Comparison of somatic mutation calling methods in amplicon and whole exome sequence data

Area of Science:

• Genomics
• Bioinformatics
• Cancer Research

Background:

• High-throughput sequencing is integral to clinical diagnosis and cancer research.
• Numerous algorithms exist for somatic single nucleotide variant (SNV) detection in matched tumor-normal DNA sequencing.
• A systematic evaluation of these algorithms using PCR-enriched amplicon data across various variant allele fractions was lacking.

Purpose of the Study:

• To systematically evaluate the performance of popular somatic SNV calling algorithms on both amplicon and exome sequencing data.
• To compare algorithm performance using a well-characterized gold standard variant set (NIST-GIAB).
• To assess the impact of variant allele fraction on algorithm sensitivity.

Main Methods:

• Utilized the NIST-GIAB gold standard variant set for performance evaluation.
• Compared five popular somatic SNV calling algorithms: GATK UnifiedGenotyper, MuTect, Strelka, SomaticSniper, and VarScan2.
• Analyzed both matched tumor-normal amplicon and exome sequencing data.

Main Results:

• All five evaluated somatic SNV calling methods are applicable to both targeted amplicon and exome sequencing data.
• The sensitivity of these methods differs depending on the mutation's allelic fraction in the tumor sample.
• Performance was assessed against a high-confidence reference standard.

Conclusions:

• Commonly used somatic SNV calling algorithms demonstrate applicability across different sequencing data types (amplicon and exome).
• Algorithm sensitivity is influenced by the variant allele fraction, a critical factor in somatic mutation detection.
• This comparative analysis provides guidance for researchers in selecting appropriate SNV calling tools based on specific experimental needs and data types.