Damage and Misrepair Signatures: Compact Representations of Pan-cancer Mutational Processes

Area of Science:

• Genomics
• Cancer Biology
• Computational Biology

Background:

• Somatic mutations, including single-base substitutions (SBSs), drive cancer development.
• Current mutational signature models do not differentiate between DNA damage and subsequent repair errors.
• This limitation hinders a complete understanding of cancer mutation processes.

Purpose of the Study:

• To develop a novel computational model, DAMUTA, to disentangle DNA damage and misrepair processes.
• To analyze their distinct contributions to mutational signatures.
• To improve cancer stratification and understand cellular responses to DNA damage.

Main Methods:

• Developed DAMUTA, a hierarchical Bayesian probabilistic model.
• Applied DAMUTA to 18,974 pan-cancer whole genome sequencing mutation catalogues from 23 cancer types.
• Compared DAMUTA's performance against existing mutational signature approaches and a mutational-burden baseline.

Main Results:

• Tissue-specificity in mutation patterns is primarily driven by variations in DNA damage processes.
• Misrepair signatures are predictive of DNA damage response deficiencies.
• DAMUTA identified a pan-cancer pattern of early clonal transition-mutations shifting to translesion synthesis-associated substitutions.

Conclusions:

• DAMUTA effectively distinguishes damage from misrepair, offering a more refined view of mutational processes.
• The model resolves redundancies in current signature models and facilitates improved tumor stratification.
• This work provides a framework for a unified pan-cancer model of DNA damage response.