ClinicalOmicsDB: exploring molecular associations of oncology drug responses in clinical trials

Area of Science:

• Computational Biology
• Bioinformatics
• Oncology

Background:

• Predictive biomarker identification for optimal cancer treatment is hindered by limited real-world clinical data.
• The integration of omics profiling in clinical trials offers a novel approach to address this challenge.

Purpose of the Study:

• To introduce ClinicalOmicsDB, a web application designed for exploring molecular associations with oncology drug responses within clinical trial data.
• To facilitate the identification of predictive biomarkers for personalized cancer therapy.

Main Methods:

• Compiled transcriptomic data from 40 clinical trials (5913 patients, 11 cancer types, 67 treatment arms) including various drug types and combinations.
• Organized data using an established ontology for intuitive navigation.
• Developed a user-friendly web interface with options to explore associations by study, treatment, or gene.
• Integrated gene set analysis for pathway and tumor microenvironment attributes.
• Implemented a Rust-based backend for performance and provided API/R package for programmatic access.

Main Results:

• ClinicalOmicsDB provides a comprehensive resource linking molecular data to drug response across diverse cancer types and treatments.
• The platform enables interactive exploration of gene expression patterns associated with treatment efficacy.
• Case studies demonstrate the utility of ClinicalOmicsDB in uncovering potential predictive biomarkers for human cancer studies.

Conclusions:

• ClinicalOmicsDB serves as a valuable, freely accessible resource for researchers in oncology and bioinformatics.
• The tool streamlines the analysis of complex omics data from clinical trials to advance personalized medicine.
• Facilitates the discovery of novel molecular associations to improve patient treatment selection.