A Compendium of Age-Related PheWAS and GWAS Traits for Human Genetic Association Studies, Their Networks and Genetic Correlations.

Seung-Soo Kim,Nir Barzilai,Sofiya Milman,Adam D Hudgins,Jan Vijg,Zhidong Tu,Brenda Gonzalez,Yousin Suh

doi:10.3389/fgene.2021.680560

Abstract

The rich data from the genome-wide association studies (GWAS) and phenome-wide association studies (PheWAS) offer an unprecedented opportunity to identify the biological underpinnings of age-related disease (ARD) risk and multimorbidity. Surprisingly, however, a comprehensive list of ARDs remains unavailable due to the lack of a clear definition and selection criteria. We developed a method to identify ARDs and to provide a compendium of ARDs for genetic association studies. Querying 1,358 electronic medical record-derived traits, we first defined ARDs and age-related traits (ARTs) based on their prevalence profiles, requiring a unimodal distribution that shows an increasing prevalence after the age of 40 years, and which reaches a maximum peak at 60 years of age or later. As a result, we identified a list of 463 ARDs and ARTs in the GWAS and PheWAS catalogs. We next translated the ARDs and ARTs to their respective 276 Medical Subject Headings diseases and 45 anatomy terms. The most abundant disease categories are neoplasms (48 terms), cardiovascular diseases (44 terms), and nervous system diseases (27 terms). Employing data from a human symptoms-disease network, we found 6 symptom-shared disease groups, representing cancers, heart diseases, brain diseases, joint diseases, eye diseases, and mixed diseases. Lastly, by overlaying our ARD and ART list with genetic correlation data from the UK Biobank, we found 54 phenotypes in 2 clusters with high genetic correlations. Our compendium of ARD and ART is a highly useful resource, with broad applicability for studies of the genetics of aging, ARD, and multimorbidity.

Highlights

In humans, physiological deterioration starts to occur at a young age (26–38 years) with loss of bone, cartilage, muscle mass and strength, and gain of abdominal fat (Belsky et al, 2015)
We found that 106 genome-wide association studies (GWAS) traits and 399 phenomewide association studies (PheWAS) traits (463 total traits) have age-associated prevalence profiles and are identified as age-related traits (ARTs) (Supplementary Table 1)
We identified 463 ARTs in the GWAS and PheWAS catalogs

Summary

Introduction

Physiological deterioration starts to occur at a young age (26–38 years) with loss of bone, cartilage, muscle mass and strength, and gain of abdominal fat (Belsky et al, 2015). An extended period of disease and dysfunction in late life is not an inevitable outcome, as studies on extremely long-lived individuals (i.e., centenarians), have found that they exhibit a significantly delayed age of onset of ARDs, resulting in a substantial compression of latelife morbidity (Partridge et al, 2018). This finding supports the recently widely embraced “geroscience hypothesis” (Kennedy et al, 2014), which posits that chronic diseases (i.e., ARDs) share a common underlying mechanism, the aging process itself, and that by targeting this process for intervention one can target multiple ARDs simultaneously. ARTs can be used as proxy phenotypes of aging, providing a useful basis for both the quantification of the health status of aged cohorts (Fried et al, 2001; Mitnitski et al, 2001; Evert et al, 2003; Terry et al, 2008; Andersen et al, 2012; Belsky et al, 2015; Cieza et al, 2015; Caballero et al, 2017), as well as for studies that aim to identify the shared genetic architectures of ARDs and longevity (Belsky et al, 2015; Fortney et al, 2015; Johnson et al, 2015; Zenin et al, 2019; Melzer et al, 2020)

Objectives

Methods

Results

Conclusion