Dindel

Dindel detects small insertions and deletions (indels) in short-read sequencing data to improve the accuracy of variant calls from mapped reads.

Key Features:

• Bayesian probabilistic realignment: Uses a Bayesian framework to realign sequencing reads against candidate haplotypes for improved indel inference.
• Candidate haplotype generation: Constructs candidate haplotypes that incorporate potential indels and single nucleotide variants (SNVs) identified during read mapping.
• Integration of known variants: Incorporates known sequence variants or candidate variants suggested by other methods into the realignment and calling process.
• Error-aware realignment model: Accounts for base-calling inaccuracies, mapping errors, and elevated sequencing error rates in long homopolymer runs.
• Designed for short-read data: Operates on short-read next-generation sequencing data to detect small indels.

Scientific Applications:

• Indel discovery in genomic studies: Detection of small insertions and deletions in resequencing experiments and population-scale projects.
• Variant calling for large cohorts: Applied in large-scale projects such as the 1000 Genomes Project call sets for calling indels.
• Method evaluation: Used to assess sensitivity and false discovery rates on simulated and real sequencing datasets.
• Genetic variation and disease studies: Supports analyses of the contribution of small indels to genetic diversity and disease-related variation.

Methodology:

Employs a Bayesian approach that realigns reads against candidate haplotypes incorporating indels and SNVs, integrates known or externally suggested variants, and uses a probabilistic realignment model that accounts for base-calling errors, mapping errors, and increased error rates in long homopolymer runs.