant SNPs Research Articles

SA74 - GENETIC UNDERPINNINGS OF SOCIAL WITHDRAWAL IN THE GENERAL POPULATION

Background Social withdrawal is present as a trait in the general population and is a common early symptom of neuropsychiatric conditions such as Alzheimer's Disease (AD), Schizophrenia (SCZ) and Major Depressive Disorder (MDD). However, the underlying, biological causes of social withdrawal (as trait or disease state) are still poorly understood and may also differ between diseases. With this study, we aim to elucidate the genetics – and hence the biological etiology – of social withdrawal within the general population and patients with AD, SCZ and MDD. Methods We conducted a Genome-Wide Association Study (GWAS) of social withdrawal in a Dutch general population sample, i.e. 2988 adult participants from the Nijmegen Biomedical Study (NBS). For the GWAS, we used the total score on a social withdrawal measure that was based on the answers to 4 items from the extraversion subscale of the Eysenck's Personality Questionnaire (EPQ) regarding social interaction avoidance, and 3 items from the Autism Spectrum Quotient (AQ) about (lack of) social skills. Based on the GWAS results, gene-wide analyses – i.e. taking into account all SNPs within each gene – were performed. Further, a molecular landscape was built based on the proteins encoded by the genes associated with the most significantly associated SNPs from the social withdrawal GWAS. Results Our social withdrawal measure showed a Cronbach's alpha of 0.693 within the NBS sample. The social withdrawal GWAS on the NBS sample did not result in genome-wide significant SNPs. The top SNPs were rs558196057 (P=1.43E-07) in an intron of AK8, and rs181663926 (P=8.11E-07) in an intron of TSC1. However, gene-wide analysis resulted in a significant association (P=2.30E-06) of the GIT1 gene with social withdrawal, which survived correction for the 17.876 genes tested. The molecular landscape provides some insights into the molecular interactions that are involved in social withdrawal in the general population. Discussion The association of social withdrawal with GIT1 is promising, because it encodes a protein involved in neurite outgrowth and synapse formation. The molecular landscape shows the interactions of GIT1 with other molecules linked to social withdrawal and as such provides clues to the mechanisms underlying social withdrawal in the general population. Future research will focus on analyzing social withdrawal in two other general population samples (i.e. the Leiden sample with 1241 subjects and the UK Biobank sample with 500.000 subjects) and the findings from these studies will be incorporated in the molecular landscape of social withdrawal. Moreover, based on the summary statistics from these GWASs, we will perform polygenic risk score-based analyses to determine the (putative) genetic link(s) between social withdrawal as a general population trait and social withdrawal as a disease state in patients with AD, SCZ and MDD.

Emerging Evidence For Overlap of ADHD Genetic Risk Factors With Those of Other Psychiatric Illnesses

This study is presented on behalf of the iPSYCH-Broad ADHD working group, the PGC ADHD working group and the PGC MDD working group.Earlier studies have shed light on the individually unique and shared genetic risk architecture of Attention Deficit Hyperactivity Disorder(ADHD), autism spectrum disorder (ASD), and major depressive disorder (MDD). This involved single trait and cross-disorder genome-wide association studies (GWAS) as well as studies targeted at estimating the degree of genetic variation within and covariation between disorders. Especially for ADHD (but also for ASD) these analyses were potentially biased due to availability of only small samples when compared to other (adult onset) disorders in the analyses. Moreover, the identification of shared risk loci for ADHD with other disorders was likely hindered by the very same limitation. Others and we have recently performed the largest GWASs to date for ADHD (~20k cases), ASD (~18k cases), and MDD (~130k cases). These studies identified numerous genome-wide significant risk loci for the individual disorders and therefore opened the door for reliable estimation of genetic covariation and cross-phenotype association studies that have enough power to detect shared risk loci. This is of interest as comorbidity with these disorders is common in patients with ADHD and identification of shared risk loci might open the door for a better understanding of disease etiology as well as treatment options and prevention. In a first step, we calculated the genetic correlation (rG) between ADHD, ASD, and MDD and found rGs of 0.35 for ADHD/ASD and 0.42 for ADHD/MDD with an rG of 0.44 for ASD and MDD. In a second step, we performed cross-phenotype GWAS analyses for the ADHD/ASD and ADHD/MDD pairs. We identified 8 regions with genome-wide significant SNPs, amongst others harboring the hostgene of miR-9, MANBA, and NUDT12 as shared risk factors for ADHD and ASD. We also identified 67 regions with genome-wide significant signals, amongst many others harboring SORCS3, NEGR1, and the hostgene of miR-9 as shared risk factors for ADHD and MDD. In addition, we are currently using, and will present results from, re-sampling procedures to study genetic correlation and individual SNP associations for individuals with ADHD and without comorbid ASD and/or MDD (and other phenotypes such as bipolar disorder, intellectual disability, and schizophrenia). We hope that these analyses together with other secondary analyses (e.g. using Mendelian Randomization) will help to identify genetically and phenotypically more homogenous subgroups of our sample in order to gain new insights in the etiology of ADHD and most importantly identify new strategies for prevention and treatment.

