The genetic architecture of multimodal human brain age.

The complex biological mechanisms underlying human brain aging remain incompletely understood. This study investigated the genetic architecture of three brain age gaps (BAG) derived from gray matter volume (GM-BAG), white matter microstructure (WM-BAG), and functional connectivity (FC-BAG). We identified sixteen genomic loci that reached genome-wide significance (P-value < 5×10-8). A gene-drug-disease network highlighted genes linked to GM-BAG for treating neurodegenerative and neuropsychiatric disorders and WM-BAG genes for cancer therapy. GM-BAG displayed the most pronounced heritability enrichment in genetic variants within conserved regions. Oligodendrocytes and astrocytes, but not neurons, exhibited notable heritability enrichment in WM and FC-BAG, respectively. Mendelian randomization identified potential causal effects of several chronic diseases on brain aging, such as type 2 diabetes on GM-BAG and AD on WM-BAG. Our results provide insights into the genetics of human brain aging, with clinical implications for potential lifestyle and therapeutic interventions. All results are publicly available at https://labs.loni.usc.edu/medicine .

Genomic loci influence patterns of structural covariance in the human brain.

Normal and pathologic neurobiological processes influence brain morphology in coordinated ways that give rise to patterns of structural covariance (PSC) across brain regions and individuals during brain aging and diseases. The genetic underpinnings of these patterns remain largely unknown. We apply a stochastic multivariate factorization method to a diverse population of 50,699 individuals (12 studies and 130 sites) and derive data-driven, multi-scale PSCs of regional brain size. PSCs were significantly correlated with 915 genomic loci in the discovery set, 617 of which are newly identified, and 72% were independently replicated. Key pathways influencing PSCs involve reelin signaling, apoptosis, neurogenesis, and appendage development, while pathways of breast cancer indicate potential interplays between brain metastasis and PSCs associated with neurodegeneration and dementia. Using support vector machines, multi-scale PSCs effectively derive imaging signatures of several brain diseases. Our results elucidate genetic and biological underpinnings that influence structural covariance patterns in the human brain.

The Genetic Architecture of Multimodal Human Brain Age.

The complex biological mechanisms underlying human brain aging remain incompletely understood, involving multiple body organs and chronic diseases. In this study, we used multimodal magnetic resonance imaging and artificial intelligence to examine the genetic architecture of the brain age gap (BAG) derived from gray matter volume (GM-BAG, N=31,557 European ancestry), white matter microstructure (WM-BAG, N=31,674), and functional connectivity (FC-BAG, N=32,017). We identified sixteen genomic loci that reached genome-wide significance (P-value<5×10-8). A gene-drug-disease network highlighted genes linked to GM-BAG for treating neurodegenerative and neuropsychiatric disorders and WM-BAG genes for cancer therapy. GM-BAG showed the highest heritability enrichment for genetic variants in conserved regions, whereas WM-BAG exhibited the highest heritability enrichment in the 5' untranslated regions; oligodendrocytes and astrocytes, but not neurons, showed significant heritability enrichment in WM and FC-BAG, respectively. Mendelian randomization identified potential causal effects of several exposure variables on brain aging, such as type 2 diabetes on GM-BAG (odds ratio=1.05 [1.01, 1.09], P-value=1.96×10-2) and AD on WM-BAG (odds ratio=1.04 [1.02, 1.05], P-value=7.18×10-5). Overall, our results provide valuable insights into the genetics of human brain aging, with clinical implications for potential lifestyle and therapeutic interventions. All results are publicly available at the MEDICINE knowledge portal: https://labs.loni.usc.edu/medicine.

Evidence of epistasis in regions of long-range linkage disequilibrium across five complex diseases in the UK Biobank and eMERGE datasets.

