Search

BACKGROUND: Single nucleotide polymorphism-based heritability is a fundamental quantity in the genetic analysis of complex traits. For case-control phenotypes, for which the continuous distribution of risk in the population is unobserved, observed-scale heritability estimates must be transformed to the more interpretable liability scale. This article describes how the field standard approach incorrectly performs the liability correction in that it does not appropriately account for variation in the proportion of cases across the cohorts comprising the meta-analysis. We propose a simple solution that incorporates cohort-specific ascertainment using the summation of effective sample sizes across cohorts. This solution is applied at the stage of single nucleotide polymorphism-based heritability estimation and does not require generating updated meta-analytic genome-wide association study summary statistics.METHODS: We began by performing a series of simulations to examine the ability of the standard approach and our proposed approach to recapture liability-scale heritability in the population. We went on to examine the differences in estimates obtained from these 2 approaches for real data for 12 major case-control genome-wide association studies of psychiatric and neurologic traits.RESULTS: We found that the field standard approach for performing the liability conversion can downwardly bias estimates by as much as approximately 50% in simulation and approximately 30% in real data.CONCLUSIONS: Prior estimates of liability-scale heritability for genome-wide association study meta-analysis may be drastically underestimated. To this end, we strongly recommend using our proposed approach of using the sum of effective sample sizes across contributing cohorts to obtain unbiased estimates.

Background Resource trade-off theory suggests that increased performance on a given trait comes at the cost of decreased performance on other traits. Methods Growth data from 1889 subjects (996 girls) were used from the GrowUp1974 Gothenburg study. Energy Trade-Off (ETO) between height and weight for individuals with extreme body types was characterized using a novel ETO-Score (ETOS). Four extreme body types were defined based on height and ETOI at early adulthood: tall-slender, short-stout, short-slender, and tall-stout; their growth trajectories assessed from ages 0.5-17.5 years.A GWAS using UK BioBank data was conducted to identify gene variants associated with height, BMI, and for the first time with ETOS. Results Height and ETOS trajectories show a two-hit pattern with profound changes during early infancy and at puberty for tall-slender and short-stout body types. Several loci (including FTO, ADCY3, GDF5, ) and pathways were identified by GWAS as being highly associated with ETOS. The most strongly associated pathways were related to “extracellular matrix,” “signal transduction,” “chromatin organization,” and “energy metabolism.” Conclusions ETOS represents a novel anthropometric trait with utility in describing body types. We discovered the multiple genomic loci and pathways probably involved in energy trade-off.

Polygenic scores (PGSs) have limited portability across different groupings of individuals (for example, by genetic ancestries and/or social determinants of health), preventing their equitable use1–3. PGS portability has typically been assessed using a single aggregate population-level statistic (for example, R2)4, ignoring inter-individual variation within the population. Here, using a large and diverse Los Angeles biobank5 (ATLAS, n = 36,778) along with the UK Biobank6 (UKBB, n = 487,409), we show that PGS accuracy decreases individual-to-individual along the continuum of genetic ancestries7 in all considered populations, even within traditionally labelled ‘homogeneous’ genetic ancestries. The decreasing trend is well captured by a continuous measure of genetic distance (GD) from the PGS training data: Pearson correlation of −0.95 between GD and PGS accuracy averaged across 84 traits. When applying PGS models trained on individuals labelled as white British in the UKBB to individuals with European ancestries in ATLAS, individuals in the furthest GD decile have 14% lower accuracy relative to the closest decile; notably, the closest GD decile of individuals with Hispanic Latino American ancestries show similar PGS performance to the furthest GD decile of individuals with European ancestries. GD is significantly correlated with PGS estimates themselves for 82 of 84 traits, further emphasizing the importance of incorporating the continuum of genetic ancestries in PGS interpretation. Our results highlight the need to move away from discrete genetic ancestry clusters towards the continuum of genetic ancestries when considering PGSs.

Motivation Measuring genetic diversity is an important problem because increasing genetic diversity is a key to making new genetic discoveries, while also being a major source of confounding to be aware of in genetics studies. Results Using the UK Biobank data, a prospective cohort study with deep genetic and phenotypic data collected on almost 500 000 individuals from across the UK, we carefully define 21 distinct ancestry groups from all four corners of the world. These ancestry groups can serve as a global reference of worldwide populations, with a handful of applications. Here, we develop a method that uses allele frequencies and principal components derived from these ancestry groups to effectively measure ancestry proportions from allele frequencies of any genetic dataset. Availability and implementation This method is implemented in function snp_ancestry_summary of R package bigsnpr. Supplementary information Supplementary data are available at Bioinformatics online.

