- Email: [email protected]
FUMA: FUNCTIONAL MAPPING AND ANNOTATION OF GENETIC ASSOCIATIONS
EXOME-WIDE ASSOCIATION STUDY IDENTIFIES NOVEL SUSCEPTIBILITY GENES FOR PERSONALITY TRAITS 3 4
Virginia Commonwealth University University of Melbourne
Background: Mates are similar to one another across a large number of traits, including liabilities underlying psychiatric disorders. It has long been known that when such phenotypic similarity implies genetic similarity (hereafter, assortative mating), heritability estimates from family and twin studies can be biased. However, the effects of assortative mating, if any, on estimates of heritability from all SNPs in unrelated individuals (SNP-based heritability, h2SNP) have not been investigated to date. Methods: We derived mathematically the expected behavior of estimates from Haseman-Elston regression and GREML, and assessed these predictions via simulation. We then assessed in the UK Biobank whether the predicted signatures of assortative mating on h2SNP were observed. Results: We show analytically and via simulation that estimates of h2SNP from both GREML and Haseman-Elston regression are typically biased upwards in the presence of assortative mating. However, estimates from Haseman-Elston regression are roughly constant as a function of sample size and number of markers, whereas the ratio of sample size to number of markers strongly affects those from GREML. This difference in estimate behavior allows one to assess in real data whether assortative mating is affecting estimates of h2SNP. When we did this in the UK Biobank for height and IQ, we observed that h2SNP estimates for height that behaved as predicted, but not for h2SNP estimates for IQ. Discussion: For many traits, the degree of bias in estimates of h2SNP is expected to be small, but it can be nontrivial for other traits, depending on the degree of assortative mating and the heritability of the traits. For example, based on spousal similarity, we predict that h2SNP estimates of height and IQ are 25% greater than their true equilibrium values. That we did not observe the predicted pattern of h2SNP estimates for IQ may either suggest that spousal similarity for IQ occurs via mechanisms that do not lead to genetic similarity (e.g., social homogamy) or that the effects of assortative mating on the covariance between causal variants is counter-acted by the forces of natural selection.
S789
relatively common in the population. Different mechanisms have been proposed to explain this so-called ‘evolutionary paradox’, including balancing selection, the mismatch hypothesis and the polygenic mutation-selection balance. Evolutionary analyses of individual genetic loci can be informative for monogenic diseases associated with large selection pressures, but have less utility in the analysis of complex, polygenic disorders due to associations of many low-risk alleles. Taking a polygenic approach to mental disorders, we explore the impact of evolutionary pressure on allele frequency under the hypotheses that lower-thanexpected average frequencies of disease risk alleles are indicative of negative selection, while higher-thanexpected frequencies are indicative of positive selection. Methods: Using the largest available GWAS summary statistics for psychiatric and other disorders and traits, we identified allele frequencies for the disease risk or traitincreasing alleles for all SNPs genome-wide. The SNPs were thinned according to P-value via clumping, using stringent linkage disequilibrium criteria, to ensure their independence. After ordering the clumped SNPs by association P-value or effect size in the GWAS, we tested whether mean trait-increasing allele frequencies differed from 50%, as expected by chance, and investigated whether the effect was greater for those SNPs with largest association signal. Results: We found evidence of negative selection on Schizophrenia, Crohn's disease and obesity (BMI) risk alleles and of positive selection for height, education and ADHD alleles. The direction of evolutionary pressure was inconclusive for other traits tested including bipolar disorder, depression and autism. For each trait, we estimated the number of SNPs that maximised the mean trait-increasing allele frequency deviation from 50%. Discussion: Our results suggest that evolutionary pressure has had a small but marked impact on the allele frequencies of thousands of SNPs of certain psychiatric disorders and other traits. Unlike measures of fecundity in population cohorts, which are restricted by current environmental pressures, social trends, and changes in reproductive behaviour, evidence of positive or negative selection on the genome provides insights into our understanding of the impact of disease on the population over generations.