Leveraging linkage disequilibrium (LD) patterns as representative of population substructure enables the discovery of additive association signals in genome-wide association studies (GWASs). Standard GWASs are well-powered to interrogate additive models; however, new approaches are required for invesigating other modes of inheritance such as dominance and epistasis. Epistasis, or non-additive interaction between genes, exists across the genome but often goes undetected because of a lack of statistical power. Furthermore, the adoption of LD pruning as customary in standard GWASs excludes detection of sites that are in LD but might underlie the genetic architecture of complex traits. We hypothesize that uncovering long-range interactions between loci with strong LD due to epistatic selection can elucidate genetic mechanisms underlying common diseases. To investigate this hypothesis, we tested for associations between 23 common diseases and 5,625,845 epistatic SNP-SNP pairs (determined by Ohta's D statistics) in long-range LD (>0.25 cM). Across five disease phenotypes, we identified one significant and four near-significant associations that replicated in two large genotype-phenotype datasets (UK Biobank and eMERGE). The genes that were most likely involved in the replicated associations were (1) members of highly conserved gene families with complex roles in multiple pathways, (2) essential genes, and/or (3) genes that were associated in the literature with complex traits that display variable expressivity. These results support the highly pleiotropic and conserved nature of variants in long-range LD under epistatic selection. Our work supports the hypothesis that epistatic interactions regulate diverse clinical mechanisms and might especially be driving factors in conditions with a wide range of phenotypic outcomes.

Session Introduction: SALUD: Scalable Applications of cLinical risk Utility and preDiction.

This PSB 2023 session discusses challenges in clinical implication and application of risk prediction models, which includes but is not limited to: implementation of risk models, responsible use of polygenic risk scores (PGS), and other risk prediction strategies. We focus on the development and use of new, scalable methods for harmonizing and refining risk prediction models by incorporating genetic and non-genetic risk factors, applying new phenotyping strategies, and integrating clinical factors and biomarkers. Lastly, we will discuss innovation in expanding the utility of these prediction models to underrepresented populations. This session focuses on the overarching theme of enabling early diagnosis, and treatment and preventive measures related to complex diseases and comorbidities.

Implicating genes, pleiotropy, and sexual dimorphism at blood lipid loci through multi-ancestry meta-analysis.

BACKGROUND: Genetic variants within nearly 1000 loci are known to contribute to modulation of blood lipid levels. However, the biological pathways underlying these associations are frequently unknown, limiting understanding of these findings and hindering downstream translational efforts such as drug target discovery. RESULTS: To expand our understanding of the underlying biological pathways and mechanisms controlling blood lipid levels, we leverage a large multi-ancestry meta-analysis (N = 1,654,960) of blood lipids to prioritize putative causal genes for 2286 lipid associations using six gene prediction approaches. Using phenome-wide association (PheWAS) scans, we identify relationships of genetically predicted lipid levels to other diseases and conditions. We confirm known pleiotropic associations with cardiovascular phenotypes and determine novel associations, notably with cholelithiasis risk. We perform sex-stratified GWAS meta-analysis of lipid levels and show that 3-5% of autosomal lipid-associated loci demonstrate sex-biased effects. Finally, we report 21 novel lipid loci identified on the X chromosome. Many of the sex-biased autosomal and X chromosome lipid loci show pleiotropic associations with sex hormones, emphasizing the role of hormone regulation in lipid metabolism. CONCLUSIONS: Taken together, our findings provide insights into the biological mechanisms through which associated variants lead to altered lipid levels and potentially cardiovascular disease risk.

A multi-layer functional genomic analysis to understand noncoding genetic variation in lipids.