The complement system, including complement components 3 and 4 (C3, C4), traditionally has been linked to innate immunity. More recently, complement components have also been implicated in brain development and the risk of schizophrenia. Based on a large, population-based case-cohort study, we measured the blood concentrations of C3 and C4 in 68,768 neonates. We found a strong correlation between the concentrations of C3 and C4 (phenotypic correlation = 0.65,P-value < 1.0×10−100, genetic correlation = 0.38,P-value = 1.9×10−35). A genome-wide association study (GWAS) for C4 protein concentration identified 36 independent loci, 30 of which were in or near the major histocompatibility complex on chromosome 6 (which includes theC4gene), while six loci were found on six other chromosomes. A GWAS for C3 identified 15 independent loci, seven of which were located in theC3gene on chromosome 19, and eight loci on five other chromosomes. We found no association between (a) measured neonatal C3 and C4 concentrations, imputed C4 haplotypes, or predictedC4gene expression, with (b) schizophrenia (SCZ), bipolar disorder (BIP), depression (DEP), autism spectrum disorder, attention deficit hyperactivity disorder or anorexia nervosa diagnosed in later life. Mendelian randomisation (MR) suggested a small positive association between higher C4 protein concentration and an increased risk of SCZ, BIP, and DEP, but these findings did not persist in more stringent analyses. Evidence from MR supported causal relationships between C4 concentration and several autoimmune disorders: systemic lupus erythematosus (SLE, OR and 95% confidence interval, 0.37, 0.34 – 0.42); type-1 diabetes (T1D, 0.54, 0.50 - 0.58); multiple sclerosis (MS, 0.68, 0.63 - 0.74); rheumatoid arthritis (0.85, 0.80 - 0.91); and Crohn’s disease (1.26, 1.19 - 1.34). A phenome-wide association study (PheWAS) in UK Biobank confirmed that the genetic correlates of C4 concentration were associated a range of autoimmune disorders including coeliac disease, thyrotoxicosis, hypothyroidism, T1D, sarcoidosis, psoriasis, SLE and ankylosing spondylitis. We found no evidence of associations between C3 versus mental or autoimmune disorders based on either MR or PheWAS. In general, our results do not support the hypothesis that C4 is causally associated with the risk of SCZ (nor several other mental disorders). We provide new evidence to support the hypothesis that higher C4 concentration is associated with lower risks of autoimmune disorders.

LDpred2 is a widely used Bayesian method for building polygenic scores (PGS). LDpred2-auto can infer the two parameters from the LDpred model,h2andp, so that it does not require an additional validation dataset to choose best-performing parameters. Here, we present a new version of LDpred2-auto, which adds a third parameterαto its model for modeling negative selection. Additional changes are also made to provide better sampling of these parameters. We then validate the inference of these three parameters. LDpred2-auto also provides per-variant probabilities of being causal that are well calibrated, and can therefore be used for fine-mapping purposes. We also derive a new formula to infer the out-of-sample predictive performancer2of the resulting PGS directly from the Gibbs sampler of LDpred2-auto. Finally, we extend the set of HapMap3 variants recommended to use with LDpred2 with 37% more variants to improve the coverage of this set, and show that this new set of variants captures 12% more heritability and provides 6% more predictive performance, on average.

Polygenic scores (PGS) have limited portability across different groupings of individuals (e.g., by genetic ancestries and/or social determinants of health), preventing their equitable use. PGS portability has typically been assessed using a single aggregate population-level statistic (e.g., R2), ignoring inter-individual variation within the population. Here we evaluate PGS accuracy at individual-level resolution, independent of its annotated genetic ancestries. We show that PGS accuracy varies between individuals across the genetic ancestry continuum in all ancestries, even within traditionally “homogeneous” genetic ancestry clusters. Using a large and diverse Los Angeles biobank (ATLAS, N= 36,778) along with the UK Biobank (UKBB, N= 487,409), we show that PGS accuracy decreases along a continuum of genetic ancestries in all considered populations and the trend is well-captured by a continuous measure of genetic distance (GD) from the PGS training data; Pearson correlation of −0.95 between GD and PGS accuracy averaged across 84 traits. When applying PGS models trained in UKBB “white British” individuals to European-ancestry individuals of ATLAS, individuals in the highest GD decile have 14% lower accuracy relative to the lowest decile; notably the lowest GD decile of Hispanic/Latino American ancestry individuals showed similar PGS performance as the highest GD decile of European ancestry ATLAS individuals. GD is significantly correlated with PGS estimates themselves for 82 out of 84 traits, further emphasizing the importance of incorporating the continuum of genetic ancestry in PGS interpretation. Our results highlight the need for moving away from discrete genetic ancestry clusters towards the continuum of genetic ancestries when considering PGS and their applications.

