Probabilistic inference of the genetic architecture underlying functional enrichment of complex traits

Evaluation of heritability partitioning approaches in livestock populations

BACKGROUND

Heritability partitioning approaches estimate the contribution of different functional classes, such as coding or regulatory variants, to the genetic variance. This information allows a better understanding of the genetic architecture of complex traits, including complex diseases, but can also help improve the accuracy of genomic selection in livestock species. However, methods have mainly been tested on human genomic data, whereas livestock populations have specific characteristics, such as high levels of relatedness, small effective population size or long-range levels of linkage disequilibrium.

RESULTS

Here, we used data from 14,762 cows, imputed at the whole-genome sequence level for 11,537,240 variants, to simulate traits in a typical livestock population and evaluate the accuracy of two state-of-the-art heritability partitioning methods, GREML and a Bayesian mixture model. In simulations where a single functional class had increased contribution to heritability, we observed that the estimators were unbiased but had low precision. When causal variants were enriched in variants with low (< 0.05) or high (> 0.20) minor allele frequency or low (below 1st quartile) or high (above 3rd quartile) linkage disequilibrium scores, it was necessary to partition the genetic variance into multiple classes defined on the basis of allele frequencies or LD scores to obtain unbiased results. When multiple functional classes had variable contributions to heritability, estimators showed higher levels of variation and confounding between certain categories was observed. In addition, estimators from small categories were particularly imprecise. However, the estimates and their ranking were still informative about the contribution of the classes. We also demonstrated that using methods that estimate the contribution of a single category at a time, a commonly used approach, results in an overestimation. Finally, we applied the methods to phenotypes for muscular development and height and estimated that, on average, variants in open chromatin regions had a higher contribution to the genetic variance (> 45%), while variants in coding regions had the strongest individual effects (> 25-fold enrichment on average). Conversely, variants in intergenic or intronic regions showed lower levels of enrichment (0.2 and 0.6-fold on average, respectively).

CONCLUSIONS

Heritability partitioning approaches should be used cautiously in livestock populations, in particular for small categories. Two-component approaches that fit only one functional category at a time lead to biased estimators and should not be used.

Improving the accuracy of genomic prediction in dairy cattle using the biologically annotated neural networks framework

BACKGROUND

Biologically annotated neural networks (BANNs) are feedforward Bayesian neural network models that utilize partially connected architectures based on SNP-set annotations. As an interpretable neural network, BANNs model SNP and SNP-set effects in their input and hidden layers, respectively. Furthermore, the weights and connections of the network are regarded as random variables with prior distributions reflecting the manifestation of genetic effects at various genomic scales. However, its application in genomic prediction has yet to be explored.

RESULTS

This study extended the BANNs framework to the area of genomic selection and explored the optimal SNP-set partitioning strategies by using dairy cattle datasets. The SNP-sets were partitioned based on two strategies-gene annotations and 100 kb windows, denoted as BANN_gene and BANN_100kb, respectively. The BANNs model was compared with GBLUP, random forest (RF), BayesB and BayesCπ through five replicates of five-fold cross-validation using genotypic and phenotypic data on milk production traits, type traits, and one health trait of 6,558, 6,210 and 5,962 Chinese Holsteins, respectively. Results showed that the BANNs framework achieves higher genomic prediction accuracy compared to GBLUP, RF and Bayesian methods. Specifically, the BANN_100kb demonstrated superior accuracy and the BANN_gene exhibited generally suboptimal accuracy compared to GBLUP, RF, BayesB and BayesCπ across all traits. The average accuracy improvements of BANN_100kb over GBLUP, RF, BayesB and BayesCπ were 4.86%, 3.95%, 3.84% and 1.92%, and the accuracy of BANN_gene was improved by 3.75%, 2.86%, 2.73% and 0.85% compared to GBLUP, RF, BayesB and BayesCπ, respectively across all seven traits. Meanwhile, both BANN_100kb and BANN_gene yielded lower overall mean square error values than GBLUP, RF and Bayesian methods.

CONCLUSION

Our findings demonstrated that the BANNs framework performed better than traditional genomic prediction methods in our tested scenarios, and might serve as a promising alternative approach for genomic prediction in dairy cattle.

