Polygenic Prediction Research Articles

Extremely sparse models of linkage disequilibrium in ancestrally diverse association studies.

Linkage disequilibrium (LD) is the correlation among nearby genetic variants. In genetic association studies, LD is often modeled using large correlation matrices, but this approach is inefficient, especially in ancestrally diverse studies. In the present study, we introduce LD graphical models (LDGMs), which are an extremely sparse and efficient representation of LD. LDGMs are derived from genome-wide genealogies; statistical relationships among alleles in the LDGM correspond to genealogical relationships among haplotypes. We published LDGMs and ancestry-specific LDGM precision matrices for 18 million common variants (minor allele frequency >1%) in five ancestry groups, validated their accuracy and demonstrated order-of-magnitude improvements in runtime for commonly used LD matrix computations. We implemented an extremely fast multiancestry polygenic prediction method, BLUPx-ldgm, which performs better than a similar method based on the reference LD correlation matrix. LDGMs will enable sophisticated methods that scale to ancestrally diverse genetic association data across millions of variants and individuals.

Genome-Wide Meta-analysis Identifies Risk Loci and Improves Disease Prediction of Age-Related Macular Degeneration

To identify age-related macular degeneration (AMD) risk loci and to establish a polygenic prediction model. Genome-wide association study (GWAS) and polygenic risk score (PRS) construction. We included 64 885 European patients with AMD and 568 740 control participants (with overlapped samples) in the UK Biobank, Genetic Epidemiology Research on Aging (GERA), International AMD Consortium, FinnGen, and published early AMD GWASs in meta-analyses, as well as 733 European patients with AMD and 20 487 control participants from the Canadian Longitudinal Study on Aging (CLSA) and non-Europeans from the UK Biobank and GERA for polygenic risk score validation. A multitrait meta-analysis of GWASs comprised 64 885 patients with AMD and 568 740 control participants; the multitrait approach accounted for sample overlap. We constructed a PRS for AMD based on both previously reported as well as unreported AMD loci. We applied the PRS to nonoverlapping data from the CLSA. We identified several single nucleotide polymorphisms associated with AMD and established a PRS for AMD risk prediction. We identified 63 AMD risk loci alongside the well-established AMD loci CFH and ARMS2, including 9 loci that were not reported in previous GWASs, some of which previously were linked to other eye diseases such as glaucoma (e.g., HIC1). We applied our PRS to nonoverlapping data from the CLSA. A new PRS was constructed using the PRS method, PRS-CS, and significantly improved the prediction accuracy of AMD risk compared with PRSs from previously published datasets. We further showed that even people who carry all the well-known AMD risk alleles at CFH and ARMS2 vary considerably in their AMD risk (ranging from close to 0 in individuals with low PRS to > 50% in individuals with high PRS). Although our PRS was derived in individuals of European ancestry, the PRS shows potential for predicting risk in people of East Asian, South Asian, and Latino ancestry. Our findings improve the knowledge of the genetic architecture of AMD and help achieve better accuracy in AMD prediction. Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.

Multi-PGS enhances polygenic prediction by combining 937 polygenic scores

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 increases 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 compared to a single PGS. We also generate multi-PGS for phenotypes without an existing GWAS and for case-case predictions. We benchmark the multi-PGS framework against other methods and highlight its potential application to new emerging biobanks.

Biobank-scale methods and projections for sparse polygenic prediction from machine learning

In this paper we characterize the performance of linear models trained via widely-used sparse machine learning algorithms. We build polygenic scores and examine performance as a function of training set size, genetic ancestral background, and training method. We show that predictor performance is most strongly dependent on size of training data, with smaller gains from algorithmic improvements. We find that LASSO generally performs as well as the best methods, judged by a variety of metrics. We also investigate performance characteristics of predictors trained on one genetic ancestry group when applied to another. Using LASSO, we develop a novel method for projecting AUC and correlation as a function of data size (i.e., for new biobanks) and characterize the asymptotic limit of performance. Additionally, for LASSO (compressed sensing) we show that performance metrics and predictor sparsity are in agreement with theoretical predictions from the Donoho-Tanner phase transition. Specifically, a future predictor trained in the Taiwan Precision Medicine Initiative for asthma can achieve an AUC of 0.63_{(0.02)} and for height a correlation of 0.648_{(0.009)} for a Taiwanese population. This is above the measured values of 0.61_{(0.01)} and 0.631_{(0.008)}, respectively, for UK Biobank trained predictors applied to a European population.

