MetaPlatanus: a metagenome assembler that combines long-range sequence links and species-specific features. Issue 22 (27th September 2021)

• Record Type: Journal Article
• Authors: Kajitani, Rei; Noguchi, Hideki; Gotoh, Yasuhiro; Ogura, Yoshitoshi; Yoshimura, Dai; Okuno, Miki; Toyoda, Atsushi; Kuwahara, Tomomi; Hayashi, Tetsuya; Itoh, Takehiko
• Is Part Of: Nucleic acids research. Volume 49:Issue 22(2021)
• Page Start: e130
• Page End: e130
• Publication Date: 2021-09-27
• Subjects: Nucleic acids -- Periodicals; Molecular biology -- Periodicals; 572.805
• DOI: 10.1093/nar/gkab831
• Languages: English
• ISSNs: 0305-1048

Abstract

Abstract: De novo metagenome assembly is effective in assembling multiple draft genomes, including those of uncultured organisms. However, heterogeneity in the metagenome hinders assembly and introduces interspecies misassembly deleterious for downstream analysis. For this purpose, we developed a hybrid metagenome assembler, MetaPlatanus. First, as a characteristic function, it assembles the basic contigs from accurate short reads and then iteratively utilizes long-range sequence links, species-specific sequence compositions, and coverage depth. The binning information was also used to improve contiguity. Benchmarking using mock datasets consisting of known bacteria with long reads or mate pairs revealed the high contiguity MetaPlatanus with a few interspecies misassemblies. For published human gut data with nanopore reads from potable sequencers, MetaPlatanus assembled many biologically important elements, such as coding genes, gene clusters, viral sequences, and over-half bacterial genomes. In the benchmark with published human saliva data with high-throughput nanopore reads, the superiority of MetaPlatanus was considerably more evident. We found that some high-abundance bacterial genomes were assembled only by MetaPlatanus as near-complete. Furthermore, MetaPlatanus can circumvent the limitations of highly fragmented assemblies and frequent interspecies misassembles obtained by the other tools. Overall, the study demonstrates that MetaPlatanus could be an effective approach for exploring large-scale structures in metagenomes.