MetaBCC-LR: metagenomics binning by coverage and composition for long reads

A review of neural networks for metagenomic binning

One of the main goals of metagenomic studies is to describe the taxonomic diversity of microbial communities. A crucial step in metagenomic analysis is metagenomic binning, which involves the (supervised) classification or (unsupervised) clustering of metagenomic sequences. Various machine learning models have been applied to address this task. In this review, the contributions of artificial neural networks (ANN) in the context of metagenomic binning are detailed, addressing both supervised, unsupervised, and semi-supervised approaches. 34 ANN-based binning tools are systematically compared, detailing their architectures, input features, datasets, advantages, disadvantages, and other relevant aspects. The findings reveal that deep learning approaches, such as convolutional neural networks and autoencoders, achieve higher accuracy and scalability than traditional methods. Gaps in benchmarking practices are highlighted, and future directions are proposed, including standardized datasets and optimization of architectures, for third-generation sequencing. This review provides support to researchers in identifying trends and selecting suitable tools for the metagenomic binning problem.

Solving genomic puzzles: computational methods for metagenomic binning

Metagenomics involves the study of genetic material obtained directly from communities of microorganisms living in natural environments. The field of metagenomics has provided valuable insights into the structure, diversity and ecology of microbial communities. Once an environmental sample is sequenced and processed, metagenomic binning clusters the sequences into bins representing different taxonomic groups such as species, genera, or higher levels. Several computational tools have been developed to automate the process of metagenomic binning. These tools have enabled the recovery of novel draft genomes of microorganisms allowing us to study their behaviors and functions within microbial communities. This review classifies and analyzes different approaches of metagenomic binning and different refinement, visualization, and evaluation techniques used by these methods. Furthermore, the review highlights the current challenges and areas of improvement present within the field of research.

METASEED: a novel approach to full-length 16S rRNA gene reconstruction from short read data

BACKGROUND

With the emergence of Oxford Nanopore technology, now the on-site sequencing of 16S rRNA from environments is available. Due to the error level and structure, the analysis of such data demands some database of reference sequences. However, many taxa from complex and diverse environments, have poor representation in publicly available databases. In this paper, we propose the METASEED pipeline for the reconstruction of full-length 16S sequences from such environments, in order to improve the reference for the subsequent use of on-site sequencing.

RESULTS

We show that combining high-precision short-read sequencing of both 16S and full metagenome from the same samples allow us to reconstruct high-quality 16S sequences from the more abundant taxa. A significant novelty is the carefully designed collection of metagenome reads that matches the 16S amplicons, based on a combination of uniqueness and abundance. Compared to alternative approaches this produces superior results.

CONCLUSION

Our pipeline will facilitate numerous studies associated with various unknown microorganisms, thus allowing the comprehension of the diverse environments. The pipeline is a potential tool in generating a full length 16S rRNA gene database for any environment.

Unveiling microbial diversity: harnessing long-read sequencing technology

Long-read sequencing has recently transformed metagenomics, enhancing strain-level pathogen characterization, enabling accurate and complete metagenome-assembled genomes, and improving microbiome taxonomic classification and profiling. These advancements are not only due to improvements in sequencing accuracy, but also happening across rapidly changing analysis methods. In this Review, we explore long-read sequencing's profound impact on metagenomics, focusing on computational pipelines for genome assembly, taxonomic characterization and variant detection, to summarize recent advancements in the field and provide an overview of available analytical methods to fully leverage long reads. We provide insights into the advantages and disadvantages of long reads over short reads and their evolution from the early days of long-read sequencing to their recent impact on metagenomics and clinical diagnostics. We further point out remaining challenges for the field such as the integration of methylation signals in sub-strain analysis and the lack of benchmarks.