The predictive performance of polygenic scores (PGS) is largely dependent on the number of samples available to train the PGS. Increasing the sample size for a specific phenotype is expensive and takes time, but this sample size can be effectively increased by using genetically correlated phenotypes. We propose a framework to generate multi-PGS from thousands of publicly available genome-wide association studies (GWAS) with no need to individually select the most relevant ones. In this study, the multi-PGS framework increased prediction accuracy over single PGS for all included psychiatric disorders and other available outcomes, with prediction R2 increases of up to 9-fold for attention-deficit/hyperactivity disorder (ADHD) compared to a single PGS. We also generate multi-PGS for phenotypes without an existing GWAS and for case-case predictions, with up to 15-fold increases in prediction accuracy. We benchmark the multi-PGS framework against other methods and highlight its potential application to new emerging biobanks.

Resource trade-off theory suggests that increased performance on a given trait comes at the cost of decreased performance on other traits.Growth data from 1889 subjects (996 girls) were used from the GrowUp1974 Gothenburg study. Energy Trade-Off (ETO) between height and weight for individuals with extreme body types was characterized using a novel ETO-Score (ETOS). Four extreme body types were defined based on height and ETOI at early adulthood: tall-slender, short-stout, short-slender, and tall-stout; their growth trajectories assessed from ages 0.5-17.5 years.A GWAS using UK BioBank data was conducted to identify gene variants associated with height, BMI, and for the first time with ETOS.Height and ETOS trajectories show a two-hit pattern with profound changes during early infancy and at puberty for tall-slender and short-stout body types.Several loci (including FTO, ADCY3, GDF5, ) and pathways were identified by GWAS as being highly associated with ETOS. The most strongly associated pathways were related to 'extracellular matrix', 'signal transduction', 'chromatin organization', and 'energy metabolism'.ETOS represents a novel anthropometric trait with utility in describing body types. We discovered the multiple genomic loci and pathways probably involved in energy trade-off.

Proportional hazards models have previously been proposed to analyse time-to-event phenotypes in genome-wide association studies(GWAS). While proportional hazards models have many useful applications, their ability to identify genetic associations under different generative models where ascertainment is present in the analysed data is poorly understood. This includes widely used study designs such as case-control and case-cohort designs (e.g. the iPSYCH study design) where cases are commonly ascertained.Here we examine how recently proposed and computationally efficient Cox regression for GWAS perform under different generative models with and without ascertainment. We also propose the age-dependent liability threshold model (ADuLT), first introduced as the underlying model for the LT-FH++ method, as an alternative approach for time-to-event GWAS. We then benchmark ADuLT with SPACox and standard case-control GWAS using simulated data with varying degrees of ascertainment. We find Cox regression GWAS to underperform when cases are strongly ascertained (cases are oversampled by a factor larger than 5), regardless of the generative model used. In contrast, we found ADuLT to be robust to case-control ascertainment, while being much faster to run. We then used the methods to conduct GWAS for four psychiatric disorders, ADHD, Autism, Depression, and Schizophrenia in the iPSYCH case-cohort sample, which has a strong case-ascertainment. Summarising across all four mental disorders, ADuLT found 20 independent genome-wide significant associations, while case-control GWAS found 17 and SPACox found 8, consistent with our simulation results.As more genetic data are being linked to electronic health records, robust GWAS methods that can make use of age-of-onset information have the opportunity to increase power in analyses. We find that ADuLT to be a robust time-to-event GWAS method that performs on par with or better than Cox-regression GWAS, both in simulations and real data analyses of four psychiatric disorders. ADuLT has been implemented in an R package called LTFHPlus, and is available on GitHub.

The vitamin D binding protein (DBP), encoded by the group-specific component (GC) gene, is a much-studied component of the vitamin D system. In a genome-wide association study of DBP concentration in 65,589 neonates, we identified 26 independent loci, 17 of which were in or close to the GC gene, with fine-mapping identifying 2 loci on chromosomes 12 and 17 (missense variants within SH2B3 and GSDMA, respectively). When adjusted for key GC haplotypes, we found 15 independent loci distributed over 10 chromosomes. Mendelian randomization analyses found evidence consistent with a unidirectional, causal effect of higher DBP concentration and (a) higher 25 hydroxyvitamin D (25OHD) concentration, and (b) a reduced risk of multiple sclerosis and rheumatoid arthritis. A phenome-wide association study in an external dataset confirmed that higher DBP concentration was associated with higher 25OHD concentration and a reduced risk of vitamin D deficiency. Our study provides new insights into the influence of DBP on vitamin D status and a range of health outcomes.

