Comparative genome analysis using sample-specific string detection in accurate long reads

Abstract

Motivation: Comparative genome analysis of two or more whole-genome sequenced (WGS) samples is at the core of most applications in genomics. These include discovery of genomic differences segregating in population, case-control analysis in common diseases, and rare disorders. With the current progress of accurate long-read sequencing technologies (e.g., circular consensus sequencing from PacBio sequencers) we can dive into studying repeat regions of genome (e.g., segmental duplications) and hard-to-detect variants (e.g., complex structural variants). Results: We propose a novel framework for addressing the comparative genome analysis by discovery of strings that are specific to one genome ("samples-specific" strings). We have developed an accurate and efficient novel method for discovery of samples-specific strings between two groups of WGS samples. The proposed approach will give us the ability to perform comparative genome analysis without the need to map the reads and is not hindered by shortcomings of the reference genome. We show that the proposed approach is capable of accurately finding samples-specific strings representing nearly all variation (>98%) reported across pairs or trios of WGS samples using accurate long reads (e.g., PacBio HiFi data). Availability: The proposed tool is publicly available at https://github.com/Parsoa/PingPong.

Fast detection of specific fragments against a set of sequences

“…The second result consists in the design of an algorithm to compute all the occurrences in a single sequence T of its target-specific factors against a reference R. The algorithm runs in real-time on the target sequence, independently of the number of occurrences of target-specific factors, after a standard processing of the reference. This improves on the result in [15], where the running time of the main algorithm depends on the number of occurrences of sought factors.…”

“…The motivation comes from the analysis of genomic sequences as done for example by Khorsand et al in [15] in which authors introduce the notion of sample-specific strings. To avoid alignments but however to extract interesting elements that differentiate the target from the reference, the chosen specific fragments are minimal forbidden factors, also called minimal absent factors.…”

“…In the article we continue exploring the approach of target-specific words as done in [15] by introducing new other algorithmic techniques to detect them. See also the more general view on the usefulness of formal languages to analyse several genomes using pangenomics graphs by Bonizzoni et al in [5].…”

“…The links can also be simulated with suffix arrays [17] and their implementations, for example the FM-index [13]. The algorithm in [15] uses the FMD index by Li [16]. All these data structures can accommodate the sequences and their reverse complements.…”

Can Formal Languages Help Pangenomics to Represent and Analyze Multiple Genomes?

“…Since being able to detect dissimilarities is also important in sequence comparison, here we also present a novel notion aiming at discovering differences among similar sequences. This is the notion of sample specific string (SFS) [27]. We show applications of both notions in facing problems motivated by computational pangenomics [13,20].…”