BayesTyper

BayesTyper performs unbiased probabilistic genotyping of single nucleotide polymorphisms (SNPs), insertions and deletions (indels), and complex structural variants from short-read sequencing by exact alignment of read k-mers to a graph that integrates the reference genome and candidate variants.

Key Features:

• Graph-based k-mer alignment: Exact alignment of read k-mers to a graph representation that integrates the reference genome and candidate variants.
• Probabilistic genotyping: Produces unbiased genotype estimates across SNPs, indels, and complex structural variants using probabilistic models.
• Simultaneous realignment: Realigns reads to the variant-inclusive graph to mitigate bias from aligning short-read sequencing data to a linear reference genome.
• Integration of variant sources: Integrates variants from diverse discovery methods, individuals, and databases into the analysis.
• Variation-prior database: Incorporates a variation-prior database of known variants to enhance sensitivity.
• Efficient algorithmic framework: Employs algorithms to address the computational challenges of graph realignment and genotyping.

Scientific Applications:

• Comprehensive genotyping: Genotyping of SNPs, indels, and complex structural variants with improved sensitivity and accuracy.
• Unbiased variant characterization: Detection and characterization of genetic variation while reducing alignment bias from linear reference mapping.
• Integrated-variant analyses: Analyses that combine variant calls from multiple discovery methods, individuals, and databases to improve variant detection.

Methodology:

Exact alignment of read k-mers to a graph integrating the reference genome and candidate variants; simultaneous realignment to mitigate alignment bias; probabilistic genotyping; integration of variants from diverse discovery methods, individuals, and databases; use of a variation-prior database; and an efficient algorithmic framework to handle computational complexity.