Managing differential performance of polygenic risk scores across groups: Real-world experience of the eMERGE Network.

The differential performance of polygenic risk scores (PRSs) by group is one of the major ethical barriers to their clinical use. It is also one of the main practical challenges for any implementation effort. The social repercussions of how people are grouped in PRS research must be considered in communications with research participants, including return of results. Here, we outline the decisions faced and choices made by a large multi-site clinical implementation study returning PRSs to diverse participants in handling this issue of differential performance. Our approach to managing the complexities associated with the differential performance of PRSs serves as a case study that can help future implementers of PRSs to plot an anticipatory course in response to this issue.

Opposing effects of genetic variation in MTCH2 for obesity versus heart failure.

Genetic variation in genes regulating metabolism may be advantageous in some settings but not others. The non-failing adult heart relies heavily on fatty acids as a fuel substrate and source of ATP. In contrast, the failing heart favors glucose as a fuel source. A bootstrap analysis for genes with deviant allele frequencies in cardiomyopathy cases versus controls identified the MTCH2 gene as having unusual variation. MTCH2 encodes an outer mitochondrial membrane protein, and prior genome-wide studies associated MTCH2 variants with body mass index, consistent with its role in metabolism. We identified the referent allele of rs1064608 (p.Pro290) as being overrepresented in cardiomyopathy cases compared to controls, and linkage disequilibrium analysis associated this variant with the MTCH2 cis eQTL rs10838738 and lower MTCH2 expression. To evaluate MTCH2, we knocked down Mtch in Drosophila heart tubes which produced a dilated and poorly functioning heart tube, reduced adiposity and shortened life span. Cardiac Mtch mutants generated more lactate at baseline, and they displayed impaired oxygen consumption in the presence of glucose but not palmitate. Treatment of cardiac Mtch mutants with dichloroacetate, a pyruvate dehydrogenase kinase inhibitor, reduced lactate and rescued lifespan. Deletion of MTCH2 in human cells similarly impaired oxygen consumption in the presence of glucose but not fatty acids. These data support a model in which MTCH2 reduction may be favorable when fatty acids are the major fuel source, favoring lean body mass. However, in settings like heart failure, where the heart shifts toward using more glucose, reduction of MTCH2 is maladaptive.

Partial and complete loss of myosin binding protein H-like cause cardiac conduction defects.

A premature truncation of MYBPHL in humans and a loss of Mybphl in mice is associated with dilated cardiomyopathy, atrial and ventricular arrhythmias, and atrial enlargement. MYBPHL encodes myosin binding protein H-like (MyBP-HL). Prior work in mice indirectly identified Mybphl expression in the atria and in small puncta throughout the ventricle. Because of its genetic association with human and mouse cardiac conduction system disease, we evaluated the anatomical localization of MyBP-HL and the consequences of loss of MyBP-HL on conduction system function. Immunofluorescence microscopy of normal adult mouse ventricles identified MyBP-HL-positive ventricular cardiomyocytes that co-localized with the ventricular conduction system marker contactin-2 near the atrioventricular node and in a subset of Purkinje fibers. Mybphl heterozygous ventricles had a marked reduction of MyBP-HL-positive cells compared to controls. Lightsheet microscopy of normal perinatal day 5 mouse hearts showed enrichment of MyBP-HL-positive cells within and immediately adjacent to the contactin-2-positive ventricular conduction system, but this association was not apparent in Mybphl heterozygous hearts. Surface telemetry of Mybphl-null mice revealed atrioventricular block and atrial bigeminy, while intracardiac pacing revealed a shorter atrial relative refractory period and atrial tachycardia. Calcium transient analysis of isolated Mybphl-null atrial cardiomyocytes demonstrated an increased heterogeneity of calcium release and faster rates of calcium release compared to wild type controls. Super-resolution microscopy of Mybphl heterozygous and homozygous null atrial cardiomyocytes showed ryanodine receptor disorganization compared to wild type controls. Abnormal calcium release, shorter atrial refractory period, and atrial dilation seen in Mybphl null, but not wild type control hearts, agree with the observed atrial arrhythmias, bigeminy, and atrial tachycardia, whereas the proximity of MyBP-HL-positive cells with the ventricular conduction system provides insight into how a predominantly atrial expressed gene contributes to ventricular arrhythmias and ventricular dysfunction.