Exploring high-quality microbial genomes by assembling short-reads with long-range connectivity

Although long-read sequencing enables the generation of complete genomes for unculturable microbes, its high cost limits the widespread adoption of long-read sequencing in large-scale metagenomic studies. An alternative method is to assemble short-reads with long-range connectivity, which can be a cost-effective way to generate high-quality microbial genomes. Here, we develop Pangaea, a bioinformatic approach designed to enhance metagenome assembly using short-reads with long-range connectivity. Pangaea leverages connectivity derived from physical barcodes of linked-reads or virtual barcodes by aligning short-reads to long-reads. Pangaea utilizes a deep learning-based read binning algorithm to assemble co-barcoded reads exhibiting similar sequence contexts and abundances, thereby improving the assembly of high- and medium-abundance microbial genomes. Pangaea also leverages a multi-thresholding algorithm strategy to refine assembly for low-abundance microbes. We benchmark Pangaea on linked-reads and a combination of short- and long-reads from simulation data, mock communities and human gut metagenomes. Pangaea achieves significantly higher contig continuity as well as more near-complete metagenome-assembled genomes (NCMAGs) than the existing assemblers. Pangaea also generates three complete and circular NCMAGs on the human gut microbiomes.

MarkerScan: Separation and assembly of cobionts sequenced alongside target species in biodiversity genomics projects

Contamination of public databases by mislabelled sequences has been highlighted for many years and the avalanche of novel sequencing data now being deposited has the potential to make databases difficult to use effectively. It is therefore crucial that sequencing projects and database curators perform pre-submission checks to remove obvious contamination and avoid propagating erroneous taxonomic relationships. However, it is important also to recognise that biological contamination of a target sample with unexpected species' DNA can also lead to the discovery of fascinating biological phenomena through the identification of environmental organisms or endosymbionts. Here, we present a novel, integrated method for detection and generation of high-quality genomes of all non-target genomes co-sequenced in eukaryotic genome sequencing projects. After performing taxonomic profiling of an assembly from the raw data, and leveraging the identity of small rRNA sequences discovered therein as markers, a targeted classification approach retrieves and assembles high-quality genomes. The genomes of these cobionts are then not only removed from the target species' genome but also available for further interrogation. Source code is available from https://github.com/CobiontID/MarkerScan. MarkerScan is written in Python and is deployed as a Docker container.

Unraveling metagenomics through long-read sequencing: a comprehensive review

The study of microbial communities has undergone significant advancements, starting from the initial use of 16S rRNA sequencing to the adoption of shotgun metagenomics. However, a new era has emerged with the advent of long-read sequencing (LRS), which offers substantial improvements over its predecessor, short-read sequencing (SRS). LRS produces reads that are several kilobases long, enabling researchers to obtain more complete and contiguous genomic information, characterize structural variations, and study epigenetic modifications. The current leaders in LRS technologies are Pacific Biotechnologies (PacBio) and Oxford Nanopore Technologies (ONT), each offering a distinct set of advantages. This review covers the workflow of long-read metagenomics sequencing, including sample preparation (sample collection, sample extraction, and library preparation), sequencing, processing (quality control, assembly, and binning), and analysis (taxonomic annotation and functional annotation). Each section provides a concise outline of the key concept of the methodology, presenting the original concept as well as how it is challenged or modified in the context of LRS. Additionally, the section introduces a range of tools that are compatible with LRS and can be utilized to execute the LRS process. This review aims to present the workflow of metagenomics, highlight the transformative impact of LRS, and provide researchers with a selection of tools suitable for this task.

SemiBin2: self-supervised contrastive learning leads to better MAGs for short- and long-read sequencing

MOTIVATION

Metagenomic binning methods to reconstruct metagenome-assembled genomes (MAGs) from environmental samples have been widely used in large-scale metagenomic studies. The recently proposed semi-supervised binning method, SemiBin, achieved state-of-the-art binning results in several environments. However, this required annotating contigs, a computationally costly and potentially biased process.

RESULTS

We propose SemiBin2, which uses self-supervised learning to learn feature embeddings from the contigs. In simulated and real datasets, we show that self-supervised learning achieves better results than the semi-supervised learning used in SemiBin1 and that SemiBin2 outperforms other state-of-the-art binners. Compared to SemiBin1, SemiBin2 can reconstruct 8.3-21.5% more high-quality bins and requires only 25% of the running time and 11% of peak memory usage in real short-read sequencing samples. To extend SemiBin2 to long-read data, we also propose ensemble-based DBSCAN clustering algorithm, resulting in 13.1-26.3% more high-quality genomes than the second best binner for long-read data.