Leveraging functional genomic annotations and genome coverage to improve polygenic prediction of complex traits within and between ancestries

We develop a method, SBayesRC, that integrates genome-wide association study (GWAS) summary statistics with functional genomic annotations to improve polygenic prediction of complex traits. Our method is scalable to whole-genome variant analysis and refines signals from functional annotations by allowing them to affect both causal variant probability and causal effect distribution. We analyze 50 complex traits and diseases using ∼7 million common single-nucleotide polymorphisms (SNPs) and 96 annotations. SBayesRC improves prediction accuracy by 14% in European ancestry and up to 34% in cross-ancestry prediction compared to the baseline method SBayesR, which does not use annotations, and outperforms other methods, including LDpred2, LDpred-funct, MegaPRS, PolyPred-S and PRS-CSx. Investigation of factors affecting prediction accuracy identifies a significant interaction between SNP density and annotation information, suggesting whole-genome sequence variants with annotations may further improve prediction. Functional partitioning analysis highlights a major contribution of evolutionary constrained regions to prediction accuracy and the largest per-SNP contribution from nonsynonymous SNPs.

Polygenic risk scores for cervical HPV infection, neoplasia and cancer show potential for personalised screening: comparison of two methods

The era of precision medicine requires the achievement of accurate risk assessment. Polygenic risk scores (PRSs) have strong potential for increasing the benefits of nationwide cancer screening programs. The current pool of evidence on the role of a PRS as a risk stratification model in actual practice and implementation is limited. To better understand the impact of possible method-induced variance, we constructed and validated two PRSs for cervical cancer (CC) using the Estonian Biobank female population (691 CC cases and 13,820 controls) and evaluated their utility in predicting incident cervical neoplasia (CIN), cancer, and human papillomavirus (HPV) infection using two methods (LDPred and BayesRR-RC). This study demonstrated that two genetic risk scores were significantly associated with CIN, CC, and HPV infection incidence. Independent of the method, we demonstrated that women with elevated PRS values reached the observed cumulative risk levels of CIN or CC much earlier. Our results indicated that the PRS-based discrimination rules could differ substantially when the PRSs contain similar predictive information. In summary, our analysis indicated that PRSs represent a personalized genetic component that could be an additional tool for cervical cancer risk stratification, and earlier detection of abnormalities provides invaluable information for those at high risk.

Gene expression and RNA splicing explain large proportions of the heritability for complex traits in cattle

Many quantitative trait loci (QTLs) are in non-coding regions. Therefore, QTLs are assumed to affect gene regulation. Gene expression and RNA splicing are primary steps of transcription, so DNA variants changing gene expression (eVariants) or RNA splicing (sVariants) are expected to significantly affect phenotypes. We quantify the contribution of eVariants and sVariants detected from 16 tissues (n = 4,725) to 37 traits of ∼120,000 cattle (average magnitude of genetic correlation between traits = 0.13). Analyzed in Bayesian mixture models, averaged across 37 traits, cis and trans eVariants and sVariants detected from 16 tissues jointly explain 69.2% (SE = 0.5%) of heritability, 44% more than expected from the same number of random variants. This 69.2% includes an average of 24% from trans e-/sVariants (14% more than expected). Averaged across 56 lipidomic traits, multi-tissue cis and trans e-/sVariants also explain 71.5% (SE = 0.3%) of heritability, demonstrating the essential role of proximal and distal regulatory variants in shaping mammalian phenotypes.

Genetic insights into the age-specific biological mechanisms governing human ovarian aging

There is currently little evidence that the genetic basis of human phenotype varies significantly across the lifespan. However, time-to-event phenotypes are understudied and can be thought of as reflecting an underlying hazard, which is unlikely to be constant through life when values take a broad range. Here, we find that 74% of 245 genome-wide significant genetic associations with age at natural menopause (ANM) in the UK Biobank show a form of age-specific effect. Nineteen of these replicated discoveries are identified only by our modeling framework, which determines the time dependency of DNA-variant age-at-onset associations without a significant multiple-testing burden. Across the range of early to late menopause, we find evidence for significantly different underlying biological pathways, changes in the signs of genetic correlations of ANM to health indicators and outcomes, and differences in inferred causal relationships. We find that DNA damage response processes only act to shape ovarian reserve and depletion for women of early ANM. Genetically mediated delays in ANM were associated with increased relative risk of breast cancer and leiomyoma at all ages and with high cholesterol and heart failure for late-ANM women. These findings suggest that a better understanding of the age dependency of genetic risk factor relationships among health indicators and outcomes is achievable through appropriate statistical modeling of large-scale biobank data.