A major challenge of genome-wide association studies (GWASs) is to translate phenotypic associations into biological insights. Here, we integrate a large GWAS on blood lipids involving 1.6 million individuals from five ancestries with a wide array of functional genomic datasets to discover regulatory mechanisms underlying lipid associations. We first prioritize lipid-associated genes with expression quantitative trait locus (eQTL) colocalizations and then add chromatin interaction data to narrow the search for functional genes. Polygenic enrichment analysis across 697 annotations from a host of tissues and cell types confirms the central role of the liver in lipid levels and highlights the selective enrichment of adipose-specific chromatin marks in high-density lipoprotein cholesterol and triglycerides. Overlapping transcription factor (TF) binding sites with lipid-associated loci identifies TFs relevant in lipid biology. In addition, we present an integrative framework to prioritize causal variants at GWAS loci, producing a comprehensive list of candidate causal genes and variants with multiple layers of functional evidence. We highlight two of the prioritized genes, CREBRF and RRBP1, which show convergent evidence across functional datasets supporting their roles in lipid biology.

Large-scale genomic analyses reveal insights into pleiotropy across circulatory system diseases and nervous system disorders.

Clinical and epidemiological studies have shown that circulatory system diseases and nervous system disorders often co-occur in patients. However, genetic susceptibility factors shared between these disease categories remain largely unknown. Here, we characterized pleiotropy across 107 circulatory system and 40 nervous system traits using an ensemble of methods in the eMERGE Network and UK Biobank. Using a formal test of pleiotropy, five genomic loci demonstrated statistically significant evidence of pleiotropy. We observed region-specific patterns of direction of genetic effects for the two disease categories, suggesting potential antagonistic and synergistic pleiotropy. Our findings provide insights into the relationship between circulatory system diseases and nervous system disorders which can provide context for future prevention and treatment strategies.

Characterizing Heterogeneity in Neuroimaging, Cognition, Clinical Symptoms, and Genetics Among Patients With Late-Life Depression.

Importance: Late-life depression (LLD) is characterized by considerable heterogeneity in clinical manifestation. Unraveling such heterogeneity might aid in elucidating etiological mechanisms and support precision and individualized medicine. Objective: To cross-sectionally and longitudinally delineate disease-related heterogeneity in LLD associated with neuroanatomy, cognitive functioning, clinical symptoms, and genetic profiles. Design, Setting, and Participants: The Imaging-Based Coordinate System for Aging and Neurodegenerative Diseases (iSTAGING) study is an international multicenter consortium investigating brain aging in pooled and harmonized data from 13 studies with more than 35 000 participants, including a subset of individuals with major depressive disorder. Multimodal data from a multicenter sample (N = 996), including neuroimaging, neurocognitive assessments, and genetics, were analyzed in this study. A semisupervised clustering method (heterogeneity through discriminative analysis) was applied to regional gray matter (GM) brain volumes to derive dimensional representations. Data were collected from July 2017 to July 2020 and analyzed from July 2020 to December 2021. Main Outcomes and Measures: Two dimensions were identified to delineate LLD-associated heterogeneity in voxelwise GM maps, white matter (WM) fractional anisotropy, neurocognitive functioning, clinical phenotype, and genetics. Results: A total of 501 participants with LLD (mean [SD] age, 67.39 [5.56] years; 332 women) and 495 healthy control individuals (mean [SD] age, 66.53 [5.16] years; 333 women) were included. Patients in dimension 1 demonstrated relatively preserved brain anatomy without WM disruptions relative to healthy control individuals. In contrast, patients in dimension 2 showed widespread brain atrophy and WM integrity disruptions, along with cognitive impairment and higher depression severity. Moreover, 1 de novo independent genetic variant (rs13120336; chromosome: 4, 186387714; minor allele, G) was significantly associated with dimension 1 (odds ratio, 2.35; SE, 0.15; P = 3.14 ×108) but not with dimension 2. The 2 dimensions demonstrated significant single-nucleotide variant-based heritability of 18% to 27% within the general population (N = 12 518 in UK Biobank). In a subset of individuals having longitudinal measurements, those in dimension 2 experienced a more rapid longitudinal change in GM and brain age (Cohen f2 = 0.03; P = .02) and were more likely to progress to Alzheimer disease (Cohen f2 = 0.03; P = .03) compared with those in dimension 1 (N = 1431 participants and 7224 scans from the Alzheimer's Disease Neuroimaging Initiative [ADNI], Baltimore Longitudinal Study of Aging [BLSA], and Biomarkers for Older Controls at Risk for Dementia [BIOCARD] data sets). Conclusions and Relevance: This study characterized heterogeneity in LLD into 2 dimensions with distinct neuroanatomical, cognitive, clinical, and genetic profiles. This dimensional approach provides a potential mechanism for investigating the heterogeneity of LLD and the relevance of the latent dimensions to possible disease mechanisms, clinical outcomes, and responses to interventions.

