Dataset Information

A deep siamese neural network improves metagenome-assembled genomes in microbiome datasets across different environments.

ABSTRACT: Metagenomic binning is the step in building metagenome-assembled genomes (MAGs) when sequences predicted to originate from the same genome are automatically grouped together. The most widely-used methods for binning are reference-independent, operating de novo and enable the recovery of genomes from previously unsampled clades. However, they do not leverage the knowledge in existing databases. Here, we introduce SemiBin, an open source tool that uses deep siamese neural networks to implement a semi-supervised approach, i.e. SemiBin exploits the information in reference genomes, while retaining the capability of reconstructing high-quality bins that are outside the reference dataset. Using simulated and real microbiome datasets from several different habitats from GMGCv1 (Global Microbial Gene Catalog), including the human gut, non-human guts, and environmental habitats (ocean and soil), we show that SemiBin outperforms existing state-of-the-art binning methods. In particular, compared to other methods, SemiBin returns more high-quality bins with larger taxonomic diversity, including more distinct genera and species.

SUBMITTER: Pan S

PROVIDER: S-EPMC9051138 | biostudies-literature | 2022 Apr

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

A deep siamese neural network improves metagenome-assembled genomes in microbiome datasets across different environments.

Pan Shaojun S Zhu Chengkai C Zhao Xing-Ming XM Coelho Luis Pedro LP

Nature communications 20220428 1

Metagenomic binning is the step in building metagenome-assembled genomes (MAGs) when sequences predicted to originate from the same genome are automatically grouped together. The most widely-used methods for binning are reference-independent, operating de novo and enable the recovery of genomes from previously unsampled clades. However, they do not leverage the knowledge in existing databases. Here, we introduce SemiBin, an open source tool that uses deep siamese neural networks to implement a s ...[more]

PMID: 35484115

Similar Datasets

Project description:Eukaryotic organisms evolved in a microbial world and often have intimate associations with diverse bacterial groups. Kelp, brown macroalgae in the order Laminariales, play a vital role in coastal ecosystems, yet we know little about the functional role of the microbial symbionts that cover their photosynthetic surfaces. Here, we reconstructed 79 bacterial metagenome-assembled genomes (MAGs) from blades of the bull kelp, Nereocystis luetkeana, allowing us to determine their metabolic potential and functional roles. Despite the annual life history of bull kelp, nearly half of the bacterial MAGs were detected across multiple years. Diverse members of the kelp microbiome, spanning 6 bacterial phyla, contained genes for transporting and assimilating dissolved organic matter (DOM), which is secreted by kelp in large quantities and likely fuels the metabolism of these heterotrophic bacteria. Bacterial genomes also contained alginate lyase and biosynthesis genes, involved in polysaccharide degradation and biofilm formation, respectively. Kelp-associated bacterial genomes contained genes for dissimilatory nitrate reduction and urea hydrolysis, likely providing a reduced source of nitrogen to the host kelp. The genome of the most abundant member of the kelp microbiome and common macroalgal symbiont, Granulosicoccus, contained a full suite of genes for synthesizing cobalamin (vitamin B12), suggesting that kelp-associated bacteria have the potential to provide their host kelp with vitamins. Finally, kelp-associated Granulosicoccus contained genes that typify the aerobic anoxygenic phototrophic bacteria, including genes for bacteriochlorophyll synthesis and photosystem II reaction center proteins, making them the first known photoheterotrophic representatives of this genus. IMPORTANCE Kelp (brown algae in the order Laminariales) are foundational species that create essential habitat in temperate and arctic coastal marine ecosystems. These photosynthetic giants host millions of microbial taxa whose functions are relatively unknown, despite their potential importance for host-microbe interactions and nutrient cycling in kelp forest ecosystems. We reconstructed bacterial genomes from metagenomic samples collected from blades of the bull kelp, Nereocystis luetkeana, allowing us to determine the functional gene content of specific members of the kelp microbiome. These bacterial genomes spanned 6 phyla and 19 families and included common alga-associated microbial symbionts such as Granulosicoccus. Key functions encoded in kelp-associated bacterial genomes included dissolved organic matter assimilation, alginate metabolism, vitamin B12 biosynthesis, and nitrogen reduction from nitrate and urea to ammonium, potentially providing the host kelp with vitamins and reduced nitrogen.

Project description:Metagenomic studies about cocoa fermentation have mainly reported on the analysis of short reads for determination of operational taxonomic units. However, it is also important to determine metagenome-assembled genomes (MAGs), which are genomes deriving from the assembly of metagenomics. For this research, all the cocoa metagenomes from public databases were downloaded, resulting in five data sets: one from Ghana and four from Brazil. In addition, in silico approaches were used to describe putative phenotypes and the metabolic potential of MAGs. A total of 17 high-quality MAGs were recovered from these microbiomes, as follows: (i) for fungi, Yamadazyma tenuis (n = 1); (ii) lactic acid bacteria, Limosilactobacillus fermentum (n = 5), Liquorilactobacillus cacaonum (n = 1), Liquorilactobacillus nagelli (n = 1), Leuconostoc pseudomesenteroides (n = 1), and Lactiplantibacillus plantarum subsp. plantarum (n = 1); (iii) acetic acid bacteria, Acetobacter senegalensis (n = 2) and Kozakia baliensis (n = 1); and (iv) Bacillus subtilis (n = 1), Brevundimonas sp. (n = 2), and Pseudomonas sp. (n = 1). Medium-quality MAGs were also recovered from cocoa microbiomes, including some that, to our knowledge, were not previously detected in this environment (Liquorilactobacillus vini, Komagataeibacter saccharivorans, and Komagataeibacter maltaceti) and others previously described (Fructobacillus pseudoficulneus and Acetobacter pasteurianus). Taken together, the MAGs were useful for providing an additional description of the microbiome of cocoa fermentation, revealing previously overlooked microorganisms, with prediction of key phenotypes and biochemical pathways. IMPORTANCE The production of chocolate starts with the harvesting of cocoa fruits and the spontaneous fermentation of the seeds in a microbial succession that depends on yeasts, lactic acid bacteria, and acetic acid bacteria in order to eliminate bitter and astringent compounds present in the raw material, which will be further roasted and grinded to originate the cocoa powder that will enter the food processing industry. The microbiota of cocoa fermentation is not completely known, and yet it advanced from culture-based studies to the advent of next-generation DNA sequencing, with the generation of a myriad of data that need bioinformatic approaches to be properly analyzed. Although the majority of metagenomic studies have been based on short reads (operational taxonomic units), it is also important to analyze entire genomes to determine more precisely possible ecological roles of different species. Metagenome-assembled genomes (MAGs) are very useful for this purpose; here, MAGs from cocoa fermentation microbiomes are described, and the possible implications of their phenotypic and metabolic potentials are discussed.

Dataset Information

A deep siamese neural network improves metagenome-assembled genomes in microbiome datasets across different environments.

Publications

A deep siamese neural network improves metagenome-assembled genomes in microbiome datasets across different environments.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets