WGS2IBI: a cloud-based workflow for individualized Bayesian inference from whole genome sequencing data
TL;DR
WGS2IBI provides a scalable, reproducible, and accessible workflow resource for WGS analysis, enabling efficient population- and individual-level genetic studies without local installation and lowering practical computational barriers for large-scale WGS studies.
Abstract
Scalable and reproducible genomic workflows that support both individual- and population-level analyses are critically needed in precision medicine. We present WGS2IBI, a cloud-based, modular workflow that integrates whole-genome sequencing (WGS) preprocessing, population-level variant screening, and the previously developed Individualized Bayesian Inference (IBI) framework within a reproducible analysis pipeline. Implemented using the Common Workflow Language (CWL) and Docker, WGS2IBI ensures portability and reproducibility across computational environments. Benchmarking on the Jackson Heart Study (JHS) TOPMed Freeze 9 cohort demonstrated efficient large-scale genomic processing and analysis. Preprocessing reduced 102 million variants to 18 million variants in approximately two hours at a total cost of $20.83. Population-level analyses using Global Search and Fisher’s exact test (FET) were completed for $0.44 and $11.44, respectively. Individual-level analysis using IBIwas completed for $3.66. Analyses across additional TOPMed cohorts showed predictable scaling across cohort and chromosome sizes. Deployment on both BioData Catalyst (BDC) and the Gabriella Miller Kids First (KF) Data Resource Center via CAVATICA further demonstrated portability across two major NIH cloud ecosystems. To provide a biologically meaningful use case beyond workflow benchmarking, we also applied WGS2IBI to real hypertension data from 1,821 unrelated Framingham Heart Study (FHS) participants, where IBI-prioritized variants were enriched for lower minor allele frequency and included variants mapping to genes with prior blood-pressure relevance. WGS2IBI provides a scalable, reproducible, and accessible workflow resource for WGS analysis, enabling efficient population- and individual-level genetic studies without local installation. Dual deployment on BDC and KF expands usability across diverse NIH genomic ecosystems, supporting both adult and pediatric research communities. This workflow lowers practical computational barriers for large-scale WGS studies and enables integrated evaluation of individual- and population-level genetic analyses.