#3 · Genomics

Multi-omic and structural data are converted into drug and target prioritization outputs

The dominant thread is computational conversion of heterogeneous molecular data into prioritization outputs. MultiOmicsXplorer provides an R Shiny interface for browsing CPTAC multi-omics cancer data and inferring protein activities, including custom datasets. A structure-based framework scores pH sensitivity from PROPKA pKa values and 3D descriptors, prioritizing cancer secretory and membrane proteins. Network-based approaches extend this logic to therapeutics: RepurNet screens compounds against reverse disease gene signatures using protein-interaction disease modules, and a master-regulator method designs combination therapies for coexisting diffuse midline glioma cell states, validated in vivo. A GNN-Transformer model screens over 28,000 compounds on LINCS L1000 signatures, nominating Class I HDAC inhibitors for pan-cancer use. TargetTrace is distinct as a resource, linking 2,270 drug targets to 71,323 pharmacodynamic biomarker relationships. The shared element is methodological, and validation depth varies across items.