The Harvest suite for rapid core-genome alignment and visualization of thousands of intraspecific microbial genomes

Abstract

Whole-genome sequences are now available for many microbial species and clades, however existing whole-genome alignment methods are limited in their ability to perform sequence comparisons of multiple sequences simultaneously. Here we present the Harvest suite of core-genome alignment and visualization tools for the rapid and simultaneous analysis of thousands of intraspecific microbial strains. Harvest includes Parsnp, a fast core-genome multi-aligner, and Gingr, a dynamic visual platform. Together they provide interactive core-genome alignments, variant calls, recombination detection, and phylogenetic trees. Using simulated and real data we demonstrate that our approach exhibits unrivaled speed while maintaining the accuracy of existing methods. The Harvest suite is open-source and freely available from: http://github.com/marbl/harvest. RationaleMicrobial genomes represent over 93% of past sequencing projects, with the current total over 10,000 and growing exponentially. Multiple clades of draft and complete genomes comprising hundreds of closely related strains are now available from public databases [1], largely due to an increase in sequencing-based outbreak studies [2]. The quality of future genomes is also set to improve as shortread assemblers mature [3] and long-read sequencing enables finishing at greatly reduced costs [4,5].One direct benefit of high-quality genomes is that they empower comparative genomic studies based on multiple genome alignment. Multiple genome alignment is a fundamental tool in genomics essential for tracking genome evolution [6][7][8] [26,40], recombination, homoplasy, gene conversion, mobile genetic elements, pseudogenization, and convoluted orthology relationships [25]. In addition, the computational burden of multiple sequence alignment remains very high [41] despite recent progress [42].The current influx of microbial sequencing data necessitates methods for large-scale comparative genomics and shifts the focus towards scalability. Current microbial genome alignment methods focus on all-versus-all progressive alignment [31,36] to detect subset relationships (that is, gene gain/loss), but these methods are bounded at various steps by quadratic time complexity. This exponential growth in compute time prohibits comparisons involving thousands of genomes. Chan and Ragan [43] reiterated this point, emphasizing that current phylogenomic methods, such as multiple alignment, will not scale with the increasing number of genomes, and that 'alignment-free' or exact alignment methods must be used to analyze such datasets. However, such approaches do not come without compromising phylogenetic resolution [44].Core-genome alignment is a subset of whole-genome alignment, focused on identifying the set of orthologous

Evaluation of phylogenetic reconstruction methods using bacterial whole genomes: a simulation based study

“…Parsnp 27 produces a core genome alignment by rapidly finding maximal exact matches (as in nucmer) which can include both genes and intergenic regions. In our test we found that it performed even better than FastTree while using less CPU time.…”

Emergence of a new Neisseria meningitidis clonal complex 11 lineage 11.2 clade as an effective urogenital pathogen

“…Draft genome assemblies of urogenital sourced Nm and closed reference genomes of N. meningitidis, N. gonorrheae, and N. lactamica were downloaded from PubMLST (https://pubmlst.org/neisseria/) (52). Genome assemblies were compared using Harvest (v1.1.2); a reference-independent core genome alignment and maximum-likelihood phylogeny were constructed using the Parsnp aligner and then visualized using Gingr (53). The MegAlign program in the DNASTAR Lasergene 13 suite was used for multialignments of short DNA and protein sequences by Clustal W. Allele assignments of MLST, finetyping antigens, and Bexsero Antigen Sequence Typing (54) were retrieved from the PubMLST database.…”

Environmental adaptation and vertical dissemination of ESBL‐/pAmpC‐producing Escherichia coli in an integrated broiler production chain in the absence of an antibiotic treatment

“…The bacterial core genome and the number of SNP differences were calculated between isolates of the same MLST type using Harvest suite 1.0 (Treangen et al ., 2014) and mega 7.0 (Kumar et al ., 2016). Phylogenetic trees were constructed using a pairwise distance matrix and the neighbour‐joining algorithm ( mega 7.0, 1000 bootstraps) based on the number of calculated SNP differences between the isolates.…”