Polygenic Risk Score Predicts Modified Risk in BRCA1 Pathogenic Variant c.4035del and c.5266dup Carriers in Breast Cancer Patients

The aim of this study was to assess the power of the polygenic risk score (PRS) in estimating the overall genetic risk of women carrying germline BRCA1 pathogenic variants (PVs) c.4035del or c.5266dup to develop breast (BC) or ovarian cancer (OC) due to additional genetic variations. In this study, PRSs previously developed from two joint models using summary statistics of age-at-onset (BayesW model) and case-control data (BayesRR-RC model) from a genome-wide association analysis (GWAS) were applied to 406 germline BRCA1 PV (c.4035del or c.5266dup) carriers affected by BC or OC, compared with unaffected individuals. A binomial logistic regression model was used to assess the association of PRS with BC or OC development risk. We observed that the best-fitting BayesW PRS model effectively predicted the individual's BC risk (OR = 1.37; 95% CI = 1.03-1.81, p = 0.02905 with AUC = 0.759). However, none of the applied PRS models was a good predictor of OC risk. The best-fitted PRS model (BayesW) contributed to assessing the risk of developing BC for germline BRCA1 PV (c.4035del or c.5266dup) carriers and may facilitate more precise and timely patient stratification and decision-making to improve the current BC treatment or even prevention strategies.

Liability-scale heritability estimation for biobank studies of low-prevalence disease

Theory for liability-scale models of the underlying genetic basis of complex disease provides an important way to interpret, compare, and understand results generated from biological studies. In particular, through estimation of the liability-scale heritability (LSH), liability models facilitate an understanding and comparison of the relative importance of genetic and environmental risk factors that shape different clinically important disease outcomes. Increasingly, large-scale biobank studies that link genetic information to electronic health records, containing hundreds of disease diagnosis indicators that mostly occur infrequently within the sample, are becoming available. Here, we propose an extension of the existing liability-scale model theory suitable for estimating LSH in biobank studies of low-prevalence disease. In a simulation study, we find that our derived expression yields lower mean square error (MSE) and is less sensitive to prevalence misspecification as compared to previous transformations for diseases with ≤2% population prevalence and LSH of ≤0.45, especially if the biobank sample prevalence is less than that of the wider population. Applying our expression to 13 diagnostic outcomes of ≤3% prevalence in the UK Biobank study revealed important differences in LSH obtained from the different theoretical expressions that impact the conclusions made when comparing LSH across disease outcomes. This demonstrates the importance of careful consideration for estimation and prediction of low-prevalence disease outcomes and facilitates improved inference of the underlying genetic basis of ≤2% population prevalence diseases, especially where biobank sample ascertainment results in a healthier sample population.

A saturated map of common genetic variants associated with human height

Common single-nucleotide polymorphisms (SNPs) are predicted to collectively explain 40-50% of phenotypic variation in human height, but identifying the specific variants and associated regions requires huge sample sizes1. Here, using data from a genome-wide association study of 5.4 million individuals of diverse ancestries, we show that 12,111 independent SNPs that are significantly associated with height account for nearly all of the common SNP-based heritability. These SNPs are clustered within 7,209 non-overlapping genomic segments with a mean size of around 90 kb, covering about 21% of the genome. The density of independent associations varies across the genome and the regions of increased density are enriched for biologically relevant genes. In out-of-sample estimation and prediction, the 12,111 SNPs (or all SNPs in the HapMap 3 panel2) account for 40% (45%) of phenotypic variance in populations of European ancestry but only around 10-20% (14-24%) in populations of other ancestries. Effect sizes, associated regions and gene prioritization are similar across ancestries, indicating that reduced prediction accuracy is likely to be explained by linkage disequilibrium and differences in allele frequency within associated regions. Finally, we show that the relevant biological pathways are detectable with smaller sample sizes than are needed to implicate causal genes and variants. Overall, this study provides a comprehensive map of specific genomic regions that contain the vast majority of common height-associated variants. Although this map is saturated for populations of European ancestry, further research is needed to achieve equivalent saturation in other ancestries.

Validating and automating learning of cardiometabolic polygenic risk scores from direct-to-consumer genetic and phenotypic data: implications for scaling precision health research

INTRODUCTION

A major challenge to enabling precision health at a global scale is the bias between those who enroll in state sponsored genomic research and those suffering from chronic disease. More than 30 million people have been genotyped by direct-to-consumer (DTC) companies such as 23andMe, Ancestry DNA, and MyHeritage, providing a potential mechanism for democratizing access to medical interventions and thus catalyzing improvements in patient outcomes as the cost of data acquisition drops. However, much of these data are sequestered in the initial provider network, without the ability for the scientific community to either access or validate. Here, we present a novel geno-pheno platform that integrates heterogeneous data sources and applies learnings to common chronic disease conditions including Type 2 diabetes (T2D) and hypertension.