Genetic Variation in Enhancers Modifies Cardiomyopathy Gene Expression and Progression.

BACKGROUND: Inherited cardiomyopathy associates with a range of phenotypes, mediated by genetic and nongenetic factors. Noninherited cardiomyopathy also displays varying progression and outcomes. Expression of cardiomyopathy genes is under the regulatory control of promoters and enhancers, and human genetic variation in promoters and enhancers may contribute to this variability. METHODS: We superimposed epigenomic profiling from hearts and cardiomyocytes, including promoter-capture chromatin conformation information, to identify enhancers for 2 cardiomyopathy genes, MYH7 and LMNA. Enhancer function was validated in human cardiomyocytes derived from induced pluripotent stem cells. We also conducted a genome-wide search to ascertain genomic variation in enhancers positioned to alter cardiac expression and correlated one of these variants to cardiomyopathy progression using biobank data. RESULTS: Multiple enhancers were identified and validated for LMNA and MYH7, including a key enhancer that regulates the switch from MYH6 expression to MYH7 expression. Deletion of this enhancer resulted in a dose-dependent increase in MYH6 and faster contractile rate in engineered heart tissues. We searched for genomic variation in enhancer sequences across the genome, with a focus on nucleotide changes that create or interrupt transcription factor binding sites. The sequence variant, rs875908, disrupts a T-Box Transcription Factor 5 binding motif and maps to an enhancer region 2 kilobases from the transcriptional start site of MYH7. Gene editing to remove the enhancer that harbors this variant markedly reduced MYH7 expression in human cardiomyocytes. Using biobank-derived data, rs875908 associated with longitudinal echocardiographic features of cardiomyopathy. CONCLUSIONS: Enhancers regulate cardiomyopathy gene expression, and genomic variation within these enhancer regions associates with cardiomyopathic progression over time. This integrated approach identified noncoding modifiers of cardiomyopathy and is applicable to other cardiac genes.

Genetically based atrial fibrillation: Current considerations for diagnosis and management.

Atrial fibrillation (AF) is the most common atrial arrhythmia and is subcategorized into numerous clinical phenotypes. Given its heterogeneity, investigations into the genetic mechanisms underlying AF have been pursued in recent decades, with predominant analyses focusing on early onset or lone AF. Linkage analyses, genome-wide association studies (GWAS), and single gene analyses have led to the identification of rare and common genetic variants associated with AF risk. Significant overlap with genetic variants implicated in dilated cardiomyopathy syndromes, including truncating variants of the sarcomere protein titin, have been identified through these analyses, in addition to other genes associated with cardiac structure and function. Despite this, widespread utilization of genetic testing in AF remains hindered by the unclear impact of genetic risk identification on clinical outcomes and the high prevalence of variants of unknown significance (VUS). However, genetic testing is a reasonable option for patients with early onset AF and in those with significant family history of arrhythmia. While many knowledge gaps remain, emerging data support genotyping to inform selection of AF therapeutics. In this review, we highlight the current understanding of the complex genetic basis of AF and explore the overlap of AF with inherited cardiomyopathy syndromes. We propose a set of criteria for clinical genetic testing in AF patients and outline future steps for the integration of genetics into AF care.

Mitochondrial cardiomyopathy and ventricular arrhythmias associated with biallelic variants in C1QBP.

