A unified haplotype-based method for accurate and comprehensive variant calling

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Existing haplotype-based variant callers are optimized for common germline variations in diploid populations, yielding suboptimal performance in diverse scenarios.
• There is a need for a versatile variant caller capable of handling complex genomic data, including varying ploidy and somatic mutations.

Purpose of the Study:

• To introduce Octopus, a novel variant caller designed for accurate genotyping across a range of experimental designs and ploidy levels.
• To demonstrate Octopus's capability in detecting both germline and somatic variations, including low-frequency mutations.

Main Methods:

• Development of Octopus, a variant caller employing a polymorphic Bayesian genotyping model.
• Unified haplotype-aware framework to model sequencing data from diverse experimental designs.
• Integration of sequencing reads and prior information for phasing genotypes of arbitrary ploidy, including somatic mutations.

Main Results:

• Octopus accurately calls germline variants (SNVs, indels, microinversions) in individuals.
• Demonstrated increased sensitivity for low-frequency somatic variation detection using a synthetic tumor dataset.
• Octopus significantly reduces false positives compared to other variant calling methods.
• Output of realigned evidence BAM files to facilitate validation and interpretation.

Conclusions:

• Octopus provides a unified and accurate approach for variant calling in diverse genomic contexts, including cancer research.
• The tool enhances the detection of challenging somatic mutations with improved precision.
• Octopus's ability to handle arbitrary ploidy and somatic mutations makes it a valuable tool for complex genomic analyses.