Multivariate genetic analysis of personality and cognitive traits reveals abundant pleiotropy.

Personality and cognitive function are heritable mental traits whose genetic foundations may be distributed across interconnected brain functions. Previous studies have typically treated these complex mental traits as distinct constructs. We applied the 'pleiotropy-informed' multivariate omnibus statistical test to genome-wide association studies of 35 measures of neuroticism and cognitive function from the UK Biobank (n = 336,993). We identified 431 significantly associated genetic loci with evidence of abundant shared genetic associations, across personality and cognitive function domains. Functional characterization implicated genes with significant tissue-specific expression in all tested brain tissues and brain-specific gene sets. We conditioned independent genome-wide association studies of the Big 5 personality traits and cognitive function on our multivariate findings, boosting genetic discovery in other personality traits and improving polygenic prediction. These findings advance our understanding of the polygenic architecture of these complex mental traits, indicating a prominence of pleiotropic genetic effects across higher order domains of mental function such as personality and cognitive function.

141-OR: Discrepancy between Genetically-Predicted and Actual Body Mass Index Predicts Incident Diabetes Mellitus

Background: Body mass index (BMI) is a key predictor of metabolic disorders including type 2 diabetes mellitus (T2D). Polygenic risk scores (PRS) derived from genome-wide association studies (GWAS) can predict BMI based on common variants, and genetically-predicted BMI (BMI-pred) could serve as a threshold for personal fat storage. We hypothesized that difference between BMI-pred and actual BMI (BMI-diff) could reflect severity of environmental factors affecting BMI and may predict incident T2D. Methods: From the UK Biobank cohort, we selected participants of white British descent without a history of T2D (n=383,160). The PRS for BMI was calculated using polygenic prediction via Bayesian regression and continuous shrinkage priors (PRS-CS) with summary statistics from the Genetic Investigation of ANthropometric Traits consortium GWAS. According to the BMI-diff, the 10-year risk of T2D was assessed using multivariable Cox proportional-hazards model. Results were validated in an independent cohort from South Korea (n=8,840). Results: Participants (mean age 56.8 ± 8.0, male 45.3%) were randomly divided into train (n=268,041) and test set (n=115,119) to establish BMI-pred using PRS-CS. In the test set, the BMI-pred could explain 7.1% of the variance of BMI and T2D occurred in 3,599 (3.1%) cases during 10-year follow-up. The risk of T2D increased linearly according to the decrease of BMI-diff. Participants in the higher quintiles of BMI-diff (more obese than genetically-predicted), had significantly higher risk of T2D than those in the lowest quintile (less obese than genetically-predicted) regardless of their own BMI: adjusted HR (aHR) of 1st quintile 1.61, 95% confidence interval (CI) 1.26-2.05, P <0.001; aHR of 2nd quintile 1.48, 95% CI 1.21-1.81, P <0.001. Results were consistent when validated in the South Korean cohort. Conclusion: Deviation of actual BMI from BMI-pred which reflects severity of environmental factors is a significant predictor of incident T2D. Disclosure T.Rhee: None. H.Lee: None. J.Choi: None. J.S.Lee: None. S.Kwak: None. K.Park: None.

150 risk variants for diverticular disease of intestine prioritize cell types and enable polygenic prediction of disease susceptibility

We conducted a genome-wide association study (GWAS) analysis of diverticular disease (DivD) of intestine within 724,372 individuals and identified 150 independent genome-wide significant DNA variants. Integration of the GWAS results with human gut single-cell RNA sequencing data implicated gut myocyte, mesothelial and stromal cells, and enteric neurons and glia in DivD development. Ninety-five genes were prioritized based on multiple lines of evidence, including SLC9A3, a drug target gene of tenapanor used for the treatment of the constipation subtype of irritable bowel syndrome. A DivD polygenic score (PGS) enables effective risk prediction (area under the curve [AUC], 0.688; 95% confidence interval [CI], 0.645-0.732) and the top 20% PGS was associated with ∼3.6-fold increased DivD risk relative to the remaining population. Our statistical and bioinformatic analyses suggest that the mechanism of DivD is through colon structure, gut motility, gastrointestinal mucus, and ionic homeostasis. Our analyses reinforce the link between gastrointestinal disorders and the enteric nervous system through genetics.

