Efficient Bayesian mixed-model analysis increases association power in large cohorts

A hybrid expectation maximisation and MCMC sampling algorithm to implement Bayesian mixture model based genomic prediction and QTL mapping

Background

Bayesian mixture models in which the effects of SNP are assumed to come from normal distributions with different variances are attractive for simultaneous genomic prediction and QTL mapping. These models are usually implemented with Monte Carlo Markov Chain (MCMC) sampling, which requires long compute times with large genomic data sets. Here, we present an efficient approach (termed HyB_BR), which is a hybrid of an Expectation-Maximisation algorithm, followed by a limited number of MCMC without the requirement for burn-in.

Results

To test prediction accuracy from HyB_BR, dairy cattle and human disease trait data were used. In the dairy cattle data, there were four quantitative traits (milk volume, protein kg, fat% in milk and fertility) measured in 16,214 cattle from two breeds genotyped for 632,002 SNPs. Validation of genomic predictions was in a subset of cattle either from the reference set or in animals from a third breeds that were not in the reference set. In all cases, HyB_BR gave almost identical accuracies to Bayesian mixture models implemented with full MCMC, however computational time was reduced by up to 1/17 of that required by full MCMC. The SNPs with high posterior probability of a non-zero effect were also very similar between full MCMC and HyB_BR, with several known genes affecting milk production in this category, as well as some novel genes. HyB_BR was also applied to seven human diseases with 4890 individuals genotyped for around 300 K SNPs in a case/control design, from the Welcome Trust Case Control Consortium (WTCCC). In this data set, the results demonstrated again that HyB_BR performed as well as Bayesian mixture models with full MCMC for genomic predictions and genetic architecture inference while reducing the computational time from 45 h with full MCMC to 3 h with HyB_BR.

Conclusions

The results for quantitative traits in cattle and disease in humans demonstrate that HyB_BR can perform equally well as Bayesian mixture models implemented with full MCMC in terms of prediction accuracy, but with up to 17 times faster than the full MCMC implementations. The HyB_BR algorithm makes simultaneous genomic prediction, QTL mapping and inference of genetic architecture feasible in large genomic data sets.

Electronic supplementary material

The online version of this article (doi:10.1186/s12864-016-3082-7) contains supplementary material, which is available to authorized users.

Genome-Wide Association Analyses in 128,266 Individuals Identifies New Morningness and Sleep Duration Loci

Disrupted circadian rhythms and reduced sleep duration are associated with several human diseases, particularly obesity and type 2 diabetes, but until recently, little was known about the genetic factors influencing these heritable traits. We performed genome-wide association studies of self-reported chronotype (morning/evening person) and self-reported sleep duration in 128,266 white British individuals from the UK Biobank study. Sixteen variants were associated with chronotype (P<5x10^-8), including variants near the known circadian rhythm genes RGS16 (1.21 odds of morningness, 95% CI [1.15, 1.27], P = 3x10^-12) and PER2 (1.09 odds of morningness, 95% CI [1.06, 1.12], P = 4x10^-10). The PER2 signal has previously been associated with iris function. We sought replication using self-reported data from 89,283 23andMe participants; thirteen of the chronotype signals remained associated at P<5x10^-8 on meta-analysis and eleven of these reached P<0.05 in the same direction in the 23andMe study. We also replicated 9 additional variants identified when the 23andMe study was used as a discovery GWAS of chronotype (all P<0.05 and meta-analysis P<5x10^-8). For sleep duration, we replicated one known signal in PAX8 (2.6 minutes per allele, 95% CI [1.9, 3.2], P = 5.7x10^-16) and identified and replicated two novel associations at VRK2 (2.0 minutes per allele, 95% CI [1.3, 2.7], P = 1.2x10^-9; and 1.6 minutes per allele, 95% CI [1.1, 2.2], P = 7.6x10^-9). Although we found genetic correlation between chronotype and BMI (rG = 0.056, P = 0.05); undersleeping and BMI (rG = 0.147, P = 1x10^-5) and oversleeping and BMI (rG = 0.097, P = 0.04), Mendelian Randomisation analyses, with limited power, provided no consistent evidence of causal associations between BMI or type 2 diabetes and chronotype or sleep duration. Our study brings the total number of loci associated with chronotype to 22 and with sleep duration to three, and provides new insights into the biology of sleep and circadian rhythms in humans.