AVAILABILITY AND IMPLEMENTATION

SemiBin2 is available as open source software at https://github.com/BigDataBiology/SemiBin/ and the analysis scripts used in the study can be found at https://github.com/BigDataBiology/SemiBin2_benchmark.

A novel and diverse group of Candidatus Patescibacteria from bathypelagic Lake Baikal revealed through long-read metagenomics

BACKGROUND

Lake Baikal, the world's deepest freshwater lake, contains important numbers of Candidatus Patescibacteria (formerly CPR) in its deepest reaches. However, previously obtained CPR metagenome-assembled genomes recruited very poorly indicating the potential of other groups being present. Here, we have applied for the first time a long-read (PacBio CCS) metagenomic approach to analyze in depth the Ca. Patescibacteria living in the bathypelagic water column of Lake Baikal at 1600 m.

RESULTS

The retrieval of nearly complete 16S rRNA genes before assembly has allowed us to detect the presence of a novel and a likely endemic group of Ca. Patescibacteria inhabiting bathypelagic Lake Baikal. This novel group seems to possess extremely high intra-clade diversity, precluding complete genomes' assembly. However, read binning and scaffolding indicate that these microbes are similar to other Ca. Patescibacteria (i.e. parasites or symbionts), although they seem to carry more anabolic pathways, likely reflecting the extremely oligotrophic habitat they inhabit. The novel bins have not been found anywhere, but one of the groups appears in small amounts in an oligotrophic and deep alpine Lake Thun. We propose this novel group be named Baikalibacteria.

CONCLUSION

The recovery of 16S rRNA genes via long-read metagenomics plus the use of long-read binning to uncover highly diverse "hidden" groups of prokaryotes are key strategies to move forward in ecogenomic microbiology. The novel group possesses enormous intraclade diversity akin to what happens with Ca. Patescibacteria at the interclade level, which is remarkable in an environment that has changed little in the last 25 million years.

Recovery and Analysis of Long-Read Metagenome-Assembled Genomes

The development of long-read nucleic acid sequencing is beginning to make very substantive impact on the conduct of metagenome analysis, particularly in relation to the problem of recovering the genomes of member species of complex microbial communities. Here we outline bioinformatics workflows for the recovery and characterization of complete genomes from long-read metagenome data and some complementary procedures for comparison of cognate draft genomes and gene quality obtained from short-read sequencing and long-read sequencing.

Metagenomic binning with assembly graph embeddings

MOTIVATION

Despite recent advancements in sequencing technologies and assembly methods, obtaining high-quality microbial genomes from metagenomic samples is still not a trivial task. Current metagenomic binners do not take full advantage of assembly graphs and are not optimized for long-read assemblies. Deep graph learning algorithms have been proposed in other fields to deal with complex graph data structures. The graph structure generated during the assembly process could be integrated with contig features to obtain better bins with deep learning.

RESULTS

We propose GraphMB, which uses graph neural networks to incorporate the assembly graph into the binning process. We test GraphMB on long-read datasets of different complexities, and compare the performance with other binners in terms of the number of High Quality (HQ) genome bins obtained. With our approach, we were able to obtain unique bins on all real datasets, and obtain more bins on most datasets. In particular, we obtained on average 17.5% more HQ bins when compared with state-of-the-art binners and 13.7% when aggregating the results of our binner with the others. These results indicate that a deep learning model can integrate contig-specific and graph-structure information to improve metagenomic binning.

AVAILABILITY AND IMPLEMENTATION

GraphMB is available from https://github.com/MicrobialDarkMatter/GraphMB.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Binning long reads in metagenomics datasets using composition and coverage information

Background

Advancements in metagenomics sequencing allow the study of microbial communities directly from their environments. Metagenomics binning is a key step in the species characterisation of microbial communities. Next-generation sequencing reads are usually assembled into contigs for metagenomics binning mainly due to the limited information within short reads. Third-generation sequencing provides much longer reads that have lengths similar to the contigs assembled from short reads. However, existing contig-binning tools cannot be directly applied on long reads due to the absence of coverage information and the presence of high error rates. The few existing long-read binning tools either use only composition or use composition and coverage information separately. This may ignore bins that correspond to low-abundance species or erroneously split bins that correspond to species with non-uniform coverages. Here we present a reference-free binning approach, LRBinner, that combines composition and coverage information of complete long-read datasets. LRBinner also uses a distance-histogram-based clustering algorithm to extract clusters with varying sizes.

