Binning microbial genomes using deep learning

Jn, Nissen; Ck, Sønderby; Jja, Armenteros; Grønbech, Christopher Heje; H, Bjørn Nielsen; Tn, Petersen; Winther, Ole; Rasmussen, Simon

doi:10.1101/490078

Cited by 21 publications

(21 citation statements)

References 37 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, we show that the sequences as short as 100 bp formed separable clusters using high-level features extracted by DNNs. The feature type generated by supervised learning depends on training targets, which is more focused and task-relevant than auto-encoder methods 27 . Future researches might investigate how to utilize these features, and also incorporate them with co-occurrence or coverage information to build a powerful metagenome binning tool.…”

Section: Discussionmentioning

confidence: 99%

DeepMicrobes: taxonomic classification for metagenomics with deep learning

Liang

Bible

Liu

et al. 2019

Preprint

View full text Add to dashboard Cite

10Taxonomic classification is a crucial step for metagenomics applications 11 including disease diagnostics, microbiome analyses, and outbreak tracing. Yet 12it is unknown what deep learning architecture can capture microbial genome-13 wide features relevant to this task. We report DeepMicrobes 14 (https://github.com/MicrobeLab/DeepMicrobes), a computational framework 15 that can perform large-scale training on > 10,000 RefSeq complete microbial 16 genomes and accurately predict the species-of-origin of whole metagenome 17 shotgun sequencing reads. We show the advantage of DeepMicrobes over 18 state-of-the-art tools in precisely identifying species from microbial community 19 sequencing data. Therefore, DeepMicrobes expands the toolbox of taxonomic 20 classification for metagenomics and enables the development of further deep 21 learning-based bioinformatics algorithms for microbial genomic sequence 22 analysis. 23 4 hypothesize that deep learning can automatically discover taxonomic 45 classification-relevant and genome-wide shared features appearing in short 46 metagenomics sequencing reads given a well-designed deep neural network 47 (DNN) architecture. 48Deep learning has made tremendous recent advances in genomics 5 . 49Taking one-hot encoded DNA sequences as input, the DNNs that have been 50 employed to genomic data fall into two major categories, convolutional neural 51 networks (CNNs) and a hybrid of CNNs and recurrent neural networks (RNNs). 52For example, DeepSEA 6 , PrimateAI 7 and SpliceAI 8 used CNNs to predict the 53 impact of genetic variation. Seq2species 9 also adopted CNNs to predict the 54 species-of-origin of 16S data. DeeperBind 10 and DanQ 11 used hybrid 55 architectures to predict transcription factor binding and DNA accessibility. 56Despite the success of these applications, it remains unknown what DNN 57 architecture and DNA encoding method are suitable for taxonomic classification 58 of metagenomics data. 59Here we describe DeepMicrobes, a k-mer embedding-based recurrent 60 network with attention mechanism (Fig. 1a). We trained the DNN on synthetic 61 reads from RefSeq complete bacterial and archaeal genomes. The first layer of 62DeepMicrobes is designed to encode k-mers to dense vectors through 63 embedding. The vectors are fed into a bidirectional long short-term memory 64 network (BiLSTM) followed by self-attention and a multilayer perceptron (MLP). 65

show abstract

Section: Discussionmentioning

confidence: 99%

DeepMicrobes: taxonomic classification for metagenomics with deep learning

Liang

Bible

Liu

et al. 2019

Preprint

View full text Add to dashboard Cite

show abstract

“…Anvi'o v5.1.0 [23] was also applied to bin contigs >5 kb using CONCOCT v1.0.0 proceeded by manual refinement with redundancy cut-offs of 2.5% for MAGs with 50-75% completeness, 5% for MAGs with 75-90 % completeness, and 10% for MAGs with >90 % completeness. In addition, Vamb v1.0.1 [24] and BinSanity v0.2.8 [25] were used to bin contigs with a minimum length cutoff of 4 kb. All bins generated by the six binners and the MetaWRAP refined bins were further refined using DAS_Tool v1.1.1 [26] with custom penalty parameters (--duplicate_penalty 0.4, --megabin_penalty 0.4) and a score threshold of 0.3.…”