Numerous studies have identified links between too little or too much sleep and circadian misalignment with metabolic disorders such as obesity and type 2 diabetes. However, cause-and-effect is not easily determined, because of multiple confounding factors affecting both sleep patterns and disease risk. Using the first release of the UK Biobank study, which combines detailed measurements and questionnaire data with genetic data, we investigate the genetics of two self-report sleep measures, chronotype and average sleep duration, in 128,266 white British individuals. We replicate previous genetic associations and identify seven and two novel genetic variants influencing chronotype and sleep duration, respectively. Associated variants are located near genes implicated in circadian rhythm regulation (RGS16, PER2), near a serotonin receptor gene (HTR6) and another gene (INADL) encoding a protein thought to be important in photosensitive retinal cells, cells known to communicate with the brain’s primary circadian pacemaker. Using the genetic risk factors, we estimate the unconfounded causal associations of BMI and type 2 diabetes on sleep patterns (and vice versa) through Mendelian Randomisation. However, we find no evidence for causal associations in either direction. The full UK Biobank release of 500,000 individuals will boost our power to detect causal associations.

Transethnic Genetic-Correlation Estimates from Summary Statistics

The increasing number of genetic association studies conducted in multiple populations provides an unprecedented opportunity to study how the genetic architecture of complex phenotypes varies between populations, a problem important for both medical and population genetics. Here, we have developed a method for estimating the transethnic genetic correlation: the correlation of causal-variant effect sizes at SNPs common in populations. This methods takes advantage of the entire spectrum of SNP associations and uses only summary-level data from genome-wide association studies. This avoids the computational costs and privacy concerns associated with genotype-level information while remaining scalable to hundreds of thousands of individuals and millions of SNPs. We applied our method to data on gene expression, rheumatoid arthritis, and type 2 diabetes and overwhelmingly found that the genetic correlation was significantly less than 1. Our method is implemented in a Python package called Popcorn.

Enabling Privacy-Preserving GWASs in Heterogeneous Human Populations

The proliferation of large genomic databases offers the potential to perform increasingly larger-scale genome-wide association studies (GWASs). Due to privacy concerns, however, access to these data is limited, greatly reducing their usefulness for research. Here, we introduce a computational framework for performing GWASs that adapts principles of differential privacy-a cryptographic theory that facilitates secure analysis of sensitive data-to both protect private phenotype information (e.g., disease status) and correct for population stratification. This framework enables us to produce privacy-preserving GWAS results based on EIGENSTRAT and linear mixed model (LMM)-based statistics, both of which correct for population stratification. We test our differentially private statistics, PrivSTRAT and PrivLMM, on simulated and real GWAS datasets and find they are able to protect privacy while returning meaningful results. Our framework can be used to securely query private genomic datasets to discover which specific genomic alterations may be associated with a disease, thus increasing the availability of these valuable datasets.

Physical and neurobehavioral determinants of reproductive onset and success

The ages of puberty, first sexual intercourse and first birth signify the onset of reproductive ability, behavior and success, respectively. In a genome-wide association study of 125,667 UK Biobank participants, we identify 38 loci associated (P < 5 × 10(-8)) with age at first sexual intercourse. These findings were taken forward in 241,910 men and women from Iceland and 20,187 women from the Women's Genome Health Study. Several of the identified loci also exhibit associations (P < 5 × 10(-8)) with other reproductive and behavioral traits, including age at first birth (variants in or near ESR1 and RBM6-SEMA3F), number of children (CADM2 and ESR1), irritable temperament (MSRA) and risk-taking propensity (CADM2). Mendelian randomization analyses infer causal influences of earlier puberty timing on earlier first sexual intercourse, earlier first birth and lower educational attainment. In turn, likely causal consequences of earlier first sexual intercourse include reproductive, educational, psychiatric and cardiometabolic outcomes.