M98 - EVENT-RELATED P300 AMPLITUDE AS AN ENDOPHENOTYPE OF SCHIZOPHRENIA: ASSOCIATION WITH GENOME WIDE SIGNIFICANT INDEX SINGLE NUCLEOTIDE POLYMORPHISMS

Background Reduced amplitude of the event related P300 potential is a prominent finding in schizophrenia patients and a promising endophenotype for genetic research. As yet, little is known on common genetic risk factors for P300 amplitude reduction and the clinical manifestation of schizophrenia. Recently, the Psychiatric Genomics Consortium (PGC2) identified 128 independent associations spanning 108 conservatively defined loci that meet genome-wide significance. We used data from 269 normal subjects with both electrophysiological and genotype data to explore the possible relevance of these schizophrenia index SNPs on the P300 amplitude. Methods The sample comprised 269 unrelated healthy volunteers of German descent from the general population of Munich, Germany. Subjects with neuropsychiatric disorders were excluded. Evoked potentials were recorded using an auditory oddball paradigm with 80% non-target stimuli (540 tones, 500 Hz) and 20% target stimuli (135 tones, 1000 Hz) presented binaurally through headphones in a pseudo randomized order. The P300 amplitude was detected semiautomatically as the most positive value in the timeframe 250–500ms post-stimulus with a visual control afterwards at the electrode position Fz. The sample was genotyped on different platforms following established quality control and imputation protocols. After excluding two SNPs with insufficient quality of imputation, 126 genome-wide significant schizophrenia index SNPs from PGC2 could be evaluated for association with P300 amplitude by linear regression. Results There were 12 associations significant on the p=0.05 level (chr5_140143664_I, rs6002655, rs12887734, rs11191419, rs75059851, rs7907645, rs7432375, rs12421382, rs36068923, rs6065094, chr10_104957618_I, rs6670165). None of the associations survived Bonferroni correction for 126 comparisons. When employing the exact binomial test the number of significant associations observed exceeded chance expectation (p=0,024). Only one half of the associations showed effects in the expected direction (lower P300 amplitude associated with schizophrenia risk allele or amplitude associated with protective allele). Discussion Although we found an overrepresentation of significant associations of schizophrenia index SNPs with event related P300 amplitudes we were unable to establish the connection for any specific locus with sufficient statistical power. The pattern of observed effects was not supportive of a specific role of PGC2 schizophrenia risk alleles in P300 amplitude in normal subjects.

Functional Genomic Regulation In The Brain: (Epi)Genetic Variation, Neurodevelopment and Psychiatric Disease

Overall Abstract Epigenetic mechanisms play a role in the dynamic regulation of various genomic functions through covalent modifications to DNA and histones. These mechanisms establish and maintain cell-specific gene expression programs to orchestrate brain development, adult neurogenesis and synaptic plasticity. Given their role in the development and function of the brain, there is increasing awareness about the involvement of altered gene regulatory processes in the etiology of many neuropsychiatric phenotypes. This symposium will highlight how studies from different areas of epigenetics research can provide important insights into the molecular etiology of schizophrenia and cognitive function, and enable the identification of potential therapeutic targets. The first presentation will discuss a study of the NuRD nucleosome remodeling complex that mediates the projection of neurons across the cerebral hemispheres during development. SATB2, BCL11B and GATAD2A encode proteins within this complex and contain genome-wide significant risk SNPs for schizophrenia. Beyond these individually associated genes, this study will describe a gene set analysis of 127 genes involved in this process during cortical development that indicates strong enrichment of association signals in both schizophrenia and educational attainment GWAS data. These data point to a role for this chromatin remodeling process in the biology of schizophrenia and cognition. The second presentation will describe an analysis of dynamic DNA modifications (5mC and 5hmC) across human brain development, highlighting how the prenatal period is a time of considerable epigenomic plasticity in the brain, and the importance of neurodevelopmentally-dynamic loci in psychiatric phenotypes. It will highlight the impact of genetic variation on the epigenome during brain development showing that although most DNA methylation quantitative trait loci (mQTLs) are developmentally stable, a subset are characterized by fetal-specific effects and enriched amongst risk loci identified in recent schizophrenia GWAS. We will highlight the utility of integrating genetic and epigenetic data to fine-map GWAS regions using co-localization approaches. The third presentation will highlight results from a study of cell type-specific genome-wide maps of chromatin accessibility, which are critical resources to complement genomic sequence data and to correlate functional and genetic brain architecture. This study will describe the generation and analysis of an open chromatin atlas of different cell types in the adult human brain generated across multiple brain regions. Transcription factor-gene regulome networks were generated for each cell type and brain region and subnetworks affected by schizophrenia risk loci were identified. The final presentation will describe a study of non-coding RNAs in neuronal function. This study will highlight how epigenetic processes are potential therapeutic targets in neuropsychiatric disorders and will then detail research work showing that non-coding RNAs regulate epigenetic dynamics in neuronal systems to support cognitive function.

