DIVAS: an R package for identifying shared and individual variations of multiomics data

MotivationMultiomics data integration aims to identify biological patterns shared across different molecular modalities. Traditional methods focus on detecting either jointly shared variation (across all modalities) or individual variation (unique to single modalities), but overlook partially shared variation, shared by subsets of modalities. This limitation is critical because many biological mechanisms manifest across only some, not all, molecular modalities. Systematically identifying these partially shared patterns is essential for comprehensive understanding of complex biological systems. ResultsWe developed an open-source R package that implements DIVAS, a computational framework for systematically identifying jointly shared, partially shared and individual variations across multiple data types. DIVAS employs angle-based subspace analysis with rigorous statistical inference through rotational bootstrap. Unlike existing methods, DIVAS hierarchically searches through all possible combinations of data modalities, providing complete decomposition of multiomics data into interpretable components (modes of variation) with both scores and loadings. Application to multimodal COVID-19 data demonstrates DIVASs ability to reveal biologically meaningful partially shared immune and metabolic dysregulation patterns that underpins disease severity and would be missed by conventional integration approaches. Availability and implementationDIVAS is available as an R package on GitHub (https://github.com/ByronSyun/DIVAS_Develop). A step-by-step vignette is available on GitHub (https://byronsyun.github.io/DIVAS_COVID19_CaseStudy/).