Diffsig: Associating Risk Factors with Mutational Signatures.

Area of Science:

• Genomics
• Cancer Research
• Computational Biology

Background:

• Somatic mutational signatures offer insights into cancer's molecular vulnerabilities and therapeutic targets.
• Identifying the causes of mutational signatures is statistically challenging due to small sample sizes and algorithm variability.
• Few mutational signatures have been definitively linked to specific risk factors.

Purpose of the Study:

• To develop a statistical model for associating risk factors with DNA mutational signatures.
• To address the challenges of uncertainty in signature assignment and multiple simultaneous risk factors.

Main Methods:

• Developed Diffsig, a statistical model for estimating associations between risk factors and mutational signatures.
• The model accounts for uncertainty in signature classification and the combined effects of multiple risk factors.

Main Results:

• Applied Diffsig to breast cancer data, confirming known relationships between signatures and etiologic variables.
• Simulations demonstrated the model's accuracy in estimating associations across various scenarios.

Conclusions:

• Diffsig enables robust quantification and inference of associations between risk factors and mutational signatures.
• This tool is expected to enhance understanding of tumor development and refine tumorigenesis models.