SoloDel: a probabilistic model for detecting low-frequent somatic deletions from unmatched sequencing data

Area of Science:

• Genomics
• Computational Biology
• Cancer Research

Background:

• Somatic mutation detection is crucial for genome-based biomedical studies.
• Existing methods for somatic single nucleotide variations are robust, but somatic copy number alteration detection is less explored.
• Challenges include low-frequency mutations and lack of matched control samples.

Purpose of the Study:

• To develop a novel computational method for accurate classification of low-frequency somatic deletions.
• To address challenges of low mutation frequency and absence of matched control samples in somatic copy number alteration detection.

Main Methods:

• Developed SoloDel, a probabilistic somatic mutation progression model.
• Utilized a Gaussian mixture model and expectation-maximization algorithm to analyze read-depth ratios.
• Integrated with conventional structural variation callers for enhanced accuracy.

Main Results:

• SoloDel accurately classifies low-frequency somatic deletions from germline deletions.
• Achieved comparable performance without control samples across a 10-70% mutated subpopulation range.
• Successfully identified validated somatic deletions in neuropsychiatric whole-genome sequencing data.

Conclusions:

• SoloDel offers a robust solution for detecting somatic copy number alterations, particularly low-frequency deletions.
• The method performs well even without matched control samples, increasing its applicability.
• SoloDel enhances the accuracy of somatic mutation classification in genomic studies.