Patients with biallelic mutations in the nuclear-encoded mitochondrial gene C1QBP/p32 have been described with syndromic features and autosomal recessive cardiomyopathy. We describe the clinical course in two siblings who developed cardiomyopathy and ventricular fibrillation in infancy. We provide genomic analysis and clinical-pathologic correlation. Both siblings had profound cardiac failure with ventricular arrhythmia. One child died suddenly. The second sibling survived resuscitation but required extracorporeal cardiopulmonary support and died shortly afterward. On cardiac autopsy, the left ventricle was hypertrophied in both children. Histological examination revealed prominent cardiomyocyte cytoplasmic clearing, and electron microscopy confirmed abnormal mitochondrial structure within cardiomyocytes. DNA sequencing revealed compound heterozygous variants in C1QBP (p.Thr40Asnfs*45 and p.Phe204Leu) in both children. Family segregation analysis demonstrated each variant was inherited from an unaffected, heterozygous parent. Inherited loss of C1QBP/p32 is associated with recessive cardiomyopathy, ventricular fibrillation, and sudden death in early life. Ultrastructural mitochondrial evaluation in the second child was similar to findings in engineered C1qbp-deficient mice. Rapid trio analysis can define rare biallelic variants in genes that may be implicated in sudden death and facilitate medical management and family planning. (184/200).

The eMERGE genotype set of 83,717 subjects imputed to ~40 million variants genome wide and association with the herpes zoster medical record phenotype.

The Electronic Medical Records and Genomics (eMERGE) network is a network of medical centers with electronic medical records linked to existing biorepository samples for genomic discovery and genomic medicine research. The network sought to unify the genetic results from 78 Illumina and Affymetrix genotype array batches from 12 contributing medical centers for joint association analysis of 83,717 human participants. In this report, we describe the imputation of eMERGE results and methods to create the unified imputed merged set of genome-wide variant genotype data. We imputed the data using the Michigan Imputation Server, which provides a missing single-nucleotide variant genotype imputation service using the minimac3 imputation algorithm with the Haplotype Reference Consortium genotype reference set. We describe the quality control and filtering steps used in the generation of this data set and suggest generalizable quality thresholds for imputation and phenotype association studies. To test the merged imputed genotype set, we replicated a previously reported chromosome 6 HLA-B herpes zoster (shingles) association and discovered a novel zoster-associated loci in an epigenetic binding site near the terminus of chromosome 3 (3p29).

Novel nesprin-1 mutations associated with dilated cardiomyopathy cause nuclear envelope disruption and defects in myogenesis.

Nesprins-1 and -2 are highly expressed in skeletal and cardiac muscle and together with SUN (Sad1p/UNC84)-domain containing proteins and lamin A/C form the LInker of Nucleoskeleton-and-Cytoskeleton (LINC) bridging complex at the nuclear envelope (NE). Mutations in nesprin-1/2 have previously been found in patients with autosomal dominant Emery-Dreifuss muscular dystrophy (EDMD) as well as dilated cardiomyopathy (DCM). In this study, three novel rare variants (R8272Q, S8381C and N8406K) in the C-terminus of the SYNE1 gene (nesprin-1) were identified in seven DCM patients by mutation screening. Expression of these mutants caused nuclear morphology defects and reduced lamin A/C and SUN2 staining at the NE. GST pull-down indicated that nesprin-1/lamin/SUN interactions were disrupted. Nesprin-1 mutations were also associated with augmented activation of the ERK pathway in vitro and in hearts in vivo. During C2C12 muscle cell differentiation, nesprin-1 levels are increased concomitantly with kinesin light chain (KLC-1/2) and immunoprecipitation and GST pull-down showed that these proteins interacted via a recently identified LEWD domain in the C-terminus of nesprin-1. Expression of nesprin-1 mutants in C2C12 cells caused defects in myoblast differentiation and fusion associated with dysregulation of myogenic transcription factors and disruption of the nesprin-1 and KLC-1/2 interaction at the outer nuclear membrane. Expression of nesprin-1α2 WT and mutants in zebrafish embryos caused heart developmental defects that varied in severity. These findings support a role for nesprin-1 in myogenesis and muscle disease, and uncover a novel mechanism whereby disruption of the LINC complex may contribute to the pathogenesis of DCM.

Experimental Modeling Supports a Role for MyBP-HL as a Novel Myofilament Component in Arrhythmia and Dilated Cardiomyopathy.

