StarSignDNA: mutation-to-signature attribution for accurate representation of mutational processes in low-mutation...

Area of Science:

• Genomics
• Cancer Biology
• Bioinformatics

Background:

• Cell lineages accumulate mutations over time, increasing cancer risk.
• Tumor mutational patterns result from multiple simultaneous mutational processes.
• Disentangling these processes requires deconvolution algorithms.

Purpose of the Study:

• Introduce StarSignDNA, a novel algorithm for mutational signature analysis.
• Enable efficient re-fitting and de novo mutational signature extraction.
• Improve the deconvolution of complex mutational patterns in cancer.

Main Methods:

• Developed StarSignDNA for mutational signature analysis.
• Implemented mutation-to-signature attribution based on trinucleotide context.
• Utilized factorisation of mutation count matrices.

Main Results:

• StarSignDNA demonstrates robust performance, especially in low-mutation contexts.
• The algorithm excels in re-fitting analysis and de novo signature discovery.
• Provides clinically relevant insights into mutational processes.

Conclusions:

• StarSignDNA offers an efficient and accurate method for mutational signature analysis.
• The algorithm aids in understanding the combined impact of multiple mutational processes.
• Facilitates research into cancer genomics and etiology.