Strategies for Imputing and Analyzing Rare Variants in Association Studies

Rare genetic variants may be responsible for a significant amount of the uncharacterized genetic risk underlying many diseases. An efficient approach to characterizing the disease burden of rare variants may be to impute them into existing large datasets. It is well known that the ability to impute a rare variant is dependent both on the array choice and number of individuals in the reference panel carrying that variant, although it is still unclear exactly how well imputation will work for rare variants. Here, we review the additional challenges that arise when imputing rare variants, looking at studies that have been able to impute rare variants, methods behind merging reference panels, approaches for imputing rare variants, and methods for analyzing rare variants.

Control for Population Structure and Relatedness for Binary Traits in Genetic Association Studies via Logistic Mixed Models

Linear mixed models (LMMs) are widely used in genome-wide association studies (GWASs) to account for population structure and relatedness, for both continuous and binary traits. Motivated by the failure of LMMs to control type I errors in a GWAS of asthma, a binary trait, we show that LMMs are generally inappropriate for analyzing binary traits when population stratification leads to violation of the LMM's constant-residual variance assumption. To overcome this problem, we develop a computationally efficient logistic mixed model approach for genome-wide analysis of binary traits, the generalized linear mixed model association test (GMMAT). This approach fits a logistic mixed model once per GWAS and performs score tests under the null hypothesis of no association between a binary trait and individual genetic variants. We show in simulation studies and real data analysis that GMMAT effectively controls for population structure and relatedness when analyzing binary traits in a wide variety of study designs.

Multiple testing correction in linear mixed models

BACKGROUND

Multiple hypothesis testing is a major issue in genome-wide association studies (GWAS), which often analyze millions of markers. The permutation test is considered to be the gold standard in multiple testing correction as it accurately takes into account the correlation structure of the genome. Recently, the linear mixed model (LMM) has become the standard practice in GWAS, addressing issues of population structure and insufficient power. However, none of the current multiple testing approaches are applicable to LMM.

RESULTS

We were able to estimate per-marker thresholds as accurately as the gold standard approach in real and simulated datasets, while reducing the time required from months to hours. We applied our approach to mouse, yeast, and human datasets to demonstrate the accuracy and efficiency of our approach.

CONCLUSIONS

We provide an efficient and accurate multiple testing correction approach for linear mixed models. We further provide an intuition about the relationships between per-marker threshold, genetic relatedness, and heritability, based on our observations in real data.

Genome-wide association analysis identifies novel loci for chronotype in 100,420 individuals from the UK Biobank

Our sleep timing preference, or chronotype, is a manifestation of our internal biological clock. Variation in chronotype has been linked to sleep disorders, cognitive and physical performance, and chronic disease. Here we perform a genome-wide association study of self-reported chronotype within the UK Biobank cohort (n=100,420). We identify 12 new genetic loci that implicate known components of the circadian clock machinery and point to previously unstudied genetic variants and candidate genes that might modulate core circadian rhythms or light-sensing pathways. Pathway analyses highlight central nervous and ocular systems and fear-response-related processes. Genetic correlation analysis suggests chronotype shares underlying genetic pathways with schizophrenia, educational attainment and possibly BMI. Further, Mendelian randomization suggests that evening chronotype relates to higher educational attainment. These results not only expand our knowledge of the circadian system in humans but also expose the influence of circadian characteristics over human health and life-history variables such as educational attainment.

Height, body mass index, and socioeconomic status: mendelian randomisation study in UK Biobank

OBJECTIVE

To determine whether height and body mass index (BMI) have a causal role in five measures of socioeconomic status.

DESIGN

Mendelian randomisation study to test for causal effects of differences in stature and BMI on five measures of socioeconomic status. Mendelian randomisation exploits the fact that genotypes are randomly assigned at conception and thus not confounded by non-genetic factors.

