Abstract
Combinations of cancer drugs have the potential to overcome resistance, improve the response rate of existing drugs and reduce dose-limiting toxicity associated with single agents. Existing drug combination databases only provide response data, such as synergy scores between two drugs, without important contextual information to assist oncologists in matching their patients with these combinations in an evidence-based way. To address this gap, we developed a cancer drug combination database (named as OncoDrug+) by manually collecting and integrating drug combinations and corresponding evidences from FDA databases, clinical guidelines, clinical trials, clinical case reports, patient-derived tumor xenograft models, cell line models and bioinformatics predictions. OncoDrug+ includes 7895 data entries, covering 77 cancer types, including unique 2201 drug combination therapies, involving 1200 biomarkers, 763 published reports and seven types of evidence. Unlike many previous databases only include treatment regime and drug response data, OncoDrug+ provides detailed genetic evidences, pharmacological target information and evidence scores supporting each combination strategy, making evidence-based experimental or clinical applications of cancer drug combinations be possible.