Evolving patterns of co-mutations from tumor initiation to metastatic progression

Area of Science:

• Genomics
• Cancer Biology
• Computational Biology

Background:

• Phenotypically healthy cells can carry mutations in cancer-associated genes.
• Malignant transformation typically requires the accumulation of multiple genetic alterations.
• Predicting oncogenic potential of mutation combinations is challenging due to tumor heterogeneity and data limitations.

Purpose of the Study:

• To identify and validate co-occurring mutations (co-mutations) with oncogenic capacity.
• To compare co-mutation patterns across diverse tumor types.
• To investigate co-mutations as potential risk factors for metastatic progression.

Main Methods:

• Curated a large dataset of putative oncogenic mutations from over 70,000 human tumors across 119 subtypes.
• Developed a computational strategy, SelectSim, using in silico mutagenesis simulation to identify co-mutations.
• Validated discovered co-mutations in independent human cohorts.

Main Results:

• Discovered and validated significant co-mutations across various human tumor types.
• Identified co-mutation patterns that differ between tumor types.
• Found that significantly co-occurring mutations are rare in phenotypically normal tissues, unlike individual mutations.
• Identified potential co-mutation-associated risk factors for metastatic progression.

Conclusions:

• Significantly co-occurring mutations are largely cancer-specific and rarely found in healthy tissues.
• These cancer-specific co-mutations provide insights into the pathways of tumorigenesis.
• The SelectSim tool and curated dataset enable robust identification of oncogenic co-mutations.