Gene-Set Analysis of Gwas Data Identifies A Role For Satb2 and The Nurd Complex In Schizophrenia and Educational Attainment

SATB2, BCL11B and GATAD2A map to genomic regions containing genome-wide significant SNPs for schizophrenia. These genes regulate key stages of neurodevelopment (e.g. axon guidance) via epigenetic mechanisms. SATB2, a DNA-binding protein, specifically mediates the projection of neurons across the cerebral hemispheres by regulating the activity of BCL11B via the NuRD nucleosome remodeling complex. GATAD2A is a core component of the NuRD complex. Given the schizophrenia association signals at SATB2, BCL11B and GATAD2A, we hypothesized that genes within the NuRD complex and genes regulated by SATB2 in the pre- and post-natal brain may also contribute to schizophrenia etiology. To test this, we developed three gene sets. 1) Genes reported in studies of SATB2 during cortical development (SATB2_Cortical, n=127 genes). 2) Genes mapping to SATB2 ChIP-seq peaks generated from mouse cortices at E15.5 (SATB2_Pre-natal, n=779). 3) Genes mapping to SATB2 ChIP-seq peaks generated from mouse P0 primary hippocampal neurons (SATB2_Post-natal, n=4,138). We performed competitive gene set analysis (GSA) using MAGMA to jointly analyse all SNPs within a gene set to test if those genes were more strongly associated with schizophrenia than other genes in the genome. We applied GSA to the largest available GWAS results dataset for schizophrenia (PGC2; 36,989 cases, 113,075 controls). Based on the genetic overlap between schizophrenia and cognition and the role for SATB2 in memory function, we also investigated these gene-sets for a genetic contribution to educational attainment (EA) using the largest available GWAS dataset (405,072 samples). After multiple test correction, we observed significant associations for (1) the SATB2_Cortical gene-set with schizophrenia (P=0.00012) and with EA (P=0.00049), (2) the SATB2_Pre-natal gene-set with EA (P=0.00675) and (3) the SATB2_Post-natal gene-set with EA (P=2.03x10-06). These data support a role for genes in the NuRD complex and genes regulated by SATB2 in the etiology of schizophrenia and EA.

SA85 - PHARMACOGENOMICS OF METHYLPHENIDATE TREATMENT OF DANISH CHILDREN WITH ADHD

Background The first choice drug for treatment of Attention Deficit Hyperactivity Disorder (ADHD) in Denmark and several other countries is Methylphenidate (MPH). More than 30% of the children treated with MPH have been reported to switch to another drug due to lack of efficacy or adverse reactions. Most likely, a significant fraction of this individual variation in treatment outcome is caused by genetic differences. Previous studies have not identified genes or genetic variants conclusively implicated in the response to MPH. Methods We performed a pharmacogenomics study of the efficacy and adverse reactions of MPH in a sample of 207 Danish children with ADHD, who were drug-naive at baseline and were monitored for 12 weeks. Each week the ADHD symptom severity was scored on the ADHD-RS questionnaire, and any adverse reactions were recorded. We divided the ADHD symptom scores into inattention and hyperactivity/impulsive sub-scores, and performed GWAS on the outcome of MPH treatment defined as the difference in symptom severity from baseline to end of study. This was followed by calculation of pair wise epistatic interaction coefficients on the top SNPs from the GWAS. Moreover, we used functional principal component analysis to derive proxies that describe the longitudinal nature of the data for both the efficacy measurements and the adverse reactions. Subsequently, we performed GWAS using the proxies as phenotypes for the MPH outcome across the entire treatment period. Results None of the GWAS resulted in genome-wide associated hits. However, among the top SNPs for the inattention sub-score were some that fell in: RAS guanyl releasing protein 1 (RASGRP1), sodium/potassium transporting ATPase interacting 2 (NKAIN2) and tetratricopeptide repeat domain 12 (TTC12). For the hyperactivity/impulsive sub-score one of the top SNPs is in the vicinity of gamma-aminobutyric acid type A receptor gamma3 subunit (GABRG3). In the interaction analysis we identified the SNPs with the largest difference in connectivity between the inattention and hyperactivity/impulsive sub-scores and found that some of these were located in or near: myelin transcription factor 1 like (MYT1L), synaptotagmin 17 (SYT17), potassium voltage-gated channel subfamily Q member 5 (KCNQ5) and potassium voltage-gated channel modifier subfamily S member 3 (KCNS3). Discussion Despite the lack of genome-wide significant SNPs, some of the top hits from the GWAS reside in or near genes that could modulate the response to MPH. Some of these genes have previously been implicated in psychiatric disorders (RASGRP1, NKAIN2 and MYT1L), are involved in neurotransmission (GABRG3, KCNQ5, KCNS3 and NKAIN2) or has been implicated in substance abuse, e.g. TTC12 has been implicated in heroin dependence and SYT17 in alcoholism.

19 - INVESTIGATING CAUSAL RELATIONSHIPS BETWEEN MAJOR DEPRESSIVE DISORDER AND GENETICALLY CORRELATED TRAITS USING MENDELIAN RANDOMISATION