METHODS

We collected genotyped data from a novel DTC platform where participants upload their genotype data files and were invited to answer general health questionnaires regarding cardiometabolic traits over a period of 6 months. Quality control, imputation, and genome-wide association studies were performed on this dataset, and polygenic risk scores were built in a case-control setting using the BASIL algorithm.

RESULTS

We collected data on N = 4,550 (389 cases / 4,161 controls) who reported being affected or previously affected for T2D and N = 4,528 (1,027 cases / 3,501 controls) for hypertension. We identified 164 out of 272 variants showing identical effect direction to previously reported genome-significant findings in Europeans. Performance metric of the PRS models was AUC = 0.68, which is comparable to previously published PRS models obtained with larger datasets including clinical biomarkers.

DISCUSSION

DTC platforms have the potential of inverting research models of genome sequencing and phenotypic data acquisition. Quality control (QC) mechanisms proved to successfully enable traditional GWAS and PRS analyses. The direct participation of individuals has shown the potential to generate rich datasets enabling the creation of PRS cardiometabolic models. More importantly, federated learning of PRS from reuse of DTC data provides a mechanism for scaling precision health care delivery beyond the small number of countries who can afford to finance these efforts directly.

CONCLUSIONS

The genetics of T2D and hypertension have been studied extensively in controlled datasets, and various polygenic risk scores (PRS) have been developed. We developed predictive tools for both phenotypes trained with heterogeneous genotypic and phenotypic data generated outside of the clinical environment and show that our methods can recapitulate prior findings with fidelity. From these observations, we conclude that it is possible to leverage DTC genetic repositories to identify individuals at risk of debilitating diseases based on their unique genetic landscape so that informed, timely clinical interventions can be incorporated.

Improving GWAS discovery and genomic prediction accuracy in biobank data

Genetically informed, deep-phenotyped biobanks are an important research resource and it is imperative that the most powerful, versatile, and efficient analysis approaches are used. Here, we apply our recently developed Bayesian grouped mixture of regressions model (GMRM) in the UK and Estonian Biobanks and obtain the highest genomic prediction accuracy reported to date across 21 heritable traits. When compared to other approaches, GMRM accuracy was greater than annotation prediction models run in the LDAK or LDPred-funct software by 15% (SE 7%) and 14% (SE 2%), respectively, and was 18% (SE 3%) greater than a baseline BayesR model without single-nucleotide polymorphism (SNP) markers grouped into minor allele frequency-linkage disequilibrium (MAF-LD) annotation categories. For height, the prediction accuracy R2 was 47% in a UK Biobank holdout sample, which was 76% of the estimated [Formula: see text]. We then extend our GMRM prediction model to provide mixed-linear model association (MLMA) SNP marker estimates for genome-wide association (GWAS) discovery, which increased the independent loci detected to 16,162 in unrelated UK Biobank individuals, compared to 10,550 from BoltLMM and 10,095 from Regenie, a 62 and 65% increase, respectively. The average [Formula: see text] value of the leading markers increased by 15.24 (SE 0.41) for every 1% increase in prediction accuracy gained over a baseline BayesR model across the traits. Thus, we show that modeling genetic associations accounting for MAF and LD differences among SNP markers, and incorporating prior knowledge of genomic function, is important for both genomic prediction and discovery in large-scale individual-level studies.

Blood-based epigenome-wide analyses of cognitive abilities

BACKGROUND

Blood-based markers of cognitive functioning might provide an accessible way to track neurodegeneration years prior to clinical manifestation of cognitive impairment and dementia.

RESULTS

Using blood-based epigenome-wide analyses of general cognitive function, we show that individual differences in DNA methylation (DNAm) explain 35.0% of the variance in general cognitive function (g). A DNAm predictor explains ~4% of the variance, independently of a polygenic score, in two external cohorts. It also associates with circulating levels of neurology- and inflammation-related proteins, global brain imaging metrics, and regional cortical volumes.

CONCLUSIONS

As sample sizes increase, the ability to assess cognitive function from DNAm data may be informative in settings where cognitive testing is unreliable or unavailable.