The power of genetic diversity in genome-wide association studies of lipids.

Increased blood lipid levels are heritable risk factors of cardiovascular disease with varied prevalence worldwide owing to different dietary patterns and medication use1. Despite advances in prevention and treatment, in particular through reducing low-density lipoprotein cholesterol levels2, heart disease remains the leading cause of death worldwide3. Genome-wideassociation studies (GWAS) of blood lipid levels have led to important biological and clinical insights, as well as new drug targets, for cardiovascular disease. However, most previous GWAS4-23 have been conducted in European ancestry populations and may have missed genetic variants that contribute to lipid-level variation in other ancestry groups. These include differences in allele frequencies, effect sizes and linkage-disequilibrium patterns24. Here we conduct a multi-ancestry, genome-wide genetic discovery meta-analysis of lipid levels in approximately 1.65 million individuals, including 350,000 of non-European ancestries. We quantify the gain in studying non-European ancestries and provide evidence to support the expansion of recruitment of additional ancestries, even with relatively small sample sizes. We find that increasing diversity rather than studying additional individuals of European ancestry results in substantial improvements in fine-mapping functional variants and portability of polygenic prediction (evaluated in approximately 295,000 individuals from 7 ancestry groupings). Modest gains in the number of discovered loci and ancestry-specific variants were also achieved. As GWAS expand emphasis beyond the identification of genes and fundamental biology towards the use of genetic variants for preventive and precision medicine25, we anticipate that increased diversity of participants will lead to more accurate and equitable26 application of polygenic scores in clinical practice.

A unified framework identifies new links between plasma lipids and diseases from electronic medical records across large-scale cohorts.

Plasma lipids are known heritable risk factors for cardiovascular disease, but increasing evidence also supports shared genetics with diseases of other organ systems. We devised a comprehensive three-phase framework to identify new lipid-associated genes and study the relationships among lipids, genotypes, gene expression and hundreds of complex human diseases from the Electronic Medical Records and Genomics (347 traits) and the UK Biobank (549 traits). Aside from 67 new lipid-associated genes with strong replication, we found evidence for pleiotropic SNPs/genes between lipids and diseases across the phenome. These include discordant pleiotropy in the HLA region between lipids and multiple sclerosis and putative causal paths between triglycerides and gout, among several others. Our findings give insights into the genetic basis of the relationship between plasma lipids and diseases on a phenome-wide scale and can provide context for future prevention and treatment strategies.

Tissue specificity-aware TWAS (TSA-TWAS) framework identifies novel associations with metabolic, immunologic, and virologic traits in HIV-positive adults.

As a type of relatively new methodology, the transcriptome-wide association study (TWAS) has gained interest due to capacity for gene-level association testing. However, the development of TWAS has outpaced statistical evaluation of TWAS gene prioritization performance. Current TWAS methods vary in underlying biological assumptions about tissue specificity of transcriptional regulatory mechanisms. In a previous study from our group, this may have affected whether TWAS methods better identified associations in single tissues versus multiple tissues. We therefore designed simulation analyses to examine how the interplay between particular TWAS methods and tissue specificity of gene expression affects power and type I error rates for gene prioritization. We found that cross-tissue identification of expression quantitative trait loci (eQTLs) improved TWAS power. Single-tissue TWAS (i.e., PrediXcan) had robust power to identify genes expressed in single tissues, but, often found significant associations in the wrong tissues as well (therefore had high false positive rates). Cross-tissue TWAS (i.e., UTMOST) had overall equal or greater power and controlled type I error rates for genes expressed in multiple tissues. Based on these simulation results, we applied a tissue specificity-aware TWAS (TSA-TWAS) analytic framework to look for gene-based associations with pre-treatment laboratory values from AIDS Clinical Trial Group (ACTG) studies. We replicated several proof-of-concept transcriptionally regulated gene-trait associations, including UGT1A1 (encoding bilirubin uridine diphosphate glucuronosyltransferase enzyme) and total bilirubin levels (p = 3.59×10-12), and CETP (cholesteryl ester transfer protein) with high-density lipoprotein cholesterol (p = 4.49×10-12). We also identified several novel genes associated with metabolic and virologic traits, as well as pleiotropic genes that linked plasma viral load, absolute basophil count, and/or triglyceride levels. By highlighting the advantages of different TWAS methods, our simulation study promotes a tissue specificity-aware TWAS analytic framework that revealed novel aspects of HIV-related traits.

Innovative strategies for annotating the "relationSNP" between variants and molecular phenotypes.

Characterizing how variation at the level of individual nucleotides contributes to traits and diseases has been an area of growing interest since the completion of sequencing the first human genome. Our understanding of how a single nucleotide polymorphism (SNP) leads to a pathogenic phenotype on a genome-wide scale is a fruitful endeavor for anyone interested in developing diagnostic tests, therapeutics, or simply wanting to understand the etiology of a disease or trait. To this end, many datasets and algorithms have been developed as resources/tools to annotate SNPs. One of the most common practices is to annotate coding SNPs that affect the protein sequence. Synonymous variants are often grouped as one type of variant, however there are in fact many tools available to dissect their effects on gene expression. More recently, large consortiums like ENCODE and GTEx have made it possible to annotate non-coding regions. Although annotating variants is a common technique among human geneticists, the constant advances in tools and biology surrounding SNPs requires an updated summary of what is known and the trajectory of the field. This review will discuss the history behind SNP annotation, commonly used tools, and newer strategies for SNP annotation. Additionally, we will comment on the caveats that distinguish approaches from one another, along with gaps in the current state of knowledge, and potential future directions. We do not intend for this to be a comprehensive review for any specific area of SNP annotation, but rather it will be an excellent resource for those unfamiliar with computational tools used to functionally characterize SNPs. In summary, this review will help illustrate how each SNP annotation method impacts the way in which the genetic and molecular etiology of a disease is explored in-silico.

Detecting potential pleiotropy across cardiovascular and neurological diseases using univariate, bivariate, and multivariate methods on 43,870 individuals from the eMERGE network.

The link between cardiovascular diseases and neurological disorders has been widely observed in the aging population. Disease prevention and treatment rely on understanding the potential genetic nexus of multiple diseases in these categories. In this study, we were interested in detecting pleiotropy, or the phenomenon in which a genetic variant influences more than one phenotype. Marker-phenotype association approaches can be grouped into univariate, bivariate, and multivariate categories based on the number of phenotypes considered at one time. Here we applied one statistical method per category followed by an eQTL colocalization analysis to identify potential pleiotropic variants that contribute to the link between cardiovascular and neurological diseases. We performed our analyses on ~530,000 common SNPs coupled with 65 electronic health record (EHR)-based phenotypes in 43,870 unrelated European adults from the Electronic Medical Records and Genomics (eMERGE) network. There were 31 variants identified by all three methods that showed significant associations across late onset cardiac- and neurologic- diseases. We further investigated functional implications of gene expression on the detected "lead SNPs" via colocalization analysis, providing a deeper understanding of the discovered associations. In summary, we present the framework and landscape for detecting potential pleiotropy using univariate, bivariate, multivariate, and colocalization methods. Further exploration of these potentially pleiotropic genetic variants will work toward understanding disease causing mechanisms across cardiovascular and neurological diseases and may assist in considering disease prevention as well as drug repositioning in future research.