Motivation Principal component analysis (PCA) of genetic data is routinely used to infer ancestry and control for population structure in various genetic analyses. However, conducting PCA analyses can be complicated and has several potential pitfalls. These pitfalls include (i) capturing linkage disequilibrium (LD) structure instead of population structure, (ii) projected PCs that suffer from shrinkage bias, (iii) detecting sample outliers and (iv) uneven population sizes. In this work, we explore these potential issues when using PCA, and present efficient solutions to these. Following applications to the UK Biobank and the 1000 Genomes project datasets, we make recommendations for best practices and provide efficient and user-friendly implementations of the proposed solutions in R packages bigsnpr and bigutilsr. Results For example, we find that PC19–PC40 in the UK Biobank capture complex LD structure rather than population structure. Using our automatic algorithm for removing long-range LD regions, we recover 16 PCs that capture population structure only. Therefore, we recommend using only 16–18 PCs from the UK Biobank to account for population structure confounding. We also show how to use PCA to restrict analyses to individuals of homogeneous ancestry. Finally, when projecting individual genotypes onto the PCA computed from the 1000 Genomes project data, we find a shrinkage bias that becomes large for PC5 and beyond. We then demonstrate how to obtain unbiased projections efficiently using bigsnpr. Overall, we believe this work would be of interest for anyone using PCA in their analyses of genetic data, as well as for other omics data. Availability and implementation R packages bigsnpr and bigutilsr can be installed from either CRAN or GitHub (see https://github.com/privefl/bigsnpr). A tutorial on the steps to perform PCA on 1000G data is available at https://privefl.github.io/bigsnpr/articles/bedpca.html. All code used for this paper is available at https://github.com/privefl/paper4-bedpca/tree/master/code. Supplementary information Supplementary data are available at Bioinformatics online.

Objective: To estimate phenotypic and familial association between early-life injuries and attention-deficit/hyperactivity disorder (ADHD) and the genetic contribution to the association using polygenic risk score for ADHD (PRS-ADHD) and genetic correlation analyses.Methods: Children born in Denmark between 1995-2010 (n = 786,543) were followed from age 5 years until a median age of 14 years (interquartile range: 10-18 years). Using ICD-10 diagnoses, we estimated hazard ratios (HRs) and absolute risks of ADHD by number of hospital/emergency ward-treated injuries by age 5. In a subset of ADHD cases and controls born 1995 to 2005 who had genetic data available (n = 16,580), we estimated incidence rate ratios (IRRs) for the association between PRS-ADHD and number of injuries before age 5 and the genetic correlation between ADHD and any injury before age 5.Results: Injuries were associated with ADHD (HR = 1.61; 95% CI, 1.55-1.66) in males (HR = 1.59; 1.53-1.65) and females (HR = 1.65; 1.54-1.77), with a dose-response relationship with number of injuries. The absolute ADHD risk by age 15 was 8.4% (3+ injuries) vs 3.1% (no injuries). ADHD was also associated with injuries in relatives, with a stronger association in first- than second-degree relatives. PRS-ADHD was marginally associated with the number of injuries in the general population (IRR = 1.06; 1.00-1.14), with a genetic correlation of 0.53 (0.21-0.85).Conclusions: Early-life injuries in individuals and their relatives were associated with a diagnosis of ADHD. However, even in children with the most injuries, more than 90% were not diagnosed with ADHD by age 15. Despite a low positive predictive value and that the impact of unmeasured factors such as parental behavior remains unclear, results indicate that the association is partly explained by genetics, suggesting that early-life injuries may represent or herald early behavioral manifestations of ADHD.

Genome-wide association studies (GWASs) have revolutionized human genetics, allowing researchers to identify thousands of disease-related genes and possible drug targets. However, case-control status does not account for the fact that not all controls may have lived through their period of risk for the disorder of interest. This can be quantified by examining the age-of-onset distribution and the age of the controls or the age of onset for cases. The age-of-onset distribution may also depend on information such as sex and birth year. In addition, family history is not routinely included in the assessment of control status. Here, we present LT-FH++, an extension of the liability threshold model conditioned on family history (LT-FH), which jointly accounts for age of onset and sex as well as family history. Using simulations, we show that, when family history and the age-of-onset distribution are available, the proposed approach yields statistically significant power gains over LT-FH and large power gains over genome-wide association study by proxy (GWAX). We applied our method to four psychiatric disorders available in the iPSYCH data and to mortality in the UK Biobank and found 20 genome-wide significant associations with LT-FH++, compared to ten for LT-FH and eight for a standard case-control GWAS. As more genetic data with linked electronic health records become available to researchers, we expect methods that account for additional health information, such as LT-FH++, to become even more beneficial.

