BinDash, software for fast genome distance estimation on a typical personal laptop

Abstract

Supplementary data are available at Bioinformatics online.

Dashing: Fast and Accurate Genomic Distances with HyperLogLog

“…Sets were constructed to have target Jaccard coefficients ranging from 0.00022 to 0.818. Many set-size pairs were evaluated ranging from equal-size sets to sets with sizes differing by a factor of 2 12 . In total, we evaluated 36 combinations of set size and J were evaluated, with full results presented in Additional File 2.…”

“…Spurred by MinHash's utility, other groups have proposed alternatives using new ideas from search and data mining. BinDash [12] uses a b-bit one-permutation rolling MinHash to achieve greater accuracy and speed compared to Mash at a smaller memory footprint. Other theoretical improvements are proposed in the HyperMin-Hash [13] and SuperMinHash [14] studies.…”

“…We implemented the HLL in the Dashing software tool (https://github.com/ dnbaker/dashing), which is free and open source under the GPLv3 license. Dashing supports the functions available in similar tools like Mash [1], BinDash [12] and Sourmash [21]. Dashing can build a sketch of an input sequence set (dashing sketch), including FASTA files (for assembled genomes) or FASTQ files (for sequencing datasets).…”

Mash Screen: High-throughput sequence containment estimation for genome discovery

“…For estimating resemblance, Mash uses a 'bottom sketch' strategy as originally proposed by Broder [8]. More efficient techniques for estimating resemblance have since emerged [9,10,11,12], but bottom sketching is elegant in its simplicity. In short, all k-mers from a genome A are passed through a single hash function h but only the smallest m hash values are stored as the sketch S(A), where |S(A)| << |A|.…”

Large-scale sequence comparisons with sourmash

“…The many-fold size reductions gained via MinHash open the door to extremely large scale searches. While the initial k-mer MinHash implementation focused on enabling Jaccard similarity comparisons (3), it has since been modified and extended to enable k-mer abundance comparisons (4), decrease runtime and memory requirements (5), and work on streaming input data (6). Furthermore, as Jaccard similarity is impacted by the relative size of the sets being compared, containment searches (2,7,8) have been developed to enable detection of a small set within a larger set, such as a genome within a metagenome.…”