BACKGROUND: Cardiomyopathy and arrhythmias are under significant genetic influence. Here, we studied a family with dilated cardiomyopathy and associated conduction system disease in whom prior clinical cardiac gene panel testing was unrevealing. METHODS: Whole-genome sequencing and induced pluripotent stem cells were used to examine a family with dilated cardiomyopathy and atrial and ventricular arrhythmias. We also characterized a mouse model with heterozygous and homozygous deletion of Mybphl. RESULTS: Whole-genome sequencing identified a premature stop codon, R255X, in the MYBPHL gene encoding MyBP-HL (myosin-binding protein-H like), a novel member of the myosin-binding protein family. MYBPHL was found to have high atrial expression with low ventricular expression. We determined that MyBP-HL protein was myofilament associated in the atria, and truncated MyBP-HL protein failed to incorporate into the myofilament. Human cell modeling demonstrated reduced expression from the mutant MYBPHL allele. Echocardiography of Mybphl heterozygous and null mouse hearts exhibited a 36% reduction in fractional shortening and an increased diastolic ventricular chamber size. Atria weight normalized to total heart weight was significantly increased in Mybphl heterozygous and null mice. Using a reporter system, we detected robust expression of Mybphl in the atria, and in discrete puncta throughout the right ventricular wall and septum, as well. Telemetric electrocardiogram recordings in Mybphl mice revealed cardiac conduction system abnormalities with aberrant atrioventricular conduction and an increased rate of arrhythmia in heterozygous and null mice. CONCLUSIONS: The findings of reduced ventricular function and conduction system defects in Mybphl mice support that MYBPHL truncations may increase risk for human arrhythmias and cardiomyopathy.

Disruption of the lamin A and matrin-3 interaction by myopathic LMNA mutations.

The nuclear face of the nuclear membrane is enriched with the intermediate filament protein lamin A. Mutations in LMNA, the gene encoding lamin A, lead to a diverse set of inherited conditions including myopathies that affect both the heart and skeletal muscle. To gain insight about lamin A protein interactions, binding proteins associated with the tail of lamin A were characterized. Of 130 nuclear proteins found associated with the lamin A tail, 17 (13%) were previously described lamin A binding partners. One protein not previously linked to lamin A, matrin-3, was selected for further study, because like LMNA mutations, matrin-3 has also been implicated in inherited myopathy. Matrin-3 binds RNA and DNA and is a nucleoplasmic protein originally identified from the insoluble nuclear fraction, referred to as the nuclear matrix. Anti-matrin-3 antibodies were found to co-immunoprecipitate lamin A, and the lamin-A binding domain was mapped to the carboxy-terminal half of matrin-3. Three-dimensional mapping of the lamin A-matrin-3 interface showed that the LMNA truncating mutation Δ303, which lacks the matrin-3 binding domain, was associated with an increased distance between lamin A and matrin-3. LMNA mutant cells are known to have altered biophysical properties and the matrin-3-lamin A interface is positioned to contribute to these defects.

Supercomputing for the parallelization of whole genome analysis.

MOTIVATION: The declining cost of generating DNA sequence is promoting an increase in whole genome sequencing, especially as applied to the human genome. Whole genome analysis requires the alignment and comparison of raw sequence data, and results in a computational bottleneck because of limited ability to analyze multiple genomes simultaneously. RESULTS: We now adapted a Cray XE6 supercomputer to achieve the parallelization required for concurrent multiple genome analysis. This approach not only markedly speeds computational time but also results in increased usable sequence per genome. Relying on publically available software, the Cray XE6 has the capacity to align and call variants on 240 whole genomes in ∼50 h. Multisample variant calling is also accelerated. AVAILABILITY AND IMPLEMENTATION: The MegaSeq workflow is designed to harness the size and memory of the Cray XE6, housed at Argonne National Laboratory, for whole genome analysis in a platform designed to better match current and emerging sequencing volume.

Genetic Variation in Cardiomyopathy and Cardiovascular Disorders.

