The Trans-Omics for Precision Medicine (TOPMed) program is a GA4GH Driver Project funded by National Heart, Lung, and Blood Institute (NHLBI) to advance precision medicine and improve the understanding of heart, lung, blood, and sleep (HLBS) conditions. Designed as a systems biology approach to discovery, TOPMed layers whole genome sequence (WGS) and -omics data onto deep phenotype, imaging, and environmental data. TOPMed has generated 155,000+ high coverage (30x) WGS and ~100,000 -omics (RNA-seq, methylation, metabolomics, proteomics) profiles on samples from 80+ pre-existing studies span many HLBS traits and study designs. A primary goal of TOPMed is improving the diversity in genomic databases; as such, only 40% of TOPMed WGS participants have European ancestry. Because TOPMed is comprised of pre-existing studies, each study’s informed consent(s) define data use limitations that must be honored when accessing its data.
TOPMed’s infrastructure depends on its Data Coordinating Center (DCC, University of Washington) and Informatics Research Center (IRC, University of Michigan). The DCC provides data quality control, develops statistical methods, harmonizes phenotypes, and offers dbGaP support and data curation. The DCC also supports 32 working groups, 1000+ investigators, and TOPMed’s Ethical, Legal, and Social Implications (ELSI) Committee, which has developed guidance about the return of genomic results to participants and the use of data in publicly accessible variant servers and imputation reference panels. The IRC conducts joint WGS variant calling and develops crucial TOPMed resources like BRAVO (whole genome variant browser), the TOPMed imputation reference panel, and ENCORE (simplified genome- wide analyses). NHLBI is developing a cloud platform, BioData Catalyst (BDC), that provides tools, applications, and workflows to find, access, share, store, and compute TOPMed data.
TOPMed’s next 3-5 years (“TOPMed 2.0”) will focus on building collaborations with large genetic consortia, promoting molecular phenotype-driven discovery, and transitioning from generating data to generating knowledge and information.