Distinguishing cancer-associated missense mutations from common polymorphisms

Area of Science:

• Genomics
• Bioinformatics
• Cancer Biology

Background:

• Missense variants are frequent in genomic data, but identifying those driving cancer (oncogenesis) is challenging.
• Distinguishing cancer-associated missense mutations typically requires complex in vivo functional analyses.

Purpose of the Study:

• To develop and validate a computational method for accurately classifying cancer-associated missense mutations.
• To differentiate pathogenic mutations from common polymorphisms and other mutation classes.

Main Methods:

• Utilized Sorting Intolerant from Tolerant (SIFT) and Pfam-based LogR.E-value algorithms.
• Developed a classifier integrating Gene Ontology, SIFT, and LogR.E-value metrics.
• Analyzed mutation recurrence and identified a novel germ line variant (P1104A) in TYK2.

Main Results:

• Cancer-associated mutations share features with Mendelian disease mutations, distinct from complex disease or common polymorphisms.
• The developed classifier predicts cancer-associated missense mutations with high accuracy and a low false-positive rate.
• Recurrent mutations are more likely to be predicted as cancer-associated, aiding in distinguishing causal from passenger mutations.

Conclusions:

• A novel bioinformatics approach robustly classifies cancer-associated missense variants for large-scale genomic analysis.
• This method can improve the identification of mutations critical for cancer development and progression.
• The approach aids in distinguishing driver mutations from passenger mutations, with implications for targeted therapies.