Biomarker and genomic analyses reveal molecular signatures of non-cardioembolic ischemic stroke

Acute ischemic stroke (AIS) is a major cause of disability and mortality worldwide. Non-cardioembolic ischemic stroke (NCIS), which constitutes the majority of AIS cases, is highly heterogeneous, thus requiring precision medicine treatments. This study aimed to investigate the molecular mechanisms underlying NCIS heterogeneity. We integrated data from the Third China National Stroke Registry, including clinical phenotypes, biomarkers, and whole-genome sequencing data for 7695 patients with NCIS. We identified 30 molecular clusters based on 63 biomarkers and explored the comprehensive landscape of biological heterogeneity and subpopulations in NCIS. Dimensionality reduction revealed fine-scale subpopulation structures associated with specific biomarkers. The subpopulations with biomarkers for inflammation, abnormal liver and kidney function, homocysteine metabolism, lipid metabolism, and gut microbiota metabolism were associated with a high risk of unfavorable clinical outcomes, including stroke recurrence, disability, and mortality. Several genes encoding potential drug targets were identified as putative causal genes that drive the clusters, such as CDK10, ERCC3, and CHEK2. We comprehensively characterized the genetic architecture of these subpopulations, identified their molecular signatures, and revealed the potential of the polybiomarkers and polygenic prediction for assessing clinical outcomes. Our study demonstrates the power of large-scale molecular biomarkers and genomics to understand the underlying biological mechanisms of and advance precision medicine for NCIS.

Polygenic prediction of preeclampsia and gestational hypertension.

Preeclampsia and gestational hypertension are common pregnancy complications associated with adverse maternal and child outcomes. Current tools for prediction, prevention and treatment are limited. Here we tested the association of maternal DNA sequence variants with preeclampsia in 20,064 cases and 703,117 control individuals and with gestational hypertension in 11,027 cases and 412,788 control individuals across discovery and follow-up cohorts using multi-ancestry meta-analysis. Altogether, we identified 18 independent loci associated with preeclampsia/eclampsia and/or gestational hypertension, 12 of which are new (for example, MTHFR-CLCN6, WNT3A, NPR3, PGR and RGL3), including two loci (PLCE1 and FURIN) identified in the multitrait analysis. Identified loci highlight the role of natriuretic peptide signaling, angiogenesis, renal glomerular function, trophoblast development and immune dysregulation. We derived genome-wide polygenic risk scores that predicted preeclampsia/eclampsia and gestational hypertension in external cohorts, independent of clinical risk factors, and reclassified eligibility for low-dose aspirin to prevent preeclampsia. Collectively, these findings provide mechanistic insights into the hypertensive disorders of pregnancy and have the potential to advance pregnancy risk stratification.

Polygenic risk score for problematic alcohol use predicts heavy drinking and alcohol use disorder symptoms in young adulthood after accounting for adolescent alcohol use and parental alcohol use disorder

BackgroundGenome-wide association studies (GWAS) have identified hundreds of loci associated with alcohol-related traits. GWAS permit the calculation of polygenic risk scores (PRS), which aggregate genetic risk for a trait across the genome. To evaluate the usefulness of a PRS for problematic alcohol use (PAU)—which subsumes alcohol use disorder (AUD) and alcohol-related problems—we tested whether this PRS predicted heavy drinking and alcohol problems after accounting for family history of AUD and prior drinking history, robust and established predictors of PAU. MethodsParticipants (N=665) were European-ancestry members of the Michigan Longitudinal Study, a prospective family study with high rates (65%) of parental AUD. Participants reported their frequency of alcohol use, maximum drinks consumed in a 24-hour period, and alcohol use problems at four assessments in adolescence and young adulthood (11–29 years old). We used polygenic prediction via Bayesian regression and continuous shrinkage priors to create a PAU PRS using summary statistics from a meta-GWAS of PAU. ResultsAfter adjusting for demographic covariates, parental AUD, and drinking and alcohol use problems in early and mid/late adolescence, the PAU PRS was significantly associated with alcohol-related problems in young adulthood (β=.08, p=.047; R2=0.6%). The PAU PRS also had a significant indirect effect on alcohol use problems in young adulthood through earlier drinking and alcohol use problems (β=.02, p=.03). ConclusionsThe PAU PRS predicted alcohol problems in young adulthood after accounting for parental history of AUD and alcohol use in adolescence, providing evidence that genetic data uniquely inform the etiology of alcohol problems.

