Oncodomains: A protein domain-centric framework for analyzing rare variants in tumor samples

Area of Science:

• Genomics
• Computational Biology
• Cancer Research

Background:

• Cancer's heterogeneity complicates identifying driver mutations.
• Traditional gene-centric methods overlook rare somatic variants and gene family similarities.
• Existing approaches struggle to functionally correlate rare somatic mutations with cancer development.

Purpose of the Study:

• To develop a domain-centric method for identifying cancer-driving somatic mutations, especially rare variants.
• To leverage protein domain information for a more comprehensive analysis of somatic mutations in cancer.
• To improve the assessment of rare somatic variants by comparing them across similar genes within families.

Main Methods:

• Developed OncoDomain Hotspots, a domain-centric computational method.
• Utilized protein domain models to identify clusters of somatic mutations across gene families.
• Analyzed structural and functional information within protein domains for variant assessment.

Main Results:

• Identified a broad landscape of somatic variants impacting protein domain families.
• Revealed alterations in key cancer-related pathways like phosphorylation, signaling, gene regulation, and metabolism.
• Demonstrated the method's capability to assess rare somatic variants effectively.

Conclusions:

• OncoDomain Hotspots provides a novel framework for analyzing somatic variants in cancer, particularly rare ones.
• The method integrates structural and functional protein domain data into variant analysis.
• Expected to be a valuable tool for analyzing sequenced tumor genomes and complementing existing methods.