Most existing tools for constructing genetic prediction models begin with the assumption that all genetic variants contribute equally towards the phenotype. However, this represents a suboptimal model for how heritability is distributed across the genome. Therefore, we develop prediction tools that allow the user to specify the heritability model. We compare individual-level data prediction tools using 14 UK Biobank phenotypes; our new tool LDAK-Bolt-Predict outperforms the existing tools Lasso, BLUP, Bolt-LMM and BayesR for all 14 phenotypes. We compare summary statistic prediction tools using 225 UK Biobank phenotypes; our new tool LDAK-BayesR-SS outperforms the existing tools lassosum, sBLUP, LDpred and SBayesR for 223 of the 225 phenotypes. When we improve the heritability model, the proportion of phenotypic variance explained increases by on average 14%, which is equivalent to increasing the sample size by a quarter., Existing genetic prediction tools typically assume that genetic variants contribute equally towards the phenotype. The authors develop eight prediction tools that allow the user to specify the heritability model, and show that these tools enable substantially improved prediction of complex traits.

The low portability of polygenic scores (PGS) across global populations is a major concern that must be addressed before PGS can be used for everyone in the clinic. Indeed, prediction accuracy has been shown to decay as a function of the genetic distance between the training and test cohorts. However, such cohorts differ not only in their genetic distance but also in their geographical distance and their data collection and assaying, conflating multiple factors. In this study, we examine the extent to which PGS are transferable between ancestries by deriving polygenic scores for 245 curated traits from the UK Biobank data and applying them in nine ancestry groups from the same cohort. By restricting both training and testing to the UK Biobank data, we reduce the risk of environmental and genotyping confounding from using different cohorts. We define the nine ancestry groups at a high-resolution, country-specific level, based on a simple, robust and effective method that we introduce here. We then apply two different predictive methods to derive polygenic scores for all 245 phenotypes, and show a systematic and dramatic reduction in portability of PGS trained in the inferred ancestral UK population and applied to the inferred ancestral Polish - Italian - Iranian - Indian - Chinese - Caribbean - Nigerian - Ashkenazi populations, respectively. These analyses, performed at a finer scale than the usual continental scale, demonstrate that prediction already drops off within European ancestries and reduces globally in proportion to PC distance, even when all individuals reside in the same country and are genotyped and phenotyped as part of the same cohort. Our study provides high-resolution and robust insights into the PGS portability problem.

The accuracy of polygenic risk scores (PRSs) to predict complex diseases increases with the training sample size. PRSs are generally derived based on summary statistics from large meta-analyses of multiple genome-wide association studies (GWAS). However, it is now common for researchers to have access to large individual-level data as well, such as the UK biobank data. To the best of our knowledge, it has not yet been explored how to best combine both types of data (summary statistics and individual-level data) to optimize polygenic prediction. The most widely used approach to combine data is the meta-analysis of GWAS summary statistics (Meta-GWAS), but we show that it does not always provide the most accurate PRS. Through simulations and using twelve real case-control and quantitative traits from both iPSYCH and UK Biobank along with external GWAS summary statistics, we compare Meta-GWAS with two alternative data-combining approaches, stacked clumping and thresholding (SCT) and Meta-PRS. We find that, when large individual-level data is available, the linear combination of PRSs (Meta-PRS) is both a simple alternative to Meta-GWAS and often more accurate.

The effective sample size (ESS) is a quantity estimated in genome-wide association studies (GWAS) with related individuals and/or linear mixed models used in analysis. ESS originally measured relative power in family-based GWAS and has recently become important for correcting GWAS summary statistics in post-GWAS analyses. However, existing ESS approaches have been overlooked and based on empirical estimation. This work presents an analytical form of ESS in mixed-model GWAS of quantitative traits, which is derived using the expectation of quadratic form and validated in extensive simulations. We illustrate the performance and relevance of our ESS estimator in common GWAS scenarios and analytically show that (i) family-based studies are consistently underpowered compared to studies of unrelated individuals of the same sample size; (ii) conditioning on polygenic genetic effect by linear mixed models boosts power; and (iii) power of detecting gene-environment interaction can be substantially gained or lost in family-based designs depending on exposure distribution. We further analyze UK Biobank dataset in two samples of 336,347 unrelated and 68,910 related individuals. Analysis in unrelated individuals reveals a high accuracy of our ESS estimator compared to the existing empirical approach; and analysis of related individuals suggests that the loss in effective sample size due to relatedness is at most 0.94x. Overall, we provide an analytical form of ESS for guiding GWAS designs and processing summary statistics in post-GWAS analyses.

