MultiGeMS

MultiGeMS detects single nucleotide variants (SNVs) in high-throughput sequencing (HTS) data across multiple samples by applying statistical model selection and enzymatic substitution error correction to improve precision and robustness.

Key Features:

• Multiple Sample Analysis: Analyzes SNVs across multiple samples simultaneously to enable joint variant detection.
• Statistical Model Selection: Employs a statistical model selection procedure to integrate complex data patterns for more accurate SNV identification.
• Enzymatic Substitution Error Correction: Accounts for enzymatic substitution sequencing errors to improve precision and recall, including in low mapping-quality and reference-allele-dominated sites.
• Multiple Testing Problem Mitigation: Incorporates strategies to address the multiple testing problem inherent in multi-sample SNV calling for more reliable inference.
• High-Performance Computing (HPC) Utilization: Leverages HPC techniques to efficiently process large genomic datasets.

Scientific Applications:

• Precision and Recall Benchmarking: Simulation studies reported superior precision among popular multiple-sample SNV callers and strong recall for identifying common SNVs.
• Robustness to Low-Quality Data: Demonstrates maintained performance on simulated and real datasets with low-quality sequencing data.

Methodology:

Builds upon GeMS by applying statistical model selection, enzymatic substitution error correction, multiple-testing mitigation strategies, and HPC techniques.