An improved encoding of genetic variation in a Burrows-Wheeler transform.

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• High-throughput sequencing generates DNA reads that are mapped to a reference genome.
• Current read mappers, often based on the Burrows-Wheeler transform (BWT), can produce errors due to reference bias.
• Existing methods for encoding genetic variations in BWT have limitations in scope or efficiency.

Purpose of the Study:

• To develop a unified method for encoding various genetic variations within a BWT.
• To improve the accuracy and efficiency of DNA read mapping in the presence of genetic diversity.
• To address the limitations of previous approaches in handling single nucleotide polymorphisms (SNPs) and other complex variations.

Main Methods:

• A hybrid encoding strategy combining IUPAC codes for SNPs and a modified natural encoding for other variations.
• Introduction of a single additional symbol to mark variant sites and delimit multiple variants.
• Adaptation of the backward search algorithm in BWT for efficient handling of encoded genetic variations.

Main Results:

• The proposed method successfully encodes a wide range of genetic variations, including SNPs, insertions, deletions, duplications, transpositions, inversions, and copy-number variations.
• The modified backward search algorithm effectively processes the encoded variations, improving read mapping.
• Implementation and comparison with existing tools (BWBBLE, gramtools) demonstrate competitive or improved performance.

Conclusions:

• The developed method offers a comprehensive and efficient way to represent genetic variations in BWT for enhanced read mapping.
• This approach mitigates reference bias and improves the accuracy of resequencing analyses.
• The method provides a valuable tool for genomic research dealing with diverse populations and complex genetic landscapes.