PI: Nicholas Schork
Many genetically-mediated factors contributing to longevity are likely to impact disease processes in addition to some fundamental mechanism of aging and/or senescence. We believe that an understanding of how much overlap there might be between genetic variants that impact disease susceptibility, as well as disease-related processes, and genetic variants that impact longevity can be obtained by pursuing genetic association studies with large data sets made up of individuals with different diseases and individuals that have lived an exceptionally long and healthy life. We will obtain as many relevant data sets with genotype and sequencing data as possible from resources such as dbGAP and the UK Biobank, and combine them with unique Longevity Consortium (LC) data sets for different types of analyses. This will lead to unprecedentedly large combined data sets with excellent statistical power, amenable to either mega (i.e., combined raw data) and meta (i.e., only using summary statistics) data analyses. Relevant analyses will involve direct association testing or Mendelian Randomization (MR) testing leveraging imputed intermediate phenotypes for causality analysis, but will have to accommodate a harmonization of phenotypes, control for population stratification, as well as potential heterogeneity in genetic effects. Tools for dealing with phenotypic harmonization with be developed and applied, as will analytical methods for handling stratification and heterogeneity. In fact, the development and implementation of analytical methods that accommodate heterogeneity will be a main feature of the proposed research. We emphasize that all findings from other LC investigators will be tested in the proposed analyses, either directly if a genetic variant (e.g., arising from the Centenarians project), via orthology for genes arising from the Mice/Cells project that may harbor interesting human genetic variants, or via imputation, where possible, as an intermediate phenotype (e.g., as a protein arising from the Proteomics or metabolite from the Metabolomics projects) amenable to MR tests. In addition, all factors found to be of interest from the proposed analyses will also be provided to the other investigators as well as the Systems Biology and Chemoinformatics cores for further and integrated analyses. We emphasize that the proposed analyses can be pursued in a wide variety of ways to yield previously undocumented insights. For example, if diabetes and obesity share genetic determinants, then combining individuals with diabetes and obesity and comparing them to individuals who have lived a long life without diabetes or obesity should reveal genetic factors mediating general vulnerabilities to metabolic diseases in all or a subset of individuals with unprecedented power.