Improved Mutation Detection in Duplex Sequencing Data with Sample-Specific Error Profiles.

Area of Science:

• Genomics
• Computational Biology
• Molecular Biology

Background:

• Duplex sequencing offers high accuracy for rare somatic mutation detection.
• Existing variant callers face limitations in sensitivity, reproducibility, and cross-study comparability due to protocol-specific heuristics.

Purpose of the Study:

• To introduce DupCaller, a novel probabilistic variant caller for duplex sequencing data.
• To improve the detection of rare somatic mutations, including single-base substitutions (SBSs) and indels.

Main Methods:

• Development of DupCaller, a probabilistic variant caller.
• Building sample-specific error profiles.
• Application of a strand-aware statistical model for mutation detection.

Main Results:

• DupCaller identified 1.25-fold more SBSs and 1.41-fold more indels than a state-of-the-art method on synthetic datasets, with comparable or better precision.
• In aristolochic acid-treated cell lines, DupCaller detected 3.5-fold more SBSs and 2.8-fold more indels, recovering expected mutational signatures.
• Across 93 diverse tissue samples, DupCaller consistently detected 1.21- to 2.7-fold more mutations, with sensitivity scaling with sample duplication rate.

Conclusions:

• DupCaller establishes a robust and scalable solution for somatic mutation profiling using duplex sequencing.
• The tool demonstrates consistent performance gains across various biological and technical contexts.
• DupCaller enhances mutation detection, particularly in low-duplication samples where other tools may falter.