SETTING

UK Biobank.

PARTICIPANTS

119,669 men and women of British ancestry, aged between 37 and 73 years.

MAIN OUTCOME MEASURES

Age completed full time education, degree level education, job class, annual household income, and Townsend deprivation index.

RESULTS

In the UK Biobank study, shorter stature and higher BMI were observationally associated with several measures of lower socioeconomic status. The associations between shorter stature and lower socioeconomic status tended to be stronger in men, and the associations between higher BMI and lower socioeconomic status tended to be stronger in women. For example, a 1 standard deviation (SD) higher BMI was associated with a £210 (€276; $300; 95% confidence interval £84 to £420; P=6 × 10(-3)) lower annual household income in men and a £1890 (£1680 to £2100; P=6 × 10(-15)) lower annual household income in women. Genetic analysis provided evidence that these associations were partly causal. A genetically determined 1 SD (6.3 cm) taller stature caused a 0.06 (0.02 to 0.09) year older age of completing full time education (P=0.01), a 1.12 (1.07 to 1.18) times higher odds of working in a skilled profession (P=6 × 10(-7)), and a £1130 (£680 to £1580) higher annual household income (P=4 × 10(-8)). Associations were stronger in men. A genetically determined 1 SD higher BMI (4.6 kg/m(2)) caused a £2940 (£1680 to £4200; P=1 × 10(-5)) lower annual household income and a 0.10 (0.04 to 0.16) SD (P=0.001) higher level of deprivation in women only.

CONCLUSIONS

These data support evidence that height and BMI play an important partial role in determining several aspects of a person's socioeconomic status, especially women's BMI for income and deprivation and men's height for education, income, and job class. These findings have important social and health implications, supporting evidence that overweight people, especially women, are at a disadvantage and that taller people, especially men, are at an advantage.

Fast Principal-Component Analysis Reveals Convergent Evolution of ADH1B in Europe and East Asia

Searching for genetic variants with unusual differentiation between subpopulations is an established approach for identifying signals of natural selection. However, existing methods generally require discrete subpopulations. We introduce a method that infers selection using principal components (PCs) by identifying variants whose differentiation along top PCs is significantly greater than the null distribution of genetic drift. To enable the application of this method to large datasets, we developed the FastPCA software, which employs recent advances in random matrix theory to accurately approximate top PCs while reducing time and memory cost from quadratic to linear in the number of individuals, a computational improvement of many orders of magnitude. We apply FastPCA to a cohort of 54,734 European Americans, identifying 5 distinct subpopulations spanning the top 4 PCs. Using the PC-based test for natural selection, we replicate previously known selected loci and identify three new genome-wide significant signals of selection, including selection in Europeans at ADH1B. The coding variant rs1229984(∗)T has previously been associated to a decreased risk of alcoholism and shown to be under selection in East Asians; we show that it is a rare example of independent evolution on two continents. We also detect selection signals at IGFBP3 and IGH, which have also previously been associated to human disease.

Human longevity is influenced by many genetic variants: evidence from 75,000 UK Biobank participants

Variation in human lifespan is 20 to 30% heritable in twins but few genetic variants have been identified. We undertook a Genome Wide Association Study (GWAS) using age at death of parents of middle-aged UK Biobank participants of European decent (n=75,244 with father's and/or mother's data, excluding early deaths). Genetic risk scores for 19 phenotypes (n=777 proven variants) were also tested. In GWAS, a nicotine receptor locus(CHRNA3, previously associated with increased smoking and lung cancer) was associated with fathers' survival. Less common variants requiring further confirmation were also identified. Offspring of longer lived parents had more protective alleles for coronary artery disease, systolic blood pressure, body mass index, cholesterol and triglyceride levels, type-1 diabetes, inflammatory bowel disease and Alzheimer's disease. In candidate analyses, variants in the TOMM40/APOE locus were associated with longevity, but FOXO variants were not. Associations between extreme longevity (mother >=98 years, fathers >=95 years, n=1,339) and disease alleles were similar, with an additional association with HDL cholesterol (p=5.7x10-3). These results support a multiple protective factors model influencing lifespan and longevity (top 1% survival) in humans, with prominent roles for cardiovascular-related pathways. Several of these genetically influenced risks, including blood pressure and tobacco exposure, are potentially modifiable.