Section: Dna Extraction Sequencing Assembly and Metagenome-assemblmentioning

confidence: 99%

Benchmarking microbial growth rate predictions from metagenomes

et al. 2020

View full text Add to dashboard Cite

Growth rates are central to understanding microbial interactions and community dynamics. Metagenomic growth estimators have been developed, specifically codon usage bias (CUB) for maximum growth rates and “peak-to-trough ratio” (PTR) for in situ rates. Both were originally tested with pure cultures, but natural populations are more heterogeneous, especially in individual cell histories pertinent to PTR. To test these methods, we compared predictors with observed growth rates of freshly collected marine prokaryotes in unamended seawater. We prefiltered and diluted samples to remove grazers and greatly reduce virus infection, so net growth approximated gross growth. We sampled over 44 h for abundances and metagenomes, generating 101 metagenome-assembled genomes (MAGs), including Actinobacteria, Verrucomicrobia, SAR406, MGII archaea, etc. We tracked each MAG population by cell-abundance-normalized read recruitment, finding growth rates of 0 to 5.99 per day, the first reported rates for several groups, and used these rates as benchmarks. PTR, calculated by three methods, rarely correlated to growth (r ~−0.26–0.08), except for rapidly growing γ-Proteobacteria (r ~0.63–0.92), while CUB correlated moderately well to observed maximum growth rates (r = 0.57). This suggests that current PTR approaches poorly predict actual growth of most marine bacterial populations, but maximum growth rates can be approximated from genomic characteristics.

show abstract

“…For all samples, short genomic assemblies (< 1000 bp), which could have biased the subsequent analysis, were first excluded. Genomes were then binned based on their tetranucleotide frequency, differential coverage, GC content, as well as codon usage, by 6 different binning tools: MetaBAT2, MaxBin2, CONCOCT, VAMB, BMC3C, and BinSanity (34)(35)(36)(37)(38). The binning results were refined using the MetaWRAP (v 1.2.1) package based on the bin quality assessment (completeness > 70 and contamination < 20) of different binners from CheckM (39,40).…”

Section: Genome Assembly and Functional Annotationmentioning

confidence: 99%

Endosymbionts of Metazoans Dwelling in the PACManus Hydrothermal Vent: Diversity and Potential Adaptive Features Revealed by Genome Analysis

Wang

Li³

et al. 2020

Appl Environ Microbiol

View full text Add to dashboard Cite

Deep-sea hydrothermal vent communities are dominated by invertebrates, namely bathymodiolin mussels, siboglinid tubeworms, and provannid snails. Symbiosis is considered key to successful colonization by these sedentary species in such extreme environments. In the PACManus vent fields, snails, tubeworms, and mussels each colonized a niche with distinct geochemical characteristics. To better understand the metabolic potentials and genomic features contributing to host-environment adaptation, we compared genomes of the symbionts of Bathymodiolus manusensis, Arcovestia ivanovi, and Alviniconcha boucheti sampled at PACManus and discuss their environmental adaptive features. We found that B. manusensis and A. ivanovi are colonized by γ-proteobacteria from distinct clades, whereas endosymbionts of B. manusensis feature high intraspecific heterogeneity with differing metabolic potentials. A. boucheti harbored three novel ϵ-proteobacteria symbionts, suggesting potential species-level diversity of snail symbionts . Genome comparisons revealed gene families related to low pH homeostasis, metal resistance, oxidative stress resistance, environmental sensing/responses, and chemotaxis and motility were most abundant in A. ivanovi's symbiont, followed by symbionts of the vent-mouth-dwelling snail A. boucheti, and relatively low in the vent-periphery-dwelling mussel B. manusensis, which is consistent with their environmental adaptation and host-symbiont interactions. Gene families classified to host interaction/attachment, virulence factors/toxins, and eukaryotic-like proteins were most abundant in symbionts of mussels and least abundant in those of snails, indicating these symbionts may differ in their host-colonization strategy. Comparison of ϵ-symbionts to non-symbionts demonstrated the expanded gene families in symbionts were related to VB12 synthesis, toxin-antitoxin, methylation, and lipopolysaccharide biosynthesis, suggesting these are vital to symbiotic establishment and development in ϵ-proteobacteria. Importance Deep-sea hydrothermal vents are dominated by several invertebrate species. The establishment of symbiosis has long been thought to be the key to successful colonization by these sedentary species in such harsh environments. Yet the relationships between symbiotic bacteria and their hosts, and their role in environment adaptations generally remain unclear. In this paper, we show that the distribution of three host species showed characteristic niche partitioning in the Manus Basin, giving us opportunity to understand how they adapt to their particular habitats. This study also revealed three novel genomes of symbionts from the snails of A. boucheti. Combined with a dataset of other ectosymbiont and free-living bacteria, genome comparisons of the snail endosymbionts pointed to several genetic traits that may have contributed to the lifestyle shift into the epithelial cells in ϵ-proteobacteria. These findings could increase our understanding of invertebrate–endosymbiont in deep-sea ecosystems.

show abstract

Binning microbial genomes using deep learning

Cited by 21 publications

References 37 publications

DeepMicrobes: taxonomic classification for metagenomics with deep learning

DeepMicrobes: taxonomic classification for metagenomics with deep learning

Benchmarking microbial growth rate predictions from metagenomes

Endosymbionts of Metazoans Dwelling in the PACManus Hydrothermal Vent: Diversity and Potential Adaptive Features Revealed by Genome Analysis

Contact Info

Product

Resources

About