Background Major Depressive Disorder is a common, complex, heterogeneous disorder that has been found to be associated with a wide array of traits and diseases such as coronary heart disease, type 2 diabetes and cancer. Due to the limitations of observational studies, it is not possible to establish whether such associations are causal, due to reverse causality or due to confounding. Mendelian Randomisation (MR) is an analytical technique that uses genetic instruments that are associated with the exposure of interest, to evaluate whether the association between exposure and outcome is causal in observational data. We sought to investigate the causal relationship between MDD and traits showing phenotypic and genetic correlation. Methods We used a novel multi-SNP MR method called Generalised Summary-based Mendelian Randomisation (GSMR) to investigate direction of causality between MDD and 4 traits with high genetic correlation with MDD – years of education (at the genetic level this is a proxy for general intelligence together with other psychosocial factors), body mass index (BMI), schizophrenia and coronary artery disease, for which GWAS summary statistics are publically available. Bi-directional analysis was performed. Sample sizes for risk factors ranged from 150,064 to 405,072, and the number of genome-wide significant SNPs included as independent instruments ranged from 59 to 124. A primary assumption of the MR method is that the genetic instrument does not have pleiotropic effects on the exposure and outcome. The GSMR method filters out SNPs with significant evidence of pleiotropy. Results GSMR analyses provided evidence for a 1.15 fold increase in MDD per standard deviation of BMI (P =2.7×10-7) and a 0.89 fold decrease in MDD per standard deviation of years of education (P=8.8×10-7). There was no evidence of reverse causality of MDD for BMI (P=0.81) or EDY (P =0.28). For BMI there was some evidence of pleiotropy, as eight SNPs were excluded as outliers. For CAD, the MR analyses were not significant when considering MDD as an outcome (P =0.39) or as an exposure (P = 0.13). There was significant evidence of bidirectional causality with a correlation in effect sizes of schizophrenia SNPs in MDD (P = 7.7×10–46) and MDD SNPs in schizophrenia (P = 6.3 × 10–15). Results from other MR methods such as inverse-weighted meta-analysis, weighted median and MR-Egger were similar to those from GSMR Discussion Our results suggest that MDD is not a causal risk factor for decreased years of education and increased BMI. In contrast, our results suggest that decreased years of education and increased BMI could be investigated as causal risk factors for MDD recognising that an alternative explanation is that these factors are correlated with other risk factors for MDD. Evidence from randomised control trials that have shown improved mood in depressed patients who adhere to a weight-loss plan is consistent with our results. Direction of causality could not be established for the association between MDD and CAD, indicating that the association may be due to other confounding factors. Lastly, our results suggest that MDD and schizophrenia have overlapping causal genetic pathways.

SA26 - SPARSE CANONICAL CORRELATION IN APPLICATION TO BIPOLAR PSYCHOTIC PHENOTYPES AND SCHIZOPHRENIA GENOME-WIDE SIGNIFICANT GENETIC LOCI

Background The aetiology of Bipolar Disorder (BD) and Schizophrenia (SZ) involves a complex gene-environment interplay and this involves common risk alleles that are shared across the disorders. Bipolar disorder is heterogeneous in presentation and we and others have postulated that the particular clinical presentation is in part determined by the particular risk alleles carried. To investigate this, we have employed sparse Canonical Correlation Analysis (sCCA), to seek patterns of correlation between psychotic symptoms and schizophrenia risk alleles in a large (N=3903) deeply phenotyped sample of subjects with BD. Methods Canonical Correlation Analysis (CCA) seeks linear relationships between two sets of multi-dimensional variables. In the psychiatric genetics field, this approach can be adapted to enable the analysis of both phenotype and genetic variables in one framework. This has the advantage of minimizing the number of statistical tests required. A recent extension of CCA to sparse canonical correlation analysis (sCCA) allows application of the general approach to high dimensional genomic data (Witten et al., 2009). sCCA has advantages compared to CCA in terms of precision of parameter estimates and statistical power. We employed sCCA to analyse 30 psychotic symptoms assessed with the OPerational CRITeria checklist (OPCRIT) and 82 independent genome-wide significant SNPs that were identified by the SZ PGC2 study (The Psychiatric Genomics Consortium, 2014) for which we had data in our BD sample. We applied the sCCA as a purely data driven approach to individual OPCRIT psychotic symptoms and also to groups of OPCRIT psychotic symptoms (“positive”, “negative”, “disorganised”) defined by a previously published factor analysis study of schizophrenia. Results sCCA analysis of the individual OPCRIT items with the set of significantly associated with SZ 82 sCCA analysis based on individual psychotic symptoms revealed a significant association (p=0.045) with the largest weights attributed to delusions of influence, bizarre behaviour and grandiose delusions and an indel SNP on chromosome 3 (3:180594593, build 37). sCCA analysis on the same set of SNPs and OPCRIT symptom groups confirmed association with the same SNP and the “disorganised” group of OPCRIT symptoms (p=0.01). The results also show significant association when we use less stringent p-value thresholds in SZ. In this case the most heavily weighted SNP remains 3:180594593, but other SNPs also contribute to the analysis. Discussion Our results suggest the CCA canonical correlation approach might be a useful tool to explore phenotype-genotype relationships. It can be applied to generate data-driven hypotheses with the potential to provide further insights into the complex architecture of psychiatric disorders. To the best of our knowledge, this is the first study to apply this approach in analysing phenotypes and genotypes together. sCCA offers the potential to be extended to include a larger number of fine grained systematic descriptors of BD, and for testing for correlation with genetic profiles from other co-morbid disorders.

Genetic determinants of glucose levels in pregnancy: genetic risk scores analysis and GWAS in the Norwegian STORK cohort.