Disclosure: Nothing to disclose. Disclosure: Nothing to disclose. http://dx.doi.org/10.1016/j.euroneuro.2017.08.017 http://dx.doi.org/10.1016/j.euroneuro.2017.08.016
16. MEASURING EVOLUTIONARY PRESSURE IN A POLYGENIC FRAMEWORK FOR COMPLEX DISEASES
17. FUMA: FUNCTIONAL MAPPING AND ANNOTATION OF GENETIC ASSOCIATIONS n
Evangelos Vassos , Paul O'Reilly, Cathryn Lewis
Kyoko Watanabe , Erdogan Taskesen, Arjen van Bochoven, Danielle Posthuma
King's College London
VU Amsterdam
Background: Although many mental disorders are herita-
Background: A main challenge in Genome-Wide Association Studies (GWAS) is to prioritize genetic variants and
n
ble and result in reproductive disadvantage, they remain
S790 identify potential causal mechanisms of human diseases. Although a variety of bioinformatics tools are now available for downstream analyses of GWAS results, a standard, integrative approach is lacking. We developed FUMA: a web-based platform to facilitate functional annotation of GWAS results, prioritization of potential causal genetic variants and genes, and interactive visualization by incorporating 14 biological data repositories and tools. FUMA is available as an online tool at http://fuma.ctglab.nl. Methods: FUMA contains two core functions to annotate input summary statistics to prioritize potential causal genetic variants and genes; SNP2GENE and GENE2FUNC. In SNP2GENE, SNPs are annotated with their biological functionality and mapped to genes based on positional and functional information of SNPs. In this process, significantly associated SNPs from GWAS are characterized as genomic risk loci by incorporating the linkage disequilibrium structure. Functionally annotated SNPs are mapped to genes based on functional consequences on genes (positional mapping), expression quantitative trait loci (eQTLs) and chromatin interactions of phenotype relevant tissue types (eQTL and chromatin interaction mappings). By combining these three mapping strategies, FUMA enables to prioritize genes that are highly likely involved in the trait of interest. To obtain insight into putative causal mechanisms, the GENE2FUNC process annotates the prioritized genes in biological context, such as tissue specific gene expression pattern, enrichment of gene sets and direct links to well defined external biological databases such as disease associated genes and drug targets. Results: We have applied FUMA to the most recent Body Mass Index (BMI)[1]. We successfully prioritized previously reported genes from the original GWAS study, and also novel candidates by combining positional mapping of deleterious coding SNPs and eQTL mapping. For example, IRX3 was prioritized by eQTLs from FTO locus which is the most significantly associated locus with BMI. Although, FTO is the only gene reported in the original study since the genomic risk locus is located within the FTO gene and not overlapped with any other genes, IRX3 has been recently validated as whose expression is affected by variants in the FTO locus[2]. Thus, by incorporating multiple biological information, we could prioritise genes that are located outside of genomic risk loci. Chromatin interaction mapping identified additional novel candidates, including genes located outside of the risk loci, which showed share biological functions with the previously reported genes. Discussion: In summary, FUMA provides an easy-to-use tool to interpret and functionally annotate results from genetic association studies. It allows to quickly gain insight into the directional biological implications of significant genetic associations. FUMA combines information of stateof-the-art biological data sources in a single platform to facilitate the generation of hypotheses for functional follow-up analysis aimed at proving causal relations between genetic variants and diseases.
Y. Ma, M. Li
Disclosure: Nothing to disclose.
References [1] A. Locke et al., “Genetic studies of body mass index yield new insights for obesity biology”. Nature 518 (7538):197–206, 2015. [2] M. Claussnitzer et al., “FTO obesity variant circuitry and adipocyte browning in humans”. The New England of Medicine 373(10):895–907, 2015. http://dx.doi.org/10.1016/j.euroneuro.2017.08.018
18. VALIDATING GWA ASSOCIATIONS IN PSYCHIATRIC DISORDERS WITH FUNCTIONAL GENOMIC DATA n
Maciej Trzaskowski ,1, Enda Byrne2, Jian Yang1, Naomi Wray1 1 2
University of Queensland Queensland Brain Institute
Background: Genome-wide association studies have been successful at identifying regions of the genome associated with disease. Often, the identified SNPs are in regions of linkage disequilibrium that stretch over large distances, making it difficult to ascertain which gene in the region is functionally relevant to the trait of interest. For disorders of the brain, assessing the functional relevance of SNPs in the same sample in which they were discovered is, of course, difficult. Methods: Here we applied the SMR (Summary-data based Mendelian Randomisation) method to the latest GWAS results of psychiatric disorders, to evaluate whether SNPs associated with these disorders are also associated with control of gene expression of genes in the same region. The advantage of this method is that it utilises GWAS and expression QTL (eQTL) data from different samples. eQTL data were taken from whole blood and brain samples. Results: We have results from application to multiple data sets (published and submitted) and multiple eQTL data sets. For example, application of SMR using 5967 eQTL discovered in whole blood (5311 samples) and applied to the schizophrenia GWAS results we identified 34 genes that were significant after correcting for multiple testing, and of these an additional test provides confidence that 16 reflect a single causal association. Of these only two genes were the nearest gene to the SNP association from the GWAS study. Additional follow-up SMR analyses were conducted for six genes using eQTL from brain. Two novel genes were identified as associated with risk of schizophrenia, SNX19 and NMRAL1.
