Oncomatch- Optimizing Oncology Combination Therapy Prediction

Advances in precision medicine are reshaping cancer treatment by tailoring therapies to apatient’s specific genetic profile. Despite this, matching cancer mutations to effective drugsremains a complex task due to variability in mutations across cancer types and limited tools forpractical clinical application. In this project, initially developed during the BioIT Hackathon2025, we created OncoMatch—an open-data-powered web application designed to bridge thisgap by integrating genomic, transcriptomic, proteomic, and drug-target interaction data tosupport therapy selection.Building on prior work in colorectal cancer, we expanded our scope to include bladder, ovarian,and non-small cell lung cancer (NSCLC), using the COSMIC and DrugCentral databasesto identify relevant gene mutations and therapeutics. We developed two novel scoringsystems—the Cancer Precision Score (CPS) and Gene Precision Score (GPS)—to evaluatedrug specificity and potential effectiveness. Using data from DrugCentral, LINCS L1000,and DeepCoverMOA, we created a unified bioactivity dataset for over 4,000 drugs, includingmeasures such as IC50 and Kd values.The OncoMatch platform features interactive tools to visualize drug bioactivity, assess multiomic and structural similarity among compounds, and explore potential drug combinations.Users can query drugs by cancer type and gene mutation, generating insights into the mostpromising therapies and alternatives.Our open source approach not only democratizes access to high quality bioinformatics toolsbut also encourages data driven, personalized cancer care. Future directions include refiningsubtype level predictions and improving the platform’s utility for combinatorial therapy planning.We have developed a streamlit app to make it easy to access this data. That app can be foundat https://oncomatchapp-precision-medicine.streamlit.app.