Objective Hyperglycaemia during pregnancy increases the risk of adverse health outcomes in mother and child, but the genetic aetiology is scarcely studied. Our aims were to (1) assess the overlapping genetic aetiology between the pregnant and non-pregnant population and (2) assess the importance of genome-wide polygenic contributions to glucose traits during pregnancy, by exploring whether genetic risk scores (GRSs) for fasting glucose (FG), 2-h glucose (2hG), type 2 diabetes (T2D) and BMI in non-pregnant individuals were associated with glucose measures in pregnant women. Methods We genotyped 529 Norwegian pregnant women and constructed GRS from known genome-wide significant variants and SNPs weakly associated (p > 5 × 10-8) with FG, 2hG, BMI and T2D from external genome-wide association studies (GWAS) and examined the association between these scores and glucose measures at gestational weeks 14-16 and 30-32. We also performed GWAS of FG, 2hG and shape information from the glucose curve during an oral glucose tolerance test (OGTT). Results GRSFG explained similar variance during pregnancy as in the non-pregnant population (~5%). GRSBMI and GRST2D explained up to 1.3% of the variation in the glucose traits in pregnancy. If we included variants more weakly associated with these traits, GRS2hG and GRST2D explained up to 2.4% of the variation in the glucose traits in pregnancy, highlighting the importance of polygenic contributions. Conclusions Our results suggest overlap in the genetic aetiology of FG in pregnant and non-pregnant individuals. This was less apparent with 2hG, suggesting potential differences in postprandial glucose metabolism inside and outside of pregnancy.

Pipeline design to identify key features and classify the chemotherapy response on lung cancer patients using large-scale genetic data

BackgroundDuring the last decade, the interest to apply machine learning algorithms to genomic data has increased in many bioinformatics applications. Analyzing this type of data entails difficulties for managing high-dimensional data, class imbalance for knowledge extraction, identifying important features and classifying individuals. In this study, we propose a general framework to tackle these challenges with different machine learning algorithms and techniques. We apply the configuration of this framework on lung cancer patients, identifying genetic signatures for classifying response to drug treatment response. We intersect these relevant SNPs with the GWAS Catalog of the National Human Genome Research Institute and explore the Regulomedb, GTEx databases for functional analysis purposes.ResultsThe machine learning based solution proposed in this study is a scalable and flexible alternative to the classical uni-variate regression approach to analyze large-scale data. From 36 experiments executed using the machine learning framework design, we obtain good classification performance from the top 5 models with the highest cross-validation score and the smallest standard deviation. One thousand two hundred twenty four SNPs corresponding to the key features from the top 20 models (cross validation F1 mean >= 0.65) were compared with the GWAS Catalog finding no intersection with genome-wide significant reported hits. From these, new genetic signatures in MAE, CEP104, PRKCZ and ADRB2 show relevant biological regulatory functionality related to lung physiology.ConclusionsWe have defined a machine learning framework using data with an unbalanced large data-set of SNP-arrays and imputed genotyping data from a pharmacogenomics study in lung cancer patients subjected to first-line platinum-based treatment. This approach found genome signals with no genome-wide significance in the uni-variate regression approach (GWAS Catalog) that are valuable for classifying patients, only few of them with related biological function. The effect results of these variants can be explained by the recently proposed omnigenic model hypothesis, which states that complex traits can be influenced mostly by genes outside not only by the “core genes”, mainly found by the genome-wide significant SNPs, but also by the rest of genes outside of the “core pathways” with apparent unrelated biological functionality.

A genome-wide association study for harness racing success in the Norwegian-Swedish coldblooded trotter reveals genes for learning and energy metabolism

BackgroundAlthough harness racing is of high economic importance to the global equine industry, significant genomic resources have yet to be applied to mapping harness racing success. To identify genomic regions associated with harness racing success, the current study performs genome-wide association analyses with three racing performance traits in the Norwegian-Swedish Coldblooded Trotter using the 670 K Axiom Equine Genotyping Array.ResultsFollowing quality control, 613 horses and 359,635 SNPs were retained for further analysis. After strict Bonferroni correction, nine genome-wide significant SNPs were identified for career earnings. No genome-wide significant SNPs were identified for number of gallops or best km time. However, four suggestive genome-wide significant SNPs were identified for number of gallops, while 19 were identified for best km time. Multiple genes related to intelligence, energy metabolism, and immune function were identified as potential candidate genes for harness racing success.ConclusionsApart from the physiological requirements needed for a harness racing horse to be successful, the results of the current study also advocate learning ability and memory as important elements for harness racing success. Further exploration into the mental capacity required for a horse to achieve racing success is likely warranted.

A genome-wide association study identifies a susceptibility locus for biliary atresia on 2p16.1 within the gene EFEMP1.

Biliary atresia (BA) is a rare pediatric cholangiopathy characterized by fibrosclerosing obliteration of the extrahepatic bile ducts, leading to cholestasis, fibrosis, cirrhosis, and eventual liver failure. The etiology of BA remains unknown, although environmental, inflammatory, infectious, and genetic risk factors have been proposed. We performed a genome-wide association study (GWAS) in a European-American cohort of 343 isolated BA patients and 1716 controls to identify genetic loci associated with BA. A second GWAS was performed in an independent European-American cohort of 156 patients with BA and other extrahepatic anomalies and 212 controls to confirm the identified candidate BA-associated SNPs. Meta-analysis revealed three genome-wide significant BA-associated SNPs on 2p16.1 (rs10865291, rs6761893, and rs727878; P < 5 ×10−8), located within the fifth intron of the EFEMP1 gene, which encodes a secreted extracellular protein implicated in extracellular matrix remodeling, cell proliferation, and organogenesis. RNA expression analysis showed an increase in EFEMP1 transcripts from human liver specimens isolated from patients with either BA or other cholestatic diseases when compared to normal control liver samples. Immunohistochemistry demonstrated that EFEMP1 is expressed in cholangiocytes and vascular smooth muscle cells in liver specimens from patients with BA and other cholestatic diseases, but it is absent from cholangiocytes in normal control liver samples. Efemp1 transcripts had higher expression in cholangiocytes and portal fibroblasts as compared with other cell types in normal rat liver. The identification of a novel BA-associated locus, and implication of EFEMP1 as a new BA candidate susceptibility gene, could provide new insights to understanding the mechanisms underlying this severe pediatric disorder.

Abstract A37: Genome-wide association study of colorectal tumors in Latin American individuals from Colombia

Abstract Introduction: Colorectal cancer (CRC) is one of the most common cancers worldwide, with higher incidence among more developed countries. However, mortality rates are higher and increasing in less-developed regions, including Colombia. Incidence and mortality rates of CRC vary across the country, with the lowest rates among populations with increased concentration of Native Americans. Our group previously reported that increased European and African ancestry was associated with higher adenomatous polyps (AP) and CRC risk in Colombians. Based on this observation we hypothesize that genetic contribution to the observed regional variation on CRC development in Colombia might be partly explained by the influence of genetic variants with different frequency distribution in the European, Amerindian, and African ancestral components. Objective: To test our hypothesis we performed a genome-wide association study (GWAS) of colorectal tumors in Colombian CRC and AP cases and controls. Materials and Methods: We conducted a GWAS of Colombians from different regions of the country that included 162 CRC cases, 122 AP cases, and 131 controls. Only SNPs with MAF ≥ 0.05 were included. The role of these common variants was assessed by logistic regression models adjusted for age, sex, and the first 10 principal components (PCs) of the genetic variation. Analyses were performed for a total of 606410 SNPs. The P value threshold for genome-wide significant SNPs was set at the traditional value of ≤ 5 × 10−8. Results: We replicated some of the previously reported associations (8 out of 44 SNPs, P ≤ 0.05) of which 7 had directionally concordant ORs. Regarding the remaining 36 published risk variants, 25 had directionally concordant ORs, even though neither of them was significant at nominal P ≤ 0.05. None of the SNPs analyzed passed the threshold for genome-wide significance. However, we found multiple suggestive associations. One novel SNP within the region 2q11.2-q12 near the gene MAP4K4 was associated with the risk of AP (rs4280476-G; OR 0.36; 95% CI 0.23-0.56; P = 8.7 x 10-6). We also found four novel SNPs across three regions with suggestive CRC risk associations: rs10514122-A located in the locus 5q14 near the TBCA gene (OR 0.30; 95% CI 0.18-0.49; P = 3.6 x 10-6), rs10746883-G (OR 0.37; 95% CI 0.25-0.56; P = 3.2 x 10-6) and the rs10781032-A (OR 0.35; 95% CI 0.22-0.53; P = 1.7 x 10-6), in high LD with each other (R2 = 0.81), located in the locus 9q21.1-q21.2 near the genes TRPM3 and TMEM2 and the SNP rs2800641-G located in the region 6q25.2 near the MYCT1 gene (OR 2.96; 95% CI 1.83-4.77; P = 8.9 x 10-6). The inflation factors (λ), when comparing AP or CRC cases vs controls, were 1.05 and 1.04, respectively, indicating appropriate control for population stratification. Finally, it is of note that the MAF reported for the 5 novel SNPs identified in our study shows important differences between European, Asian, and African reference populations from the public database 1,000 genomes. Conclusions: In this GWAS of AP and CRC in individuals from Colombia, we replicated some of the previously reported associations and identified suggestive signals for 5 novel risk alleles located at four loci (2q11.2-q12, 5q14, 9q21.1-q21.2, and 6q25.2). Further work is needed to replicate our findings in other Latin American populations. Citation Format: Maria Carolina Sanabria-Salas, Laura Fejerman, Konrad Rawlik, Albert Tenesa, Martha L. Serrano-Lopez, Jovanny Zabaleta, Gustavo A. Hernandez-Suarez. Genome-wide association study of colorectal tumors in Latin American individuals from Colombia [abstract]. In: Proceedings of the Tenth AACR Conference on the Science of Cancer Health Disparities in Racial/Ethnic Minorities and the Medically Underserved; 2017 Sep 25-28; Atlanta, GA. Philadelphia (PA): AACR; Cancer Epidemiol Biomarkers Prev 2018;27(7 Suppl):Abstract nr A37.

