SNPs detection by eBWT positional clustering.

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Rapid advancements in sequencing technology necessitate efficient data structures for storing and analyzing raw sequencing reads.
• There is a growing demand for alignment-free and reference-free variant calling methods that operate directly on indexed raw reads.
• Current methods often rely on reference genomes, limiting their applicability in certain research scenarios.

Purpose of the Study:

• To develop a novel theoretical framework, positional clustering, for analyzing sequencing data.
• To design and implement an alignment-free and reference-free method for single nucleotide polymorphism (SNP) calling.
• To evaluate the efficacy of the proposed method on both synthetic and real sequencing data.

Main Methods:

• Development of the positional clustering theory based on the extended Burrows-Wheeler Transform (eBWT) and LCP array.
• Design and implementation of an alignment-free and reference-free SNP calling pipeline utilizing the eBWT and LCP arrays.
• Experimental validation using synthetic datasets and real sequencing data to assess performance and accuracy.

Main Results:

• The positional clustering theory accurately describes how bases covering the same genome position cluster in the eBWT.
• A simple scan of the eBWT and LCP arrays allows for the detection of SNPs within these clusters.
• The implemented tool provides an intrinsic reference-free evaluation of accuracy by reporting SNP coverage.

Conclusions:

• The positional clustering framework is effective for identifying SNPs directly from raw sequencing data.
• This approach offers a promising avenue for calling other types of genetic variants without relying on a reference genome.
• The developed software, ebwt2snp, is freely available for academic use, facilitating further research in this area.