Genome-wide association studies (GWAS) have revolutionized human genetics, allowing researchers to identify thousands of disease-related genes and possible drug targets. However, case-control status does not account for the fact that not all controls may have lived through their period of risk for the disorder of interest. This can be quantified by examining the age-of-onset distribution and the age of the controls or the age-of-onset for cases. The age-of-onset distribution may also depend on information such as sex and birth year. In addition, family history is not routinely included in the assessment of control status. Here we present LT-FH++, an extension of the liability threshold model conditioned on family history (LT-FH), that jointly accounts for age-of-onset and sex, as well as family history. Using simulations, we show that, when family history and the age-of-onset distribution are available, the proposed approach yields large power gains over both LT-FH and genome-wide association study by proxy (GWAX). We applied our method to four psychiatric disorders available in the iPSYCH data, and to mortality in the UK Biobank, finding 20 genome-wide significant associations with LT-FH++, compared to 10 for LT-FH and 8 for a standard case-control GWAS. As more genetic data with linked electronic health records become available to researchers, we expect methods that account for additional health information, such as LT-FH++, to become even more beneficial.

Publicly available genome-wide association studies (GWAS) summary statistics exhibit uneven quality, which can impact the validity of follow-up analyses. First, we present an overview of possible misspecifications that come with GWAS summary statistics. Then, in both simulations and real data analyses, we show that additional information such as imputation INFO scores, allele frequencies, and per-variant sample sizes in GWAS summary statistics can be used to detect possible issues and correct for misspecifications in the GWAS summary statistics. One important motivation for us is to improve the predictive performance of polygenic scores built from these summary statistics. Unfortunately, due to the lack of reporting standards for GWAS summary statistics, this additional information is not systematically reported. We also show that using well-matched LD references can improve model fit and translate into more accurate prediction. Finally, we discuss how to make polygenic score methods such as lassosum and LDpred2 more robust to these misspecifications to improve their predictive power.

Motivation A few algorithms have been developed for splitting the genome in nearly independent blocks of linkage disequilibrium. Due to the complexity of this problem, these algorithms rely on heuristics, which makes them suboptimal. Results Here, we develop an optimal solution for this problem using dynamic programming. Availability This is now implemented as function snp_ldsplit as part of R package bigsnpr. Supplementary information Supplementary data are available at Bioinformatics online.

Although the cohort-level accuracy of polygenic risk scores (PRSs)-estimates of genetic value at the individual level-has been widely assessed, uncertainty in PRSs remains underexplored. In the present study, we show that Bayesian PRS methods can estimate the variance of an individual's PRS and can yield well-calibrated credible intervals via posterior sampling. For 13 real traits in the UK Biobank (n = 291,273 unrelated 'white British'), we observe large variances in individual PRS estimates which impact interpretation of PRS-based stratification; averaging across traits, only 0.8% (s.d. = 1.6%) of individuals with PRS point estimates in the top decile have corresponding 95% credible intervals fully contained in the top decile. We provide an analytical estimator for the expectation of individual PRS variance as a function of SNP heritability, number of causal SNPs and sample size. Our results showcase the importance of incorporating uncertainty in individual PRS estimates into subsequent analyses.

Motivation Polygenic scores have become a central tool in human genetics research. LDpred is a popular method for deriving polygenic scores based on summary statistics and a matrix of correlation between genetic variants. However, LDpred has limitations that may reduce its predictive performance. Results Here, we present LDpred2, a new version of LDpred that addresses these issues. We also provide two new options in LDpred2: a ‘sparse’ option that can learn effects that are exactly 0, and an ‘auto’ option that directly learns the two LDpred parameters from data. We benchmark predictive performance of LDpred2 against the previous version on simulated and real data, demonstrating substantial improvements in robustness and predictive accuracy compared to LDpred1. We then show that LDpred2 also outperforms other polygenic score methods recently developed, with a mean AUC over the 8 real traits analyzed here of 65.1%, compared to 63.8% for lassosum, 62.9% for PRS-CS and 61.5% for SBayesR. Note that LDpred2 provides more accurate polygenic scores when run genome-wide, instead of per chromosome. Availability and implementation LDpred2 is implemented in R package bigsnpr. Supplementary information Supplementary data are available at Bioinformatics online.

Large-scale genome-wide association studies have enabled polygenic risk scores (PRS), which estimate the genetic value of an individual for a given trait. Since PRS accuracy is typically assessed using cohort-level metrics (e.g., R2), uncertainty in PRS estimates at individual level remains underexplored. Here we show that Bayesian PRS methods can estimate the variance of an individual’s PRS and can yield well-calibrated credible intervals for the genetic value of a single individual. For real traits in the UK Biobank (N=291,273 unrelated “white British”) we observe large variance in individual PRS estimates which impacts interpretation of PRS-based stratification; for example, averaging across 13 traits, only 0.8% (s.d. 1.6%) of individuals with PRS point estimates in the top decile have their entire 95% credible intervals fully contained in the top decile. We provide an analytical estimator for individual PRS variance—a function of SNP-heritability, number of causal SNPs, and sample size—and observe high concordance with individual variances estimated via posterior sampling. Finally as an example of the utility of individual PRS uncertainties, we explore a probabilistic approach to PRS-based stratification that estimates the probability of an individual’s genetic value to be above a prespecified threshold. Our results showcase the importance of incorporating uncertainty in individual PRS estimates into subsequent analyses.