Polygenic risk score of SERPINA6/SERPINA1 associates with diurnal and stress-induced HPA axis activity in children

PurposeCorticosteroid-binding globulin (CBG) transports glucocorticoids in blood. Variation in genes SERPINA6 encoding for CBG, SERPINA2 and SERPINA1 (serpin family A member 6, 2, and 1) have been shown to influence morning plasma cortisol and CBG in adults. However, association of this genetic variation with diurnal and stress-induced salivary cortisol remain unknown. This study aims to investigate the effect of genetic variation in SERPINA6/2/1 loci on diurnal and stress-induced salivary cortisol in children. MethodsWe studied 186, 8-year-old children with genome-wide genotyping. We generated weighted polygenic risk score (PRS) based on 6 genome-wide significant SNPs (rs11621961, rs11629171, rs7161521, rs2749527, rs3762132, rs4900229) derived from the CORNET meta-analyses. Salivary cortisol was measured across one day and in response to the Trier Social Stress Test for Children (TSST-C). ResultsMixed models, adjusted for covariates, showed that the PRS x sampling time interactions associated with diurnal (P < 0.001) and stress-induced (P = 0.009) salivary cortisol. In the high PRS group (dichotomized at median) the diurnal salivary cortisol pattern decreased less from awakening to bedtime than in the low PRS group (standardized estimates of sampling time −0.64 vs. −0.73, P < 0.0001 for both estimates). In response to stress, salivary cortisol increased in the high PRS group while it remained unchanged in the low PRS group (standardized estimates of sampling time 0.12, P = 0.015 vs. −0.06, P = 0.16). These results were mainly driven by minor alleles of rs7161521 (SERPINA6) and rs4900229 (SERPINA1). ConclusionsGenetic variation in SERPINA6/2/1loci may underpin higher hypothalamic-pituitary-adrenocortical axis activity in children.

Human Lung DNA Methylation Quantitative Trait Loci Colocalize with Chronic Obstructive Pulmonary Disease Genome-Wide Association Loci.

As the third leading cause of death in the United States, the impact of chronic obstructive pulmonary disease (COPD) makes identification of its molecular mechanisms of great importance. Genome-wide association studies (GWASs) have identified multiple genomic regions associated with COPD. However, genetic variation only explains a small fraction of the susceptibility to COPD, and sub-genome-wide significant loci may play a role in pathogenesis. Regulatory annotation with epigenetic evidence may give priority for further investigation, particularly for GWAS associations in noncoding regions. We performed integrative genomics analyses using DNA methylation profiling and genome-wide SNP genotyping from lung tissue samples from 90 subjects with COPD and 36 control subjects. We performed methylation quantitative trait loci (mQTL) analyses, testing for SNPs associated with percent DNA methylation and assessed the colocalization of these results with previous COPD GWAS findings using Bayesian methods in the R package coloc to highlight potential regulatory features of the loci. We identified 942,068 unique SNPs and 33,996 unique CpG sites among the significant (5% false discovery rate) cis-mQTL results. The genome-wide significant and subthreshold (P < 10-4) GWAS SNPs were enriched in the significant mQTL SNPs (hypergeometric test P < 0.00001). We observed enrichment for sites located in CpG shores and shelves, but not CpG islands. Using Bayesian colocalization, we identified loci in regions near KCNK3, EEFSEC, PIK3CD, DCDC2C, TCERG1L, FRMD4B, and IL27. Colocalization of mQTL and GWAS loci provides regulatory characterization of significant and subthreshold GWAS findings, supporting a role for genetic control of methylation in COPD pathogenesis.

Genome-wide association study of depressive symptoms in the Hispanic Community Health Study/Study of Latinos

Although genome-wide association studies (GWAS) have identified several variants linked to depression, few GWAS of non-European populations have been performed. We conducted a genome-wide analysis of depression in a large, population-based sample of Hispanics/Latinos. Data came from 12,310 adults in the Hispanic Community Health Study/Study of Latinos (HCHS/SOL). Past-week depressive symptoms were assessed using the 10-item Center for Epidemiological Studies of Depression Scale. Three phenotypes were examined: a total depression score, a total score modified to account for psychiatric medication use, and a score excluding anti-depressant medication users. We estimated heritability due to common variants (h2SNP), and performed a GWAS of the three phenotypes. Replication was attempted in three independent Hispanic/Latino cohorts. We also performed sex-stratified analyses, analyzed a binary trait indicating probable depression, and conducted three trans-ethnic analyses. The three phenotypes exhibited significant heritability (h2SNP = 6.3–6.9%; p = .002) in the total sample. No SNPs were genome-wide significant in analyses of the three phenotypes or the binary indicator of probable depression. In sex-stratified analyses, seven genome-wide significant SNPs (one in females; six in males) were identified, though none were supported through replication. Four out of 24 loci identified in prior GWAS were nominally associated in HCHS/SOL. There was no evidence of overlap in genetic risk factors across ancestry groups, though this may have been due to low power. We conducted the largest GWAS of depression-related phenotypes in Hispanic/Latino adults. Results underscore the genetic complexity of depressive symptoms as a phenotype in this population and suggest the need for much larger samples.

GWAS of the electrocardiographic QT interval in Hispanics/Latinos generalizes previously identified loci and identifies population-specific signals

QT interval prolongation is a heritable risk factor for ventricular arrhythmias and can predispose to sudden death. Most genome-wide association studies (GWAS) of QT were performed in European ancestral populations, leaving other groups uncharacterized. Herein we present the first QT GWAS of Hispanic/Latinos using data on 15,997 participants from four studies. Study-specific summary results of the association between 1000 Genomes Project (1000G) imputed SNPs and electrocardiographically measured QT were combined using fixed-effects meta-analysis. We identified 41 genome-wide significant SNPs that mapped to 13 previously identified QT loci. Conditional analyses distinguished six secondary signals at NOS1AP (n = 2), ATP1B1 (n = 2), SCN5A (n = 1), and KCNQ1 (n = 1). Comparison of linkage disequilibrium patterns between the 13 lead SNPs and six secondary signals with previously reported index SNPs in 1000G super populations suggested that the SCN5A and KCNE1 lead SNPs were potentially novel and population-specific. Finally, of the 42 suggestively associated loci, AJAP1 was suggestively associated with QT in a prior East Asian GWAS; in contrast BVES and CAP2 murine knockouts caused cardiac conduction defects. Our results indicate that whereas the same loci influence QT across populations, population-specific variation exists, motivating future trans-ethnic and ancestrally diverse QT GWAS.