Polygenic overlap with body-mass index improves prediction of treatment-resistant schizophrenia

Treatment resistant schizophrenia (TRS) is characterized by repeated treatment failure with antipsychotics. A recent genome-wide association study (GWAS) of TRS showed a polygenic architecture, but no significant loci were identified. Clozapine is shown to be the superior drug in terms of clinical effect in TRS; at the same time it has a serious side effect profile, including weight gain. Here, we sought to increase power for genetic discovery and improve polygenic prediction of TRS, by leveraging genetic overlap with Body Mass Index (BMI). We analysed GWAS summary statistics for TRS and BMI applying the conditional false discovery rate (cFDR) framework. We observed cross-trait polygenic enrichment for TRS conditioned on associations with BMI. Leveraging this cross-trait enrichment, we identified 2 novel loci for TRS at cFDR <0.01, suggesting a role of MAP2K1 and ZDBF2. Further, polygenic prediction based on the cFDR analysis explained more variance in TRS when compared to the standard TRS GWAS. These findings highlight putative molecular pathways which may distinguish TRS patients from treatment responsive patients. Moreover, these findings confirm that shared genetic mechanisms influence both TRS and BMI and provide new insights into the biological underpinnings of metabolic dysfunction and antipsychotic treatment.

Leveraging related health phenotypes for polygenic prediction of impulsive choice, impulsive action, and impulsive personality traits in 1534 European ancestry community adults.

Impulsivity refers to a number of conceptually related phenotypes reflecting self-regulatory capacity that are considered promising endophenotypes for mental and physical health. Measures of impulsivity can be broadly grouped into three domains, namely, impulsive choice, impulsive action, and impulsive personality traits. In a community-based sample of ancestral Europeans (n = 1534), we conducted genome-wide association studies (GWASs) of impulsive choice (delay discounting), impulsive action (behavioral inhibition), and impulsive personality traits (UPPS-P), and evaluated 11 polygenic risk scores (PRSs) of phenotypes previously linked to self-regulation. Although there were no individual genome-wide significant hits, the neuroticism PRS was positively associated with negative urgency (adjusted R2 = 1.61%, p = 3.6 × 10-7 ) and the educational attainment PRS was inversely associated with delay discounting (adjusted R2 = 1.68%, p = 2.2 × 10-7 ). There was also evidence implicating PRSs of attention-deficit/hyperactivity disorder, externalizing, risk-taking, smoking cessation, smoking initiation, and body mass index with one or more impulsivity phenotypes (adjusted R2 s: 0.35%-1.07%; FDR adjusted ps = 0.05-0.0006). These significant associations between PRSs and impulsivity phenotypes are consistent with established genetic correlations. The combined PRS explained 0.91%-2.46% of the phenotypic variance for individual impulsivity measures, corresponding to 8.7%-32.5% of their reported single-nucleotide polymorphism (SNP)-based heritability, suggesting a non-negligible portion of the SNP-based heritability can be recovered by PRSs. These results support the predictive validity and utility of PRSs, even derived from related phenotypes, to inform the genetics of impulsivity phenotypes.

Polygenic prediction of bipolar disorder in a Latin American sample.

To date, bipolar disorder (BD) genetic studies and polygenic risk scores (PRSs) for BD are based primarily on populations of European descent (EUR) and lack representation from other ancestries including Latin American (LAT). Here, we describe a new LAT cohort from the Mayo Clinic Bipolar Biobank (MCBB), a multisite collaboration with recruitment sites in the United States (EUR; 1,443 cases and 777 controls) and Mexico and Chile (LAT; 211 cases and 161 controls) and use the sample to explore the performance of a BD-PRS in a LAT population. Using results from the largest genome-wide association study of BD in EUR individuals, PRSice2 and LDpred2 were used to compute BD-PRSs in the LAT and EUR samples from the MCBB. PRSs explained up to 1.4% (PRSice) and 4% (LDpred2) of the phenotypic variance on the liability scale in the LAT sample compared to 3.8% (PRSice2) and 3.4% (LDpred2) in the EUR samples. Future larger studies should further explore the differential performance of different PRS approaches across ancestries. International multisite studies, such as this one, have the potential to address diversity-related limitations of prior genomic studies and ultimately contribute to the reduction of health disparities.