With the wider deployment of massively-parallel, next-generation sequencing, it is now possible to survey human genome data for research and clinical purposes. The reduced cost of producing short-read sequencing has now shifted the burden to data analysis. Analysis of genome sequencing remains challenged by the complexity of the human genome, including redundancy and the repetitive nature of genome elements and the large amount of variation in individual genomes. Public databases of human genome sequences greatly facilitate interpretation of common and rare genetic variation, although linking database sequence information to detailed clinical information is limited by privacy and practical issues. Genetic variation is a rich source of knowledge for cardiovascular disease because many, if not all, cardiovascular disorders are highly heritable. The role of rare genetic variation in predicting risk and complications of cardiovascular diseases has been well established for hypertrophic and dilated cardiomyopathy, where the number of genes that are linked to these disorders is growing. Bolstered by family data, where genetic variants segregate with disease, rare variation can be linked to specific genetic variation that offers profound diagnostic information. Understanding genetic variation in cardiomyopathy is likely to help stratify forms of heart failure and guide therapy. Ultimately, genetic variation may be amenable to gene correction and gene editing strategies.

Identification of Clinical Drivers of Left Atrial Enlargement Through Genomics of Left Atrial Size.

BACKGROUND: Greater left atrial size is associated with a higher incidence of cardiovascular disease and mortality, but the full spectrum of diagnoses associated with left atrial enlargement in sex-stratified clinical populations is not well known. Our study sought to identify genetic risk mechanisms affecting left atrial diameter (LAD) in a clinical cohort. METHODS: Using Vanderbilt deidentified electronic health record, we studied 6163 females and 5993 males of European ancestry who had at least 1 LAD measure and available genotyping. A sex-stratified polygenic score was constructed for LAD variation and tested for association against 1680 International Classification of Diseases code-based phenotypes. Two-sample univariable and multivariable Mendelian randomization approaches were used to assess etiologic relationships between candidate associations and LAD. RESULTS: A phenome-wide association study identified 25 International Classification of Diseases code-based diagnoses in females and 11 in males associated with a polygenic score of LAD (false discovery rate q<0.01), 5 of which were further evaluated by Mendelian randomization (waist circumference [WC], atrial fibrillation, heart failure, systolic blood pressure, and coronary artery disease). Sex-stratified differences in the genetic associations between risk factors and a polygenic score for LAD were observed (WC for females; heart failure, systolic blood pressure, atrial fibrillation, and WC for males). By multivariable Mendelian randomization, higher WC remained significantly associated with larger LAD in females, whereas coronary artery disease, WC, and atrial fibrillation remained significantly associated with larger LAD in males. CONCLUSIONS: In a clinical population, we identified, by genomic approaches, potential etiologic risk factors for larger LAD. Further studies are needed to confirm the extent to which these risk factors may be modified to prevent or reverse adverse left atrial remodeling and the extent to which sex modifies these risk factors.

Reduction of Filamin C Results in Altered Proteostasis, Cardiomyopathy, and Arrhythmias.

BACKGROUND: Many cardiomyopathy-associated FLNC pathogenic variants are heterozygous truncations, and FLNC pathogenic variants are associated with arrhythmias. Arrhythmia triggers in filaminopathy are incompletely understood. METHODS AND RESULTS: We describe an individual with biallelic FLNC pathogenic variants, p.Arg650X and c.970-4A>G, with peripartum cardiomyopathy and ventricular arrhythmias. We also describe clinical findings in probands with FLNC variants including Val2715fs87X, Glu2458Serfs71X, Phe106Leu, and c.970-4A>G with hypertrophic and dilated cardiomyopathy, atrial fibrillation, and ventricular tachycardia. Induced pluripotent stem cell-derived cardiomyocytes (iPSC-CMs) were generated. The FLNC truncation, Arg650X/c.970-4A>G, showed a marked reduction in filamin C protein consistent with biallelic loss of function mutations. To assess loss of filamin C, gene editing of a healthy control iPSC line was used to generate a homozygous FLNC disruption in the actin binding domain. Because filamin C has been linked to protein quality control, we assessed the necessity of filamin C in iPSC-CMs for response to the proteasome inhibitor bortezomib. After exposure to low-dose bortezomib, FLNC-null iPSC-CMs showed an increase in the chaperone proteins BAG3, HSP70 (heat shock protein 70), and HSPB8 (small heat shock protein B8) and in the autophagy marker LC3I/II. FLNC null iPSC-CMs had prolonged electric field potential, which was further prolonged in the presence of low-dose bortezomib. FLNC null engineered heart tissues had impaired function after low-dose bortezomib. CONCLUSIONS: FLNC pathogenic variants associate with a predisposition to arrhythmias, which can be modeled in iPSC-CMs. Reduction of filamin C prolonged field potential, a surrogate for action potential, and with bortezomib-induced proteasome inhibition, reduced filamin C led to greater arrhythmia potential and impaired function.

