TideHunter: efficient and sensitive tandem repeat detection from noisy long-reads using seed-and-chain

Area of Science:

• Genomics
• Bioinformatics

Background:

• Long-read sequencing technologies like PacBio and ONT offer extensive read lengths but suffer from high error rates.
• Rolling Circle Amplification (RCA) amplifies circularized DNA, creating tandem repeats in linear products to improve consensus accuracy.
• Current alignment-based pipelines for tandem repeat detection from long-reads are often inefficient and lack sensitivity.

Purpose of the Study:

• To develop a novel tool, TideHunter, for efficient tandem repeat detection and high-quality consensus sequence generation.
• To address the limitations of existing methods in handling noisy long-read data and large repeat patterns.

Main Methods:

• TideHunter is a C-based software tool designed for processing long-read sequencing data.
• It efficiently discovers tandem repeat patterns and collapses amplified tandem sequences into a consensus.
• The tool is capable of handling reads with error rates up to 20% without limitations on repeat pattern size.

Main Results:

• TideHunter demonstrates significant improvements in speed, sensitivity, and accuracy compared to state-of-the-art methods.
• Benchmarking on simulated and real datasets confirmed its efficacy across varying error rates and repeat pattern sizes.
• The tool successfully generates high-quality consensus sequences from noisy long-read data.

Conclusions:

• TideHunter offers a robust and efficient solution for analyzing tandemly repeated sequences in long-read data.
• Its high performance and accuracy make it a valuable asset for genomic research utilizing PacBio and ONT technologies.
• The open-source availability of TideHunter facilitates its adoption and further development in the bioinformatics community.