Influence of tissue context on gene prioritization for predicted transcriptome-wide association studies.

Transcriptome-wide association studies (TWAS) have recently gained great attention due to their ability to prioritize complex trait-associated genes and promote potential therapeutics development for complex human diseases. TWAS integrates genotypic data with expression quantitative trait loci (eQTLs) to predict genetically regulated gene expression components and associates predictions with a trait of interest. As such, TWAS can prioritize genes whose differential expressions contribute to the trait of interest and provide mechanistic explanation of complex trait(s). Tissue-specific eQTL information grants TWAS the ability to perform association analysis on tissues whose gene expression profiles are otherwise hard to obtain, such as liver and heart. However, as eQTLs are tissue context-dependent, whether and how the tissue-specificity of eQTLs influences TWAS gene prioritization has not been fully investigated. In this study, we addressed this question by adopting two distinct TWAS methods, PrediXcan and UTMOST, which assume single tissue and integrative tissue effects of eQTLs, respectively. Thirty-eight baseline laboratory traits in 4,360 antiretroviral treatment-naïve individuals from the AIDS Clinical Trials Group (ACTG) studies comprised the input dataset for TWAS. We performed TWAS in a tissue-specific manner and obtained a total of 430 significant gene-trait associations (q-value < 0.05) across multiple tissues. Single tissue-based analysis by PrediXcan contributed 116 of the 430 associations including 64 unique gene-trait pairs in 28 tissues. Integrative tissue-based analysis by UTMOST found the other 314 significant associations that include 50 unique gene-trait pairs across all 44 tissues. Both analyses were able to replicate some associations identified in past variant-based genome-wide association studies (GWAS), such as high-density lipoprotein (HDL) and CETP (PrediXcan, q-value = 3.2e-16). Both analyses also identified novel associations. Moreover, single tissue-based and integrative tissuebased analysis shared 11 of 103 unique gene-trait pairs, for example, PSRC1-low-density lipoprotein (PrediXcan's lowest q-value = 8.5e-06; UTMOST's lowest q-value = 1.8e-05). This study suggests that single tissue-based analysis may have performed better at discovering gene-trait associations when combining results from all tissues. Integrative tissue-based analysis was better at prioritizing genes in multiple tissues and in trait-related tissue. Additional exploration is needed to confirm this conclusion. Finally, although single tissue-based and integrative tissue-based analysis shared significant novel discoveries, tissue context-dependency of eQTLs impacted TWAS gene prioritization. This study provides preliminary data to support continued work on tissue contextdependency of eQTL studies and TWAS.

Modeling Heterogeneity in the Genetic Architecture of Ethnically Diverse Groups Using Random Effect Interaction Models.

In humans, most genome-wide association studies have been conducted using data from Caucasians and many of the reported findings have not replicated in other populations. This lack of replication may be due to statistical issues (small sample sizes or confounding) or perhaps more fundamentally to differences in the genetic architecture of traits between ethnically diverse subpopulations. What aspects of the genetic architecture of traits vary between subpopulations and how can this be quantified? We consider studying effect heterogeneity using Bayesian random effect interaction models. The proposed methodology can be applied using shrinkage and variable selection methods, and produces useful information about effect heterogeneity in the form of whole-genome summaries (e.g., the proportions of variance of a complex trait explained by a set of SNPs and the average correlation of effects) as well as SNP-specific attributes. Using simulations, we show that the proposed methodology yields (nearly) unbiased estimates when the sample size is not too small relative to the number of SNPs used. Subsequently, we used the methodology for the analyses of four complex human traits (standing height, high-density lipoprotein, low-density lipoprotein, and serum urate levels) in European-Americans (EAs) and African-Americans (AAs). The estimated correlations of effects between the two subpopulations were well below unity for all the traits, ranging from 0.73 to 0.50. The extent of effect heterogeneity varied between traits and SNP sets. Height showed less differences in SNP effects between AAs and EAs whereas HDL, a trait highly influenced by lifestyle, exhibited a greater extent of effect heterogeneity. For all the traits, we observed substantial variability in effect heterogeneity across SNPs, suggesting that effect heterogeneity varies between regions of the genome.

Using Adipose Measures from Health Care Provider-Based Imaging Data for Discovery.

The location and type of adipose tissue is an important factor in metabolic syndrome. A database of picture archiving and communication system (PACS) derived abdominal computerized tomography (CT) images from a large health care provider, Geisinger, was used for large-scale research of the relationship of volume of subcutaneous adipose tissue (SAT) and visceral adipose tissue (VAT) with obesity-related diseases and clinical laboratory measures. Using a "greedy snake" algorithm and 2,545 CT images from the Geisinger PACS, we measured levels of VAT, SAT, total adipose tissue (TAT), and adipose ratio volumes. Sex-combined and sex-stratified association testing was done between adipose measures and 1,233 disease diagnoses and 37 clinical laboratory measures. A genome-wide association study (GWAS) for adipose measures was also performed. SAT was strongly associated with obesity and morbid obesity. VAT levels were strongly associated with type 2 diabetes-related diagnoses (p = 1.5 × 10-58), obstructive sleep apnea (p = 7.7 × 10-37), high-density lipoprotein (HDL) levels (p = 1.42 × 10-36), triglyceride levels (p = 1.44 × 10-43), and white blood cell (WBC) counts (p = 7.37 × 10-9). Sex-stratified tests revealed stronger associations among women, indicating the increased influence of VAT on obesity-related disease outcomes particularly among women. The GWAS identified some suggestive associations. This study supports the utility of pursuing future clinical and genetic discoveries with existing imaging data-derived adipose tissue measures deployed at a larger scale.

Author Correction: Rare variants in drug target genes contributing to complex diseases, phenome-wide.

A correction to this article has been published and is linked from the HTML and PDF versions of this paper. The error has not been fixed in the paper.

Collective feature selection to identify crucial epistatic variants.

BACKGROUND: Machine learning methods have gained popularity and practicality in identifying linear and non-linear effects of variants associated with complex disease/traits. Detection of epistatic interactions still remains a challenge due to the large number of features and relatively small sample size as input, thus leading to the so-called "short fat data" problem. The efficiency of machine learning methods can be increased by limiting the number of input features. Thus, it is very important to perform variable selection before searching for epistasis. Many methods have been evaluated and proposed to perform feature selection, but no single method works best in all scenarios. We demonstrate this by conducting two separate simulation analyses to evaluate the proposed collective feature selection approach. RESULTS: Through our simulation study we propose a collective feature selection approach to select features that are in the "union" of the best performing methods. We explored various parametric, non-parametric, and data mining approaches to perform feature selection. We choose our top performing methods to select the union of the resulting variables based on a user-defined percentage of variants selected from each method to take to downstream analysis. Our simulation analysis shows that non-parametric data mining approaches, such as MDR, may work best under one simulation criteria for the high effect size (penetrance) datasets, while non-parametric methods designed for feature selection, such as Ranger and Gradient boosting, work best under other simulation criteria. Thus, using a collective approach proves to be more beneficial for selecting variables with epistatic effects also in low effect size datasets and different genetic architectures. Following this, we applied our proposed collective feature selection approach to select the top 1% of variables to identify potential interacting variables associated with Body Mass Index (BMI) in ~ 44,000 samples obtained from Geisinger's MyCode Community Health Initiative (on behalf of DiscovEHR collaboration). CONCLUSIONS: In this study, we were able to show that selecting variables using a collective feature selection approach could help in selecting true positive epistatic variables more frequently than applying any single method for feature selection via simulation studies. We were able to demonstrate the effectiveness of collective feature selection along with a comparison of many methods in our simulation analysis. We also applied our method to identify non-linear networks associated with obesity.