Susceptibility to innate immune activation in genetically mediated myocarditis.

Myocarditis is clinically characterized by chest pain, arrhythmias, and heart failure, and treatment is often supportive. Mutations in DSP, a gene encoding the desmosomal protein desmoplakin, have been increasingly implicated in myocarditis. To model DSP-associated myocarditis and assess the role of innate immunity, we generated engineered heart tissues (EHTs) using human induced pluripotent stem cell-derived cardiomyocytes (hiPSC-CMs) from patients with heterozygous DSP truncating variants (DSPtvs) and a gene-edited homozygous deletion cell line (DSP-/-). At baseline, DSP-/- EHTs displayed a transcriptomic signature of innate immune activation, which was mirrored by cytokine release. Importantly, DSP-/- EHTs were hypersensitive to Toll-like receptor (TLR) stimulation, demonstrating more contractile dysfunction compared with isogenic controls. Relative to DSP-/- EHTs, heterozygous DSPtv EHTs had less functional impairment. DSPtv EHTs displayed heightened sensitivity to TLR stimulation, and when subjected to strain, DSPtv EHTs developed functional deficits, indicating reduced contractile reserve compared with healthy controls. Colchicine or NF-κB inhibitors improved strain-induced force deficits in DSPtv EHTs. Genomic correction of DSP p.R1951X using adenine base editing reduced inflammatory biomarker release from EHTs. Thus, EHTs replicate electrical and contractile phenotypes seen in human myocarditis, implicating cytokine release as a key part of the myogenic susceptibility to inflammation. The heightened innate immune activation and sensitivity are targets for clinical intervention.

The impact of damaging epilepsy and cardiac genetic variant burden in sudden death in the young.

BACKGROUND: Sudden unexpected death in children is a tragic event. Understanding the genetics of sudden death in the young (SDY) enables family counseling and cascade screening. The objective of this study was to characterize genetic variation in an SDY cohort using whole genome sequencing. METHODS: The SDY Case Registry is a National Institutes of Health/Centers for Disease Control and Prevention surveillance effort to discern the prevalence, causes, and risk factors for SDY. The SDY Case Registry prospectively collected clinical data and DNA biospecimens from SDY cases < 20 years of age. SDY cases were collected from medical examiner and coroner offices spanning 13 US jurisdictions from 2015 to 2019. The cohort included 211 children (median age 0.33 year; range 0-20 years), determined to have died suddenly and unexpectedly and from whom DNA biospecimens for DNA extractions and next-of-kin consent were ascertained. A control cohort consisted of 211 randomly sampled, sex- and ancestry-matched individuals from the 1000 Genomes Project. Genetic variation was evaluated in epilepsy, cardiomyopathy, and arrhythmia genes in the SDY and control cohorts. American College of Medical Genetics/Genomics guidelines were used to classify variants as pathogenic or likely pathogenic. Additionally, pathogenic and likely pathogenic genetic variation was identified using a Bayesian-based artificial intelligence (AI) tool. RESULTS: The SDY cohort was 43% European, 29% African, 3% Asian, 16% Hispanic, and 9% with mixed ancestries and 39% female. Six percent of the cohort was found to harbor a pathogenic or likely pathogenic genetic variant in an epilepsy, cardiomyopathy, or arrhythmia gene. The genomes of SDY cases, but not controls, were enriched for rare, potentially damaging variants in epilepsy, cardiomyopathy, and arrhythmia-related genes. A greater number of rare epilepsy genetic variants correlated with younger age at death. CONCLUSIONS: While damaging cardiomyopathy and arrhythmia genes are recognized contributors to SDY, we also observed an enrichment in epilepsy-related genes in the SDY cohort and a correlation between rare epilepsy variation and younger age at death. These findings emphasize the importance of considering epilepsy genes when evaluating SDY.

