Driver gene mutations based clustering of tumors: methods and applications

Area of Science:

• Computational biology and bioinformatics
• Genomics and cancer research

Background:

• Somatic mutations in proto-oncogenes and tumor suppressor genes are key drivers of cancer.
• Large-scale sequencing projects like The Cancer Genome Atlas (TCGA) have generated extensive tumor mutation data.
• Identifying genetically similar tumor subtypes is crucial for precision medicine and understanding cancer heterogeneity.

Purpose of the Study:

• To propose and evaluate two novel computational methods, ccpwModel and xGeneModel, for clustering tumors based on mutation spectra.
• To assess the clinical relevance of identified tumor clusters by correlating them with patient outcomes and characteristics.
• To develop mutation-characterized transcriptomic prognostic signatures for cancer.

Main Methods:

• ccpwModel: Clusters tumors using binary status of cancer driver genes and their pathway involvement.
• xGeneModel: Integrates functional similarities and confidence scores of driver genes with mutation spectra to compute genetic distances.
• Application of both methods to TCGA data across 16 cancer types.

Main Results:

• Both ccpwModel and xGeneModel successfully clustered tumors.
• The identified tumor clusters showed significant associations with patient race and survival time, outperforming state-of-the-art methods.
• Novel mutation-characterized transcriptomic prognostic signatures were detected.

Conclusions:

• The proposed methods provide effective tools for mutation-based tumor subtyping.
• These subtypes have clinical relevance and can inform precision oncology strategies.
• The findings contribute to a deeper understanding of cancer etiology and prognosis.