A new method for rapid genome classification, clustering, visualization, and novel taxa discovery from metagenome

Abstract

Classifying taxa, including those that have not previously been identified, is a key task in characterizing the microbial communities of under-described habitats, including permanently ice-covered lakes in the dry valleys of the Antarctic. Current supervised phylogeny-based methods fall short on recognizing species assembled from metagenomic datasets from such habitats, as they are often incomplete or lack closely known relatives. Here, we report an efficient software suite, "Genome Constellation", that is capable of rapidly characterizing a large number of metagenome-assembled genomes. Genome Constellation estimates similarities between genomes based on their k-mer matches, and subsequently uses these similarities for classification, clustering, and visualization. The clusters of reference genomes formed by Genome Constellation closely resemble known phylogenetic relationships while simultaneously revealing unexpected connections. In a dataset containing 1,693 draft genomes assembled from the Antarctic lake communities where only 40% could be placed in a phylogenetic tree, Genome Constellation improves taxa assignment to 61%. The clustering-based analysis revealed several novel taxa groups, including six clusters that may represent new bacterial phyla. Remarkably, we discovered 63 new giant viruses, 3 of which could not be found by using the traditional marker-based approach. In summary, we demonstrate that Genome Constellation provides an unbiased option to rapidly analyze a large number of microbial genomes and visually explore their relatedness. The software is available under BSD license at: https://bitbucket.org/berkeleylab/jgi-genomeconstellation/.

Persistent Memory as an Effective Alternative to Random Access Memory in Metagenome Assembly

“…This dataset has a total uncompressed size of 164.8GB. The Antarctic Lake Metagenome Dataset was downloaded from the JGI GOLD database with the accession no Gs0118069 (https://gold.jgi.doe.gov/study?id=Gs0118069). It was derived from samples taken from two meromictic lakes in Antartica (Wang et al, 2019). This dataset has twelve files totalling 1.37TB.…”

“…meta-RAY (Boisvert et al, 2012) uses MPI to distribute large metagenome assembly to multiple computer nodes. To overcome its limitation that it only assembles very abundant species, hybrid strategies have been developed to first use meta-RAY in a computer cluster to assemble abundant species (which often comprise most of the sequencing data), followed by MEGAHIT or metaSPAdes in a single node to assemble unassembled reads (Wang et al, 2019). Recently, MetaHipMer used UPC++ to assemble very large metagenome datasets with high accuracy and efficiency (Hofmeyr et al, 2020), but it runs best on a supercomputer that is not readily available to most researchers.…”

Persistent memory as an effective alternative to random access memory in metagenome assembly

“…id= Gs011 8069). It was derived from samples taken from two meromictic lakes in Antartica [24]. This dataset has twelve files totalling 1.37 TB.…”

“…meta-RAY [3] uses MPI to distribute a large metagenome assembly to multiple computer nodes. To overcome its limitation that it only assembles very abundant species, hybrid strategies have been developed to first use meta-RAY in a computer cluster to assemble abundant species (which often comprise most of the sequencing data), followed by MEGAHIT or metaS-PAdes in a single node to assemble unassembled reads [24]. Recently, MetaHipMer used UPC++ to assemble very large metagenome datasets with high accuracy and efficiency [6], but it runs best on a supercomputer that is not readily available to most researchers.…”