Both R packages snpnet and bigstatsr allow for fitting penalized regressions on individual-level genetic data as large as the UK Biobank. Here we benchmark bigstatsr against snpnet for fitting penalized regressions on large genetic data. We find bigstatsr to be an order of magnitude faster than snpnet when applied to the UK Biobank data (from 4.5x to 35x). We also discuss the similarities and differences between the two packages, provide theoretical insights, and make recommendations on how to fit penalized regressions in the context of genetic data.

Here we propose a simple, robust and effective method for global ancestry inference and grouping from Principal Component Analysis (PCA) of genetic data. The proposed approach is particularly useful for methods that need to be applied in homogeneous samples. First, we show that Euclidean distances in the PCA space are proportional to FST between populations. Then, we show how to use this PCA-based distance to infer ancestry in the UK Biobank and the POPRES datasets. We propose two solutions, either relying on projection of PCs to reference populations such as from the 1000 Genomes Project, or by directly using the internal data. Finally, we conclude that our method and the community would benefit from having an easy access to a reference dataset with an even better coverage of the worldwide genetic diversity than the 1000 Genomes Project.

The objective of the present study was to investigate whether the polygenic liability for attention-deficit/hyperactivity disorder (ADHD) and the psychosocial environment impact the risk of ADHD in interaction or independently of each other. We conducted a register- and biobank-based cohort study of 13,725 individuals with ADHD and 20,147 randomly drawn population-based controls. These 33,872 cohort members were genotyped on the Infinium PsychChip v1.0 array (Illumina). Subsequently, we calculated the polygenic risk score (PRS) for ADHD and extracted register data regarding the following risk factors pertaining to the psychosocial environment for each cohort member at the time of birth: maternal/paternal history of mental disorders, maternal/paternal education, maternal/paternal work status, and maternal/paternal income. We used logistic regression analyses to assess the main effects of the PRS for ADHD and the psychosocial environment on the risk of ADHD. Subsequently, we evaluated whether the effect of the PRS and the psychosocial environment act independently or in interaction upon the risk of ADHD. We found that ADHD was strongly associated with the PRS (odds ratio: 6.03, 95%CI: 4.74–7.70 for highest vs. lowest 2% liability). All risk factors pertaining to the psychosocial environment were associated with an increased risk of ADHD. These associations were only slightly attenuated after mutual adjustments. We found no statistically significant interaction between the polygenic liability and the psychosocial environment upon the risk of ADHD. In conclusion, we found main effects of both polygenic liability and risk factors pertaining to the psychosocial environment on the risk of ADHD—in the expected direction.

R package pcadapt is a user-friendly R package for performing genome scans for local adaptation. Here, we present version 4 of pcadapt which substantially improves computational efficiency while providing similar results. This improvement is made possible by using a different format for storing genotypes and a different algorithm for computing principal components of the genotype matrix, which is the most computationally demanding step in method pcadapt. These changes are seamlessly integrated into the existing pcadapt package, and users will experience a large reduction in computation time (by a factor of 20–60 in our analyses) as compared with previous versions.

Polygenic prediction has the potential to contribute to precision medicine. Clumping and Thresh-olding (C+T) is a widely used method to derive polygenic scores. When using C+T, it is common to test several p-value thresholds to maximize predictive ability of the derived polygenic scores. Along with this p-value threshold, we propose to tune three other hyper-parameters for C+T. We implement an efficient way to derive thousands of different C+T polygenic scores corresponding to a grid over four hyper-parameters. For example, it takes a few hours to derive 123,200 different C+T scores for 300K individuals and 1M variants on a single node with 16 cores.We find that optimizing over these four hyper-parameters improves the predictive performance of C+T in both simulations and real data applications as compared to tuning only the p-value threshold. A particularly large increase can be noted when predicting depression status, from an AUC of 0.557 (95% CI: [0.544-0.569]) when tuning only the p-value threshold in C+T to an AUC of 0.592 (95% CI: [0.580-0.604]) when tuning all four hyper-parameters we propose for C+T.We further propose Stacked Clumping and Thresholding (SCT), a polygenic score that results from stacking all derived C+T scores. Instead of choosing one set of hyper-parameters that maximizes prediction in some training set, SCT learns an optimal linear combination of all C+T scores by using an efficient penalized regression. We apply SCT to 8 different case-control diseases in the UK biobank data and find that SCT substantially improves prediction accuracy with an average AUC increase of 0.035 over standard C+T.