A multi-trait meta-analysis with imputed sequence variants reveals twelve QTL for mammary gland morphology in Fleckvieh cattle

Background

The availability of whole-genome sequence data from key ancestors in bovine populations provides an exhaustive catalogue of polymorphic sites that segregate within and across cattle breeds. Sequence variants identified from the sequenced genome of key ancestors can be imputed into animals that have been genotyped using medium- and high-density genotyping arrays. Association analysis with imputed sequences, particularly when applied to multiple traits simultaneously, is a very powerful approach to detect candidate causal variants that underlie complex phenotypes.

Results

We used whole-genome sequence data from 157 key ancestors of the German Fleckvieh cattle population to impute 20,561,798 sequence variants into 10,363 animals that had (partly imputed) genotypes based on 634,109 single nucleotide polymorphisms (SNPs). Rare variants were more frequent among the sequence-derived than the array-derived genotypes. Association studies with imputed sequence variants were performed using seven correlated udder conformation traits as response variables. The calculation of an approximate multi-trait test statistic enabled us to detect 12 quantitative trait loci (QTL) (P < 2.97 × 10⁻⁹) that affect different morphological features of the mammary gland. Among the tested variants, the most significant associations were found for imputed sequence variants at 11 QTL, whereas the top association signal was observed for an array-derived variant at a QTL on bovine chromosome 14. Seven QTL were associated with multiple phenotypes. Most QTL were located in non-coding regions of the genome but in close proximity of candidate genes that could be involved in mammary gland morphology (SP5, GC, NPFFR2, CRIM1, RXFP2, TBX5, RBM19 and ADAM12).

Conclusions

Using imputed sequence variants in association analyses allows the detection of QTL at maximum resolution. Multi-trait approaches can reveal QTL that are not detected in single-trait association studies. Most QTL for udder conformation traits were located in non-coding regions of the genome, which suggests that mutations in regulatory sequences are the major determinants of variation in mammary gland morphology in cattle.

Electronic supplementary material

The online version of this article (doi:10.1186/s12711-016-0190-4) contains supplementary material, which is available to authorized users.

Iterative Usage of Fixed and Random Effect Models for Powerful and Efficient Genome-Wide Association Studies

False positives in a Genome-Wide Association Study (GWAS) can be effectively controlled by a fixed effect and random effect Mixed Linear Model (MLM) that incorporates population structure and kinship among individuals to adjust association tests on markers; however, the adjustment also compromises true positives. The modified MLM method, Multiple Loci Linear Mixed Model (MLMM), incorporates multiple markers simultaneously as covariates in a stepwise MLM to partially remove the confounding between testing markers and kinship. To completely eliminate the confounding, we divided MLMM into two parts: Fixed Effect Model (FEM) and a Random Effect Model (REM) and use them iteratively. FEM contains testing markers, one at a time, and multiple associated markers as covariates to control false positives. To avoid model over-fitting problem in FEM, the associated markers are estimated in REM by using them to define kinship. The P values of testing markers and the associated markers are unified at each iteration. We named the new method as Fixed and random model Circulating Probability Unification (FarmCPU). Both real and simulated data analyses demonstrated that FarmCPU improves statistical power compared to current methods. Additional benefits include an efficient computing time that is linear to both number of individuals and number of markers. Now, a dataset with half million individuals and half million markers can be analyzed within three days.

Genetic evidence that lower circulating FSH levels lengthen menstrual cycle, increase age at menopause and impact female reproductive health

STUDY QUESTION

How does a genetic variant in the FSHB promoter, known to alter FSH levels, impact female reproductive health?

