Modeling the relationship between gene expression and mutational signature.

Area of Science:

• Genomics
• Cancer Biology
• Bioinformatics

Background:

• Somatic mutations and their signatures offer insights into tumorigenesis and potential early cancer prevention.
• Gene expression dysregulation is known to be influenced by somatic mutations.

Purpose of the Study:

• To investigate the hypothesized associations between mutational signatures and gene expression.
• To develop models predicting mutational signatures from gene expression data.

Main Methods:

• Utilized RNA-seq data to model 49 established mutational signatures across 33 cancer types.
• Employed five-fold cross-validation with accuracy and area under the curve as performance metrics.

Main Results:

• Selected 475 models using unconstrained genes and 112 models using protein-coding genes.
• Validated models on an independent lung cancer dataset, achieving >80% accuracy and area under the curve for predicting smoking status.

Conclusions:

• Demonstrated that associations between gene expression and somatic mutations can be translated into associations between gene expression and mutational signatures.
• Suggests gene expression analysis as a viable approach for inferring mutational signatures.