Virginia Commonwealth University University of Melbourne
Background: Mates are similar to one another across a large number of traits, including liabilities underlying psychiatric disorders. It has long been known that when such phenotypic similarity implies genetic similarity (hereafter, assortative mating), heritability estimates from family and twin studies can be biased. However, the effects of assortative mating, if any, on estimates of heritability from all SNPs in unrelated individuals (SNP-based heritability, h2SNP) have not been investigated to date. Methods: We derived mathematically the expected behavior of estimates from Haseman-Elston regression and GREML, and assessed these predictions via simulation. We then assessed in the UK Biobank whether the predicted signatures of assortative mating on h2SNP were observed. Results: We show analytically and via simulation that estimates of h2SNP from both GREML and Haseman-Elston regression are typically biased upwards in the presence of assortative mating. However, estimates from Haseman-Elston regression are roughly constant as a function of sample size and number of markers, whereas the ratio of sample size to number of markers strongly affects those from GREML. This difference in estimate behavior allows one to assess in real data whether assortative mating is affecting estimates of h2SNP. When we did this in the UK Biobank for height and IQ, we observed that h2SNP estimates for height that behaved as predicted, but not for h2SNP estimates for IQ. Discussion: For many traits, the degree of bias in estimates of h2SNP is expected to be small, but it can be nontrivial for other traits, depending on the degree of assortative mating and the heritability of the traits. For example, based on spousal similarity, we predict that h2SNP estimates of height and IQ are 25% greater than their true equilibrium values. That we did not observe the predicted pattern of h2SNP estimates for IQ may either suggest that spousal similarity for IQ occurs via mechanisms that do not lead to genetic similarity (e.g., social homogamy) or that the effects of assortative mating on the covariance between causal variants is counter-acted by the forces of natural selection.
S789
relatively common in the population. Different mechanisms have been proposed to explain this so-called ‘evolutionary paradox’, including balancing selection, the mismatch hypothesis and the polygenic mutation-selection balance. Evolutionary analyses of individual genetic loci can be informative for monogenic diseases associated with large selection pressures, but have less utility in the analysis of complex, polygenic disorders due to associations of many low-risk alleles. Taking a polygenic approach to mental disorders, we explore the impact of evolutionary pressure on allele frequency under the hypotheses that lower-thanexpected average frequencies of disease risk alleles are indicative of negative selection, while higher-thanexpected frequencies are indicative of positive selection. Methods: Using the largest available GWAS summary statistics for psychiatric and other disorders and traits, we identified allele frequencies for the disease risk or traitincreasing alleles for all SNPs genome-wide. The SNPs were thinned according to P-value via clumping, using stringent linkage disequilibrium criteria, to ensure their independence. After ordering the clumped SNPs by association P-value or effect size in the GWAS, we tested whether mean trait-increasing allele frequencies differed from 50%, as expected by chance, and investigated whether the effect was greater for those SNPs with largest association signal. Results: We found evidence of negative selection on Schizophrenia, Crohn's disease and obesity (BMI) risk alleles and of positive selection for height, education and ADHD alleles. The direction of evolutionary pressure was inconclusive for other traits tested including bipolar disorder, depression and autism. For each trait, we estimated the number of SNPs that maximised the mean trait-increasing allele frequency deviation from 50%. Discussion: Our results suggest that evolutionary pressure has had a small but marked impact on the allele frequencies of thousands of SNPs of certain psychiatric disorders and other traits. Unlike measures of fecundity in population cohorts, which are restricted by current environmental pressures, social trends, and changes in reproductive behaviour, evidence of positive or negative selection on the genome provides insights into our understanding of the impact of disease on the population over generations.