SUMMARY ANSWER

The T allele of the FSHB promoter polymorphism (rs10835638; c.-211G>T) results in longer menstrual cycles and later menopause and, while having detrimental effects on fertility, is protective against endometriosis.

WHAT IS KNOWN ALREADY

The FSHB promoter polymorphism (rs10835638; c.-211G>T) affects levels of FSHB transcription and, as a result, circulating levels of FSH. FSH is required for normal fertility and genetic variants at the FSHB locus are associated with age at menopause and polycystic ovary syndrome (PCOS).

STUDY DESIGN, SIZE, DURATION

We used cross-sectional data from the UK Biobank to look at associations between the FSHB promoter polymorphism and reproductive traits, and performed a genome-wide association study (GWAS) for length of menstrual cycle.

PARTICIPANTS/MATERIALS, SETTING, METHODS

We included white British individuals aged 40-69 years in 2006-2010, in the May 2015 release of genetic data from UK Biobank. We tested the FSH-lowering T allele of the FSHB promoter polymorphism (rs10835638; c.-211G>T) for associations with 29, mainly female, reproductive phenotypes in up to 63 350 women and 56 608 men. We conducted a GWAS in 9534 individuals to identify genetic variants associated with length of menstrual cycle.

MAIN RESULTS AND THE ROLE OF CHANCE

The FSH-lowering T allele of the FSHB promoter polymorphism (rs10835638; MAF 0.16) was associated with longer menstrual cycles [0.16 SD (c. 1 day) per minor allele; 95% confidence interval (CI) 0.12-0.20; P = 6 × 10(-16)], later age at menopause (0.13 years per minor allele; 95% CI 0.04-0.22; P = 5.7 × 10(-3)), greater female nulliparity [odds ratio (OR) = 1.06; 95% CI 1.02-1.11; P = 4.8 × 10(-3)] and lower risk of endometriosis (OR = 0.79; 95% CI 0.69-0.90; P = 4.1 × 10(-4)). The FSH-lowering T allele was not associated with other female reproductive illnesses or conditions in our study and we did not replicate associations with male infertility or PCOS. In the GWAS for menstrual cycle length, only variants near the FSHB gene reached genome-wide significance (P < 5 × 10(-9)).

LIMITATIONS, REASONS FOR CAUTION

The data included might be affected by recall bias. Cycle length was not available for 25% of women still cycling (1% did not answer, 6% did not know and for 18% cycle length was recorded as 'irregular'). Women with a cycle length recorded were aged over 40 and were approaching menopause; however, we did not find evidence that this affected the results. Many of the groups with illnesses had relatively small sample sizes and so the study may have been under-powered to detect an effect.

WIDER IMPLICATIONS OF THE FINDINGS

We found a strong novel association between a genetic variant that lowers FSH levels and longer menstrual cycles, at a locus previously robustly associated with age at menopause. The variant was also associated with nulliparity and endometriosis risk. These findings should now be verified in a second independent group of patients. We conclude that lifetime differences in circulating levels of FSH between individuals can influence menstrual cycle length and a range of reproductive outcomes, including menopause timing, infertility, endometriosis and PCOS.

STUDY FUNDING/COMPETING INTERESTS

None.

TRIAL REGISTRATION NUMBER

Not applicable.

A new tool called DISSECT for analysing large genomic data sets using a Big Data approach

Large-scale genetic and genomic data are increasingly available and the major bottleneck in their analysis is a lack of sufficiently scalable computational tools. To address this problem in the context of complex traits analysis, we present DISSECT. DISSECT is a new and freely available software that is able to exploit the distributed-memory parallel computational architectures of compute clusters, to perform a wide range of genomic and epidemiologic analyses, which currently can only be carried out on reduced sample sizes or under restricted conditions. We demonstrate the usefulness of our new tool by addressing the challenge of predicting phenotypes from genotype data in human populations using mixed-linear model analysis. We analyse simulated traits from 470,000 individuals genotyped for 590,004 SNPs in ∼4 h using the combined computational power of 8,400 processor cores. We find that prediction accuracies in excess of 80% of the theoretical maximum could be achieved with large sample sizes.