Risk factors affecting polygenic score performance across diverse cohorts.

Apart from ancestry, personal or environmental covariates may contribute to differences in polygenic score (PGS) performance. We analyzed effects of covariate stratification and interaction on body mass index (BMI) PGS (PGSBMI) across four cohorts of European (N=491,111) and African (N=21,612) ancestry. Stratifying on binary covariates and quintiles for continuous covariates, 18/62 covariates had significant and replicable R2 differences among strata. Covariates with the largest differences included age, sex, blood lipids, physical activity, and alcohol consumption, with R2 being nearly double between best and worst performing quintiles for certain covariates. 28 covariates had significant PGSBMI-covariate interaction effects, modifying PGSBMI effects by nearly 20% per standard deviation change. We observed overlap between covariates that had significant R2 differences among strata and interaction effects - across all covariates, their main effects on BMI were correlated with their maximum R2 differences and interaction effects (0.56 and 0.58, respectively), suggesting high-PGSBMI individuals have highest R2 and increase in PGS effect. Using quantile regression, we show the effect of PGSBMI increases as BMI itself increases, and that these differences in effects are directly related to differences in R2 when stratifying by different covariates. Given significant and replicable evidence for context-specific PGSBMI performance and effects, we investigated ways to increase model performance taking into account non-linear effects. Machine learning models (neural networks) increased relative model R2 (mean 23%) across datasets. Finally, creating PGSBMI directly from GxAge GWAS effects increased relative R2 by 7.8%. These results demonstrate that certain covariates, especially those most associated with BMI, significantly affect both PGSBMI performance and effects across diverse cohorts and ancestries, and we provide avenues to improve model performance that consider these effects.

Quantifying factors that affect polygenic risk score performance across diverse ancestries and age groups for body mass index.

Polygenic risk scores (PRS) have led to enthusiasm for precision medicine. However, it is well documented that PRS do not generalize across groups differing in ancestry or sample characteristics e.g., age. Quantifying performance of PRS across different groups of study participants, using genome-wide association study (GWAS) summary statistics from multiple ancestry groups and sample sizes, and using different linkage disequilibrium (LD) reference panels may clarify which factors are limiting PRS transferability. To evaluate these factors in the PRS generation process, we generated body mass index (BMI) PRS (PRSBMI) in the Electronic Medical Records and Genomics (eMERGE) network (N=75,661). Analyses were conducted in two ancestry groups (European and African) and three age ranges (adult, teenagers, and children). For PRSBMI calculations, we evaluated five LD reference panels and three sets of GWAS summary statistics of varying sample size and ancestry. PRSBMI performance increased for both African and European ancestry individuals using cross-ancestry GWAS summary statistics compared to European-only summary statistics (6.3% and 3.7% relative R2 increase, respectively, pAfrican=0.038, pEuropean=6.26x10-4). The effects of LD reference panels were more pronounced in African ancestry study datasets. PRSBMI performance degraded in children; R2 was less than half of teenagers or adults. The effect of GWAS summary statistics sample size was small when modeled with the other factors. Additionally, the potential of using a PRS generated for one trait to predict risk for comorbid diseases is not well understood especially in the context of cross-ancestry analyses - we explored clinical comorbidities from the electronic health record associated with PRSBMI and identified significant associations with type 2 diabetes and coronary atherosclerosis. In summary, this study quantifies the effects that ancestry, GWAS summary statistic sample size, and LD reference panel have on PRS performance, especially in cross-ancestry and age-specific analyses.