Genomewide association studies of suicide attempts in US soldiers.

Suicide is a global public health problem with particular resonance for the US military. Genetic risk factors for suicidality are of interest as indicators of susceptibility and potential targets for intervention. We utilized population-based nonclinical cohorts of US military personnel (discovery: N = 473 cases and N = 9778 control subjects; replication: N = 135 cases and N = 6879 control subjects) and a clinical case-control sample of recent suicide attempters (N = 51 cases and N = 112 control subjects) to conduct GWAS of suicide attempts (SA). Genomewide association was evaluated within each ancestral group (European-, African-, Latino-American) and study using logistic regression models. Meta-analysis of the European ancestry discovery samples revealed a genomewide significant locus in association with SA near MRAP2 (melanocortin 2 receptor accessory protein 2) and CEP162 (centrosomal protein 162); 12 genomewide significant SNPs in the region; peak SNP rs12524136-T, OR = 2.88, p = 5.24E-10. These findings were not replicated in the European ancestry subsamples of the replication or suicide attempters samples. However, the association of the peak SNP remained significant in a meta-analysis of all studies and ancestral subgroups (OR = 2.18, 95%CI 1.70, 2.80). Polygenic risk score (PRS) analyses showed some association of SA with bipolar disorder. The association with SNPs encompassing MRAP2, a gene expressed in brain and adrenal cortex and involved in neural control of energy homeostasis, points to this locus as a plausible susceptibility gene for suicidality that should be further studied. Larger sample sizes will be needed to confirm and extend these findings.

Genome-wide association study reveals putative role of gga-miR-15a in controlling feed conversion ratio in layer chickens

BackgroundEfficient use of feed resources for farm animals is a critical concern in animal husbandry. Numerous genetic and nutritional studies have been conducted to investigate feed efficiency during the regular laying cycle of chickens. However, by prolonging the laying period of layers, the performance of feed utilization in the late-laying period becomes increasingly important. In the present study, we measured daily feed intake (FI), residual feed intake (RFI) and feed conversion ratio (FCR) of 808 hens during 81–82 weeks of age to evaluate genetic properties and then used a genome-wide association study (GWAS) to reveal the genetic determinants.ResultsThe heritability estimates for the investigated traits were medium and between 0.15 and 0.28 in both pedigree- and genomic-based estimates, whereas the genetic correlations among these traits were high and ranged from 0.49 to 0.90. Three genome-wide significant SNPs located on chromosome 1 (GGA1) were detected for FCR. Linkage disequilibrium (LD) and conditional GWA analysis indicated that these 3 SNPs were highly correlated with one another, located at 13.55–45.16 Kb upstream of gga-miR-15a. Results of quantitative real-time polymerase chain reaction (qRT-PCR) analysis in liver tissue showed that the expression of gga-miR-15a was significantly higher in the high FCR birds than that in the medium or low FCR birds. Bioinformatics analysis further revealed that gga-mir-15a could act on many target genes, such as forkhead box O1 (FOXO1) that is involved in the insulin-signaling pathway, which influences nutrient metabolism in many organisms. Additionally, some suggestively significant variants, located on GGA3 and GGA9, were identified to associate with FI and RFI.ConclusionsThis GWA analysis was conducted on feed intake and efficiency traits for chickens and was innovative for application in the late laying period. Our findings can be used as a reference in the genomic breeding programs for increasing the efficiency performance of old hens and to improve our understanding of the molecular determinants for feed efficiency.

A genome-wide association study suggests new candidate genes for milk production traits in Chinese Holstein cattle.

A genome-wide association study (GWAS) was conducted on 15 milk production traits in Chinese Holstein. The experimental population consisted of 445 cattle, each genotyped by the GGP (GeneSeek genomic profiling)-BovineLD V3 SNP chip, which had 26151 public SNPs in its manifest file. After data cleaning, 20326 SNPs were retained for the GWAS. The phenotypes were estimated breeding values of traits, provided by a public dairy herd improvement program center that had been collected once a month for 3years. Two statistical models, a fixed-effect linear regression model and a mixed-effect linear model, were used to estimate the association effects of SNPs on each of the phenotypes. Genome-wide significant and suggestive thresholds were set at 2.46E-06 and 4.95E-05 respectively. The two statistical models concurrently identified two genome-wide significant (P<0.05) SNPs on milk production traits in this Chinese Holstein population. The positional candidate genes, which were the ones closest to these two identified SNPs, were EEF2K (eukaryotic elongation factor 2 kinase) and KLHL1 (kelch like family member 1). These two genes could serve as new candidate genes for milk yield and lactation persistence, yet their roles need to be verified in further function studies.