Availability of computing power can limit computational analysis of large genetic and genomic datasets. Here, Canela-Xandri, et al. describe a software called DISSECT that is capable of analyzing large-scale genetic data by distributing the work across thousands of networked computers.

Contrasting genetic architectures of schizophrenia and other complex diseases using fast variance-components analysis

Heritability analyses of genome-wide association study (GWAS) cohorts have yielded important insights into complex disease architecture, and increasing sample sizes hold the promise of further discoveries. Here we analyze the genetic architectures of schizophrenia in 49,806 samples from the PGC and nine complex diseases in 54,734 samples from the GERA cohort. For schizophrenia, we infer an overwhelmingly polygenic disease architecture in which ≥71% of 1-Mb genomic regions harbor ≥1 variant influencing schizophrenia risk. We also observe significant enrichment of heritability in GC-rich regions and in higher-frequency SNPs for both schizophrenia and GERA diseases. In bivariate analyses, we observe significant genetic correlations (ranging from 0.18 to 0.85) for several pairs of GERA diseases; genetic correlations were on average 1.3 tunes stronger than the correlations of overall disease liabilities. To accomplish these analyses, we developed a fast algorithm for multicomponent, multi-trait variance-components analysis that overcomes prior computational barriers that made such analyses intractable at this scale.

Two-Variance-Component Model Improves Genetic Prediction in Family Datasets

Genetic prediction based on either identity by state (IBS) sharing or pedigree information has been investigated extensively with best linear unbiased prediction (BLUP) methods. Such methods were pioneered in plant and animal-breeding literature and have since been applied to predict human traits, with the aim of eventual clinical utility. However, methods to combine IBS sharing and pedigree information for genetic prediction in humans have not been explored. We introduce a two-variance-component model for genetic prediction: one component for IBS sharing and one for approximate pedigree structure, both estimated with genetic markers. In simulations using real genotypes from the Candidate-gene Association Resource (CARe) and Framingham Heart Study (FHS) family cohorts, we demonstrate that the two-variance-component model achieves gains in prediction r(2) over standard BLUP at current sample sizes, and we project, based on simulations, that these gains will continue to hold at larger sample sizes. Accordingly, in analyses of four quantitative phenotypes from CARe and two quantitative phenotypes from FHS, the two-variance-component model significantly improves prediction r(2) in each case, with up to a 20% relative improvement. We also find that standard mixed-model association tests can produce inflated test statistics in datasets with related individuals, whereas the two-variance-component model corrects for inflation.

PSIKO2: a fast and versatile tool to infer population stratification on various levels in GWAS

Genome-wide association studies are an invaluable tool for identifying genotypic loci linked with agriculturally important traits or certain diseases. The signal on which such studies rely upon can, however, be obscured by population stratification making it necessary to account for it in some way. Population stratification is dependent on when admixture happened and thus can occur at various levels. To aid in its inference at the genome level, we recently introduced psiko, and comparison with leading methods indicates that it has attractive properties. However, until now, it could not be used for local ancestry inference which is preferable in cases of recent admixture as the genome level tends to be too coarse to properly account for processes acting on small segments of a genome. To also bring the powerful ideas underpinning psiko to bear in such studies, we extended it to psiko2, which we introduce here.

AVAILABILITY AND IMPLEMENTATION

Source code, binaries and user manual are freely available at https://www.uea.ac.uk/computing/psiko.

CONTACT

Andrei-Alin.Popescu@uea.ac.uk or Katharina.Huber@cmp.uea.ac.uk

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Phenotypic variance explained by local ancestry in admixed African Americans

We surveyed 26 quantitative traits and disease outcomes to understand the proportion of phenotypic variance explained by local ancestry in admixed African Americans. After inferring local ancestry as the number of African-ancestry chromosomes at hundreds of thousands of genotyped loci across all autosomes, we used a linear mixed effects model to estimate the variance explained by local ancestry in two large independent samples of unrelated African Americans. We found that local ancestry at major and polygenic effect genes can explain up to 20 and 8% of phenotypic variance, respectively. These findings provide evidence that most but not all additive genetic variance is explained by genetic markers undifferentiated by ancestry. These results also inform the proportion of health disparities due to genetic risk factors and the magnitude of error in association studies not controlling for local ancestry.