Disclosure: Nothing to disclose. Disclosure: Nothing to disclose. http://dx.doi.org/10.1016/j.euroneuro.2017.08.017 http://dx.doi.org/10.1016/j.euroneuro.2017.08.016
16. MEASURING EVOLUTIONARY PRESSURE IN A POLYGENIC FRAMEWORK FOR COMPLEX DISEASES
17. FUMA: FUNCTIONAL MAPPING AND ANNOTATION OF GENETIC ASSOCIATIONS n
Evangelos Vassos , Paul O'Reilly, Cathryn Lewis
Kyoko Watanabe , Erdogan Taskesen, Arjen van Bochoven, Danielle Posthuma
King's College London
VU Amsterdam
Background: Although many mental disorders are herita-
Background: A main challenge in Genome-Wide Association Studies (GWAS) is to prioritize genetic variants and
n
ble and result in reproductive disadvantage, they remain
S790 identify potential causal mechanisms of human diseases. Although a variety of bioinformatics tools are now available for downstream analyses of GWAS results, a standard, integrative approach is lacking. We developed FUMA: a web-based platform to facilitate functional annotation of GWAS results, prioritization of potential causal genetic variants and genes, and interactive visualization by incorporating 14 biological data repositories and tools. FUMA is available as an online tool at http://fuma.ctglab.nl. Methods: FUMA contains two core functions to annotate input summary statistics to prioritize potential causal genetic variants and genes; SNP2GENE and GENE2FUNC. In SNP2GENE, SNPs are annotated with their biological functionality and mapped to genes based on positional and functional information of SNPs. In this process, significantly associated SNPs from GWAS are characterized as genomic risk loci by incorporating the linkage disequilibrium structure. Functionally annotated SNPs are mapped to genes based on functional consequences on genes (positional mapping), expression quantitative trait loci (eQTLs) and chromatin interactions of phenotype relevant tissue types (eQTL and chromatin interaction mappings). By combining these three mapping strategies, FUMA enables to prioritize genes that are highly likely involved in the trait of interest. To obtain insight into putative causal mechanisms, the GENE2FUNC process annotates the prioritized genes in biological context, such as tissue specific gene expression pattern, enrichment of gene sets and direct links to well defined external biological databases such as disease associated genes and drug targets. Results: We have applied FUMA to the most recent Body Mass Index (BMI)[1]. We successfully prioritized previously reported genes from the original GWAS study, and also novel candidates by combining positional mapping of deleterious coding SNPs and eQTL mapping. For example, IRX3 was prioritized by eQTLs from FTO locus which is the most significantly associated locus with BMI. Although, FTO is the only gene reported in the original study since the genomic risk locus is located within the FTO gene and not overlapped with any other genes, IRX3 has been recently validated as whose expression is affected by variants in the FTO locus[2]. Thus, by incorporating multiple biological information, we could prioritise genes that are located outside of genomic risk loci. Chromatin interaction mapping identified additional novel candidates, including genes located outside of the risk loci, which showed share biological functions with the previously reported genes. Discussion: In summary, FUMA provides an easy-to-use tool to interpret and functionally annotate results from genetic association studies. It allows to quickly gain insight into the directional biological implications of significant genetic associations. FUMA combines information of stateof-the-art biological data sources in a single platform to facilitate the generation of hypotheses for functional follow-up analysis aimed at proving causal relations between genetic variants and diseases.
Y. Ma, M. Li
Disclosure: Nothing to disclose.
References [1] A. Locke et al., “Genetic studies of body mass index yield new insights for obesity biology”. Nature 518 (7538):197–206, 2015. [2] M. Claussnitzer et al., “FTO obesity variant circuitry and adipocyte browning in humans”. The New England of Medicine 373(10):895–907, 2015. http://dx.doi.org/10.1016/j.euroneuro.2017.08.018
18. VALIDATING GWA ASSOCIATIONS IN PSYCHIATRIC DISORDERS WITH FUNCTIONAL GENOMIC DATA n
Maciej Trzaskowski ,1, Enda Byrne2, Jian Yang1, Naomi Wray1 1 2
University of Queensland Queensland Brain Institute
Background: Genome-wide association studies have been successful at identifying regions of the genome associated with disease. Often, the identified SNPs are in regions of linkage disequilibrium that stretch over large distances, making it difficult to ascertain which gene in the region is functionally relevant to the trait of interest. For disorders of the brain, assessing the functional relevance of SNPs in the same sample in which they were discovered is, of course, difficult. Methods: Here we applied the SMR (Summary-data based Mendelian Randomisation) method to the latest GWAS results of psychiatric disorders, to evaluate whether SNPs associated with these disorders are also associated with control of gene expression of genes in the same region. The advantage of this method is that it utilises GWAS and expression QTL (eQTL) data from different samples. eQTL data were taken from whole blood and brain samples. Results: We have results from application to multiple data sets (published and submitted) and multiple eQTL data sets. For example, application of SMR using 5967 eQTL discovered in whole blood (5311 samples) and applied to the schizophrenia GWAS results we identified 34 genes that were significant after correcting for multiple testing, and of these an additional test provides confidence that 16 reflect a single causal association. Of these only two genes were the nearest gene to the SNP association from the GWAS study. Additional follow-up SMR analyses were conducted for six genes using eQTL from brain. Two novel genes were identified as associated with risk of schizophrenia, SNX19 and NMRAL1.