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

Abstract

Motivation: Pacific Biosciences (PacBio) and Oxford Nanopore Technologies (ONT) sequencing technologies can produce long-reads up to tens of kilobases, but with high error rates. In order to reduce sequencing error, Rolling Circle Amplification (RCA) has been used to improve library preparation by amplifying circularized template molecules. Linear products of the RCA contain multiple tandem copies of the template molecule. By integrating additional in silico processing steps, these tandem sequences can be collapsed into a consensus sequence with a higher accuracy than the original raw reads. Existing pipelines using alignment-based methods to discover the tandem repeat patterns from the long-reads are either inefficient or lack sensitivity.

Results: We present a novel tandem repeat detection and consensus calling tool, TideHunter, to efficiently discover tandem repeat patterns and generate high-quality consensus sequences from amplified tandemly repeated long-read sequencing data. TideHunter works with noisy long-reads (PacBio and ONT) at error rates of up to 20% and does not have any limitation of the maximal repeat pattern size. We benchmarked TideHunter using simulated and real datasets with varying error rates and repeat pattern sizes. TideHunter is tens of times faster than state-of-the-art methods and has a higher sensitivity and accuracy.

Availability and implementation: TideHunter is written in C, it is open source and is available at https://github.com/yangao07/TideHunter.

Fig. 1.

Chaining of tandem repeat anchors. Three arrows represent three copies of a template sequence. Vertical line represents seed for each k-mer. The same height between seeds indicates identical k-mers. Horizontal line represents tandem repeat hit of identical k-mers. Solid and dashed lines indicate their hit distances are likely and unlikely, respectively, to be the true repeat pattern size. After the dynamic programming, the optimal chain is expected to consist of anchors that have hit distances close to the repeat pattern size

Fig. 2.

Searching for repeat unit boundary based on the global alignment. s and e are the current repeat boundaries. Two anchors $A_1(s_1,e_1)$ and $A_2(s_2,e_2)$ are selected as their starting positions are the closest to e. Two subsequences starting from s₁ to s₂ and from e₁ to e₂ are extracted to perform an end-to-end global alignment. The next repeat unit boundary $e\prime$ can be calculated based on the alignment result. In this example, the base G of e is matched with the G in subsequence $[e_1:e_2]$, whose coordinate is then considered as the putative next boundary $e\prime$

Fig. 3.

Dynamic programming matrix of sequence-to-graph alignment and three types of operations. SIMD parallelization is applicable for the match and deletion operations as they only rely on the previous rows. Insertion operation must be processed linearly as it depends on the left cell, which is in the same row