Quality control, imputation and analysis of genome-wide genotyping data from the Illumina HumanCoreExome microarray

The decreasing cost of performing genome-wide association studies has made genomics widely accessible. However, there is a paucity of guidance for best practice in conducting such analyses. For the results of a study to be valid and replicable, multiple biases must be addressed in the course of data preparation and analysis. In addition, standardizing methods across small, independent studies would increase comparability and the potential for effective meta-analysis. This article provides a discussion of important aspects of quality control, imputation and analysis of genome-wide data from a low-coverage microarray, as well as a straight-forward guide to performing a genome-wide association study. A detailed protocol is provided online, with example scripts available at https://github.com/JoniColeman/gwas_scripts.

Modeling Linkage Disequilibrium Increases Accuracy of Polygenic Risk Scores

Polygenic risk scores have shown great promise in predicting complex disease risk and will become more accurate as training sample sizes increase. The standard approach for calculating risk scores involves linkage disequilibrium (LD)-based marker pruning and applying a p value threshold to association statistics, but this discards information and can reduce predictive accuracy. We introduce LDpred, a method that infers the posterior mean effect size of each marker by using a prior on effect sizes and LD information from an external reference panel. Theory and simulations show that LDpred outperforms the approach of pruning followed by thresholding, particularly at large sample sizes. Accordingly, predicted R(2) increased from 20.1% to 25.3% in a large schizophrenia dataset and from 9.8% to 12.0% in a large multiple sclerosis dataset. A similar relative improvement in accuracy was observed for three additional large disease datasets and for non-European schizophrenia samples. The advantage of LDpred over existing methods will grow as sample sizes increase.

Efficient set tests for the genetic analysis of correlated traits

Set tests are a powerful approach for genome-wide association testing between groups of genetic variants and quantitative traits. We describe mtSet (http://github.com/PMBio/limix), a mixed-model approach that enables joint analysis across multiple correlated traits while accounting for population structure and relatedness. mtSet effectively combines the benefits of set tests with multi-trait modeling and is computationally efficient, enabling genetic analysis of large cohorts (up to 500,000 individuals) and multiple traits.

Mixed model with correction for case-control ascertainment increases association power

We introduce a liability-threshold mixed linear model (LTMLM) association statistic for case-control studies and show that it has a well-controlled false-positive rate and more power than existing mixed-model methods for diseases with low prevalence. Existing mixed-model methods suffer a loss in power under case-control ascertainment, but no solution has been proposed. Here, we solve this problem by using a χ(2) score statistic computed from posterior mean liabilities (PMLs) under the liability-threshold model. Each individual's PML is conditional not only on that individual's case-control status but also on every individual's case-control status and the genetic relationship matrix (GRM) obtained from the data. The PMLs are estimated with a multivariate Gibbs sampler; the liability-scale phenotypic covariance matrix is based on the GRM, and a heritability parameter is estimated via Haseman-Elston regression on case-control phenotypes and then transformed to the liability scale. In simulations of unrelated individuals, the LTMLM statistic was correctly calibrated and achieved higher power than existing mixed-model methods for diseases with low prevalence, and the magnitude of the improvement depended on sample size and severity of case-control ascertainment. In a Wellcome Trust Case Control Consortium 2 multiple sclerosis dataset with >10,000 samples, LTMLM was correctly calibrated and attained a 4.3% improvement (p = 0.005) in χ(2) statistics over existing mixed-model methods at 75 known associated SNPs, consistent with simulations. Larger increases in power are expected at larger sample sizes. In conclusion, case-control studies of diseases with low prevalence can achieve power higher than that in existing mixed-model methods.