Genetic diversity fuels gene discovery for tobacco and alcohol use

Tobacco and alcohol use are heritable behaviours associated with 15% and 5.3% of worldwide deaths, respectively, due largely to broad increased risk for disease and injury1–4. These substances are used across the globe, yet genome-wide association studies have focused largely on individuals of European ancestries5. Here we leveraged global genetic diversity across 3.4 million individuals from four major clines of global ancestry (approximately 21% non-European) to power the discovery and fine-mapping of genomic loci associated with tobacco and alcohol use, to inform function of these loci via ancestry-aware transcriptome-wide association studies, and to evaluate the genetic architecture and predictive power of polygenic risk within and across populations. We found that increases in sample size and genetic diversity improved locus identification and fine-mapping resolution, and that a large majority of the 3,823 associated variants (from 2,143 loci) showed consistent effect sizes across ancestry dimensions. However, polygenic risk scores developed in one ancestry performed poorly in others, highlighting the continued need to increase sample sizes of diverse ancestries to realize any potential benefit of polygenic prediction.

The contribution of common and rare genetic variants to variation in metabolic traits in 288,137 East Asians

Metabolic traits are heritable phenotypes widely-used in assessing the risk of various diseases. We conduct a genome-wide association analysis (GWAS) of nine metabolic traits (including glycemic, lipid, liver enzyme levels) in 125,872 Korean subjects genotyped with the Korea Biobank Array. Following meta-analysis with GWAS from Biobank Japan identify 144 novel signals (MAF ≥ 1%), of which 57.0% are replicated in UK Biobank. Additionally, we discover 66 rare (MAF < 1%) variants, 94.4% of them co-incident to common loci, adding to allelic series. Although rare variants have limited contribution to overall trait variance, these lead, in carriers, substantial loss of predictive accuracy from polygenic predictions of disease risk from common variant alone. We capture groups with up to 16-fold variation in type 2 diabetes (T2D) prevalence by integration of genetic risk scores of fasting plasma glucose and T2D and the I349F rare protective variant. This study highlights the need to consider the joint contribution of both common and rare variants on inherited risk of metabolic traits and related diseases.

Genetic determinants of polygenic prediction accuracy within a population.

Genomic risk prediction is on the emerging path toward personalized medicine. However, the accuracy of polygenic prediction varies strongly in different individuals. Based on up to 352,277 European ancestry participants in the UK Biobank, we constructed polygenic risk scores for 15 physiological and biochemical quantitative traits. We identified a total of 185 polygenic prediction variability quantitative trait loci for 11 traits by Levene's test among 254,376 unrelated individuals. We validated the effects of prediction variability quantitative trait loci using an independent test set of 58,927 individuals. For instance, a score aggregating 51 prediction variability quantitative trait locus variants for triglycerides had the strongest Spearman correlation of 0.185 (P-value <1.0 × 10-300) with the squared prediction errors. We found a strong enrichment of complex genetic effects conferred by prediction variability quantitative trait loci compared to risk loci identified in genome-wide association studies, including 89 prediction variability quantitative trait loci exhibiting dominance effects. Incorporation of dominance effects into polygenic risk scores significantly improved polygenic prediction for triglycerides, low-density lipoprotein cholesterol, vitamin D, and platelet. In conclusion, we have discovered and profiled genetic determinants of polygenic prediction variability for 11 quantitative biomarkers. These findings may assist interpretation of genomic risk prediction in various contexts and encourage novel approaches for constructing polygenic risk scores with complex genetic effects.

Polygenic power calculator: Statistical power and polygenic prediction accuracy of genome-wide association studies of complex traits.