Psychiatric disorders are highly genetically correlated, but little research has been conducted on the genetic differences between disorders. We developed a new method (CC-GWAS) to test for differences in allele frequency among cases of two disorders using summary statistics from the respective case-control GWAS, transcending current methods that require individual-level data. Simulations and analytical computations confirm that CC-GWAS is well-powered with effective control of type I error. We applied CC-GWAS to publicly available summary statistics for schizophrenia, bipolar disorder, major depressive disorder, and five other psychiatric disorders. CC-GWAS identified 196 independent case-case loci, including 72 CC-GWAS-specific loci that were not genome-wide significant in the input case-control summary statistics; two of the CC-GWAS-specific loci implicate the genes KLF6 and KLF16 (from the Kruppel-like family of transcription factors), which have been linked to neurite outgrowth and axon regeneration. CC-GWAS loci replicated convincingly in applications to data sets with independent replication data.

Polygenic risk scores (PRS) combine many single-nucleotide polymorphisms into a score reflecting the genetic risk of developing a disease. Privé, Aschard, and Blum present an efficient implementation of penalized logistic regression..., Polygenic Risk Scores (PRS) combine genotype information across many single-nucleotide polymorphisms (SNPs) to give a score reflecting the genetic risk of developing a disease. PRS might have a major impact on public health, possibly allowing for screening campaigns to identify high-genetic risk individuals for a given disease. The “Clumping+Thresholding” (C+T) approach is the most common method to derive PRS. C+T uses only univariate genome-wide association studies (GWAS) summary statistics, which makes it fast and easy to use. However, previous work showed that jointly estimating SNP effects for computing PRS has the potential to significantly improve the predictive performance of PRS as compared to C+T. In this paper, we present an efficient method for the joint estimation of SNP effects using individual-level data, allowing for practical application of penalized logistic regression (PLR) on modern datasets including hundreds of thousands of individuals. Moreover, our implementation of PLR directly includes automatic choices for hyper-parameters. We also provide an implementation of penalized linear regression for quantitative traits. We compare the performance of PLR, C+T and a derivation of random forests using both real and simulated data. Overall, we find that PLR achieves equal or higher predictive performance than C+T in most scenarios considered, while being scalable to biobank data. In particular, we find that improvement in predictive performance is more pronounced when there are few effects located in nearby genomic regions with correlated SNPs; for instance, in simulations, AUC values increase from 83% with the best prediction of C+T to 92.5% with PLR. We confirm these results in a data analysis of a case-control study for celiac disease where PLR and the standard C+T method achieve AUC values of 89% and of 82.5%. Applying penalized linear regression to 350,000 individuals of the UK Biobank, we predict height with a larger correlation than with the best prediction of C+T (∼65% instead of ∼55%), further demonstrating its scalability and strong predictive power, even for highly polygenic traits. Moreover, using 150,000 individuals of the UK Biobank, we are able to predict breast cancer better than C+T, fitting PLR in a few minutes only. In conclusion, this paper demonstrates the feasibility and relevance of using penalized regression for PRS computation when large individual-level datasets are available, thanks to the efficient implementation available in our R package bigstatsr.

Polygenic risk scores are en vogue. This ubiquitous statistic quantifies disease liability for an individual by aggregating risk contributions from a large number of genetic variants into a single score. Recent publications have argued that risk models currently used in clinical settings for coronary artery disease can be improved by including polygenic risk scores.1,2 Similarly, polygenic risk scores have shown promise in improving breast cancer risk prediction3 and are already routinely used by direct-to-consumer genetic testing companies, such as 23andMe, to estimate disease risk. Now, in this issue of Neurology® Genetics , Kogelman et al.4 find that the polygenic risk score for migraine, a common headache disorder that is thought to affect about 18% of the population5 and estimated to be quite heritable (between 34% and 57%),6 correlates with triptans treatment response when treating migraine. However, there is no reason for people having migraines to rush and get genotyped. First, the treatment response effect was found to be small (but significant). Second, although Kogelman et al. accounted for population structure in their statistical analysis, it is hard to rule out other sources of confounding. Third, it is very difficult to estimate population risk, or conditional risk, as the sample used in this study (and most other studies) is of course ascertained (nonrandom sample). Fourth, if in doubt about the treatment response, why not try the drug?

International audience; Software is essential for data science. However, several software tools remain out of reach for many users due to a lack of documentation, thus limiting progress in the field. Tutorial development by authors and users can greatly improve a tool's accessibility and accelerate its adoption. In this article, we explore hackathons such as hackseq as a venue for authors and users to develop tutorials to address the lack of documented software. We describe four advantages of hackathon-driven tutorial development as well as three challenges that we faced. We also discuss our experience with remote participation. In short, if properly prepared, hackathons can provide a productive venue for assembling a group of passionate people, including remote participants, to develop a suite of related tutorials and address the growing need for accessible software.