SigsPack

SigsPack estimates sample exposures to known mutational signatures and evaluates exposure stability to support inference of mutational processes in cancer genomics.

Key Features:

• Mutational Signature Estimation: Quantifies a sample’s exposure to predefined mutational signatures, including signatures cataloged in the COSMIC database.
• Exposure Stability Quantification: Assesses stability of estimated exposures using bootstrapping to evaluate reliability across replicates.
• Normalization Tools: Normalizes mutation frequencies with respect to tri-nucleotide contexts within genomic regions to account for context-dependent detection biases.
• Validation of Performance: Validates exposure estimation and stability quantification using both synthetic datasets and real-world data.
• Implementation: Provided as an R/Bioconductor package for integration with R-based bioinformatics workflows.

Scientific Applications:

• Cancer mutational process inference: Identifies active mutational processes in tumor samples by quantifying signature exposures.
• Oncogenic driver inference: Supports inference of potential oncogenic mechanisms by linking exposure patterns to known etiologies.
• Comparative and normalized analyses: Enables more accurate comparisons across samples and regions by applying tri-nucleotide context normalization.

Methodology:

Compares sample mutational catalogs against predefined signature sets (e.g., COSMIC), normalizes mutation counts by tri-nucleotide context across genomic regions, quantifies signature exposures, and uses bootstrapping to assess exposure stability.