Power calculation is a necessary step when planning genome-wide association studies (GWAS) to ensure meaningful findings. Statistical power of GWAS depends on the genetic architecture of phenotype, sample size, and study design. While several computer programs have been developed to perform power calculation for single SNP association testing, it might be more appropriate for GWAS power calculation to address the probability of detecting any number of associated SNPs. In this paper, we derive the statistical power distribution across causal SNPs under the assumption of a point-normal effect size distribution. We demonstrate how key outcome indices of GWAS are related to the genetic architecture (heritability and polygenicity) of the phenotype through the power distribution. We also provide a fast, flexible and interactive power calculation tool which generates predictions for key GWAS outcomes including the number of independent significant SNPs, the phenotypic variance explained by these SNPs, and the predictive accuracy of resulting polygenic scores. These results could also be used to explore the future behaviour of GWAS as sample sizes increase further. Moreover, we present results from simulation studies to validate our derivation and evaluate the agreement between our predictions and reported GWAS results.

Testing the generalizability of ancestry-specific polygenic risk scores to predict prostate cancer in sub-Saharan Africa

BackgroundGenome-wide association studies do not always replicate well across populations, limiting the generalizability of polygenic risk scores (PRS). Despite higher incidence and mortality rates of prostate cancer in men of African descent, much of what is known about cancer genetics comes from populations of European descent. To understand how well genetic predictions perform in different populations, we evaluated test characteristics of PRS from three previous studies using data from the UK Biobank and a novel dataset of 1298 prostate cancer cases and 1333 controls from Ghana, Nigeria, Senegal, and South Africa.ResultsAllele frequency differences cause predicted risks of prostate cancer to vary across populations. However, natural selection is not the primary driver of these differences. Comparing continental datasets, we find that polygenic predictions of case vs. control status are more effective for European individuals (AUC 0.608–0.707, OR 2.37–5.71) than for African individuals (AUC 0.502–0.585, OR 0.95–2.01). Furthermore, PRS that leverage information from African Americans yield modest AUC and odds ratio improvements for sub-Saharan African individuals. These improvements were larger for West Africans than for South Africans. Finally, we find that existing PRS are largely unable to predict whether African individuals develop aggressive forms of prostate cancer, as specified by higher tumor stages or Gleason scores.ConclusionsGenetic predictions of prostate cancer perform poorly if the study sample does not match the ancestry of the original GWAS. PRS built from European GWAS may be inadequate for application in non-European populations and perpetuate existing health disparities.

Non-linear machine learning models incorporating SNPs and PRS improve polygenic prediction in diverse human populations

Polygenic risk scores (PRS) are commonly used to quantify the inherited susceptibility for a trait, yet they fail to account for non-linear and interaction effects between single nucleotide polymorphisms (SNPs). We address this via a machine learning approach, validated in nine complex phenotypes in a multi-ancestry population. We use an ensemble method of SNP selection followed by gradient boosted trees (XGBoost) to allow for non-linearities and interaction effects. We compare our results to the standard, linear PRS model developed using PRSice, LDpred2, and lassosum2. Combining a PRS as a feature in an XGBoost model results in a relative increase in the percentage variance explained compared to the standard linear PRS model by 22% for height, 27% for HDL cholesterol, 43% for body mass index, 50% for sleep duration, 58% for systolic blood pressure, 64% for total cholesterol, 66% for triglycerides, 77% for LDL cholesterol, and 100% for diastolic blood pressure. Multi-ancestry trained models perform similarly to specific racial/ethnic group trained models and are consistently superior to the standard linear PRS models. This work demonstrates an effective method to account for non-linearities and interaction effects in genetics-based prediction models.

Polygenic influences associated with adolescent cognitive skills

Genes play an important role in children's cognitive ability through adolescence and into adulthood. Recent advances in genomics have enabled us to test the effect of various genetic predispositions on measured cognitive outcomes. Here, we leveraged summary statistics from the most recent genome-wide association studies of eleven cognitive and mental health traits to build polygenic prediction models of measured intelligence and academic skills in a cohort of Australian adolescent twins (N = 2335, 57% female). We show that polygenic risk scores for educational attainment, intelligence, and cognitive performance explained up to 10% of the variance in academic skills and 7% in intelligence test scores in our cohort. Additionally, we found that a genetic predisposition for ADHD was negatively associated with all cognitive measures and a genetic predisposition for schizophrenia was negatively associated with performance IQ but no other cognitive measure. In this study, we provide evidence that a genetic vulnerability to some mental health disorders is associated with poorer performance on a variety of cognitive and academic tests, regardless of whether the individual has developed the disorder.