Recovery of Metagenome-Assembled Genomes from a Human Fecal Sample with Pacific Biosciences High-Fidelity Sequencing.

Here, we report the recovery of 89 metagenome-assembled genomes (MAGs) derived from a human fecal sample subjected to Pacific Biosciences (PacBio) high-fidelity (HiFi) sequencing. A total of 9 MAGs consisted of complete circular contigs, and 45 MAGs were high-quality draft genomes according to the minimum information about a metagenome-assembled genome (MIMAG) standards.

ANNOUNCEMENT

Recent studies have shown that long-read sequencing technologies improve the contiguity of metagenomic assemblies and enable the recovery of repeated regions compared with short-read alternatives (1, 2). To assess long-read technologies and evaluate dedicated bioinformatics tools, we performed deep sequencing of a human fecal metagenome with the PacBio Sequel II system.

A fecal sample was self-collected by a 30-year-old healthy French volunteer for whom written consent was obtained. The sample was stored immediately in a stabilizing solution (RNAlater) following the International Human Microbiome Standards (IHMS) SOP_05_V2 (3, 4). DNA was extracted from 200 mg of fecal material following the IHMS SOP_07_V2 (3, 5). A total of 500 ng of high-molecular-weight DNA was used to build an unamplified nonmultiplexed library with the SMRTbell express template prep kit 2.0 (Pacific Biosciences) following manufacturer recommendations for metagenomics (6). Then, a 30-hour sequencing run was performed on a Sequel II device using Chemistry v2.0. Finally, removal of adapter sequences, read quality control, and generation of circular consensus sequencing (CCS) reads were performed through a dedicated pipeline (7).

Below, default parameters were used for all software unless otherwise specified. Reads shorter than 1,000 bp or aligned to the human genome (GenBank accession number GCA_009914755.3) with minimap v2.24 (8) (parameters, -x asm20) were discarded. In total, 1,645,079 reads with a median quality value of 40 were obtained for a cumulative length of 13,012,430,198 bp. The median length of the sequencing reads was 7,620 bp (Q1 = 5,936 bp; Q3 = 9,635 bp). Metagenomic assembly was performed with Flye v2.9 (9) (parameters: ‐‐pacbio-hifi ‐‐meta), and obtained contigs shorter than 2,500 bp were filtered out. The assembly consisted of 9,253 contigs (including 9 circular contigs of ≥1 Mb) with a cumulative length of 596,522,308 bp. N50 and L50 values were 164,736 bp and 628, respectively. Contig binning was performed with MetaBAT v2.12.1 (10) and SemiBin v0.5.0 (11) (parameters: ‐‐environment human_gut). Results from both tools were combined with the bin_refinement module implemented in metaWRAP v1.3.2 (12). Metagenome-assembled genome (MAG) quality was assessed with checkm v1.1.3 (13) except for one eukaryotic MAG for which BUSCO v5.2.2 (14) was used. A total of 89 MAGs with an estimated completeness of ≥70%, with a contamination of ≤5%, and passing the chimera detection implemented in GUNC v1.0.5 (15) were selected (Fig. 1). These MAGs were annotated subsequently with Prokka v1.14.5 (16), and taxonomic classification was performed with GTDB-Tk v1.5.0 (17). In total, 9 MAGs consisted of complete circular contigs, and 45 MAGs were high-quality draft sequences according to the minimum information about metagenome-assembled genome (MIMAG) standards (18). Notably, all 88 prokaryotic MAGs had at least 1 complete 16S rRNA gene. MAGs were mainly bacteria (87/89), 1 was an archaea (min17_bin38), and the largest (11.6 Mb) was an unicellular eukaryote of the genus Blastocystis (min17_eukbin1). We compared our prokaryotic MAGs with the representative genomes of the UHGG catalogue v2 (19) using fastANI v1.33 (20). Three MAGs corresponded to species not represented in the Unified Human Gastrointestinal Genome (UHGG) collection (average nucleotide identity cutoff = 95%). Remarkably, 30 MAGs had better assembly statistics than the UHGG representatives according to a composite score defined as completeness – (5 × contamination) + log(N50).

FIG 1

Rank order plots comparing the quality of MAGs produced by the three binning methods. Estimated completeness (A), estimated contamination (B), and N50 statistics (C) taken from the CheckM output report are shown. Only MAGs with completeness of ≥70% and contamination of ≤5% are considered.

Data availability.

Sequencing data (accession ERX7722845) and primary metagenome assembly (accession ERZ4963561) were deposited in the European Nucleotide Archive (ENA) under BioProject accession number PRJEB50473. Prokka annotation reports, MAG sequences, and related metadata were deposited in the INRAE data portal (data set S63W9S).