vi-HMM: a novel HMM-based method for sequence variant identification in short-read data.

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Accurate identification of sequence variants like single nucleotide polymorphisms (SNPs) and insertion-deletion polymorphisms (INDELs) is crucial for next-generation sequencing (NGS).
• Current variant calling methods often assume positional independence and do not utilize the linkage disequilibrium (LD) between nearby loci.

Purpose of the Study:

• To develop a novel hidden Markov model (HMM)-based method for accurate SNP and INDEL calling in mapped short-read sequencing data.
• To leverage the dependence between genotypes at nearby loci caused by LD for improved variant identification.

Main Methods:

• Proposed vi-HMM, a hidden Markov model (HMM) that allows transitions between hidden states (SNP, Ins, Del, Match) of adjacent genomic bases.
• Utilized the Viterbi algorithm to determine the optimal hidden state path for identifying SNPs and INDELs.

Main Results:

• Simulation studies demonstrated that vi-HMM outperforms existing variant calling methods in sensitivity and F1 score across various sequencing depths.
• Application to real data confirmed vi-HMM's higher accuracy in calling both SNPs and INDELs.

Conclusions:

• vi-HMM provides a more accurate and reliable approach for variant calling in NGS data compared to existing methods.
• The method effectively accounts for linkage disequilibrium, leading to improved SNP and INDEL identification.