Results

The experimental results on both simulated and real datasets show that LRBinner achieves the best binning accuracy in most cases while handling the complete datasets without any sampling. Moreover, we show that binning reads using LRBinner prior to assembly reduces computational resources required for assembly while attaining satisfactory assembly qualities.

Conclusion

LRBinner shows that deep-learning techniques can be used for effective feature aggregation to support the metagenomics binning of long reads. Furthermore, accurate binning of long reads supports improvements in metagenomics assembly, especially in complex datasets. Binning also helps to reduce the resources required for assembly. Source code for LRBinner is freely available at https://github.com/anuradhawick/LRBinner.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13015-022-00221-z.

Inferring Species Compositions of Complex Fungal Communities from Long- and Short-Read Sequence Data

Our study is unique in that it provides an in-depth comparative study of a real-life complex fungal community analyzed with multiple long- and short-read sequencing approaches. These technologies and their application are currently of great interest to diverse biologists as they seek to characterize the community compositions of microbiomes.

Disentangle genus microdiversity within a complex microbial community by using a multi-distance long-read binning method: example of Candidatus Accumulibacter

Complete genomes can be recovered from metagenomes by assembling and binning DNA sequences into metagenome assembled genomes (MAGs). Yet, the presence of microdiversity can hamper the assembly and binning processes, possibly yielding chimeric, highly fragmented and incomplete genomes. Here, the metagenomes of four samples of aerobic granular sludge bioreactors containing Candidatus (Ca.) Accumulibacter, a phosphate-accumulating organism of interest for wastewater treatment, were sequenced with both PacBio and Illumina. Different strategies of genome assembly and binning were investigated, including published protocols and a binning procedure adapted to the binning of long contigs (MuLoBiSC). Multiple criteria were considered to select the best strategy for Ca. Accumulibacter, whose multiple strains in every sample represent a challenging microdiversity. In this case, the best strategy relies on long-read only assembly and a custom binning procedure including MuLoBiSC in metaWRAP. Several high-quality Ca. Accumulibacter MAGs, including a novel species, were obtained independently from different samples. Comparative genomic analysis showed that MAGs retrieved in different samples harbour genomic rearrangements in addition to accumulation of point mutations. The microdiversity of Ca. Accumulibacter, likely driven by mobile genetic elements, causes major difficulties in recovering MAGs, but it is also a hallmark of the panmictic lifestyle of these bacteria.

GraphPlas: Refined Classification of Plasmid Sequences Using Assembly Graphs

Plasmids are extra-chromosomal genetic materials with important markers that affect the function and behaviour of the microorganisms supporting their environmental adaptations. Hence the identification and recovery of such plasmid sequences from assemblies is a crucial task in metagenomics analysis. In the past, machine learning approaches have been developed to separate chromosomes and plasmids. However, there is always a compromise between precision and recall in the existing classification approaches. The similarity of compositions between chromosomes and their plasmids makes it difficult to separate plasmids and chromosomes with high accuracy. However, high confidence classifications are accurate with a significant compromise of recall, and vice versa. Hence, the requirement exists to have more sophisticated approaches to separate plasmids and chromosomes accurately while retaining an acceptable trade-off between precision and recall. We present GraphPlas, a novel approach for plasmid recovery using coverage, composition and assembly graph topology. We evaluated GraphPlas on simulated and real short read assemblies with varying compositions of plasmids and chromosomes. Our experiments show that GraphPlas is able to significantly improve accuracy in detecting plasmid and chromosomal contigs on top of popular state-of-the-art plasmid detection tools. The source code is freely available at: https://github.com/anuradhawick/GraphPlas.

OUP accepted manuscript

Despite recent methodology and reference database improvements for taxonomic profiling tools, metagenomic assembly and genomic binning remain important pillars of metagenomic analysis workflows. In case reference information is lacking, genomic binning is considered to be a state-of-the-art method in mixed culture metagenomic data analysis. In this light, our previously published tool BusyBee Web implements a composition-based binning method efficient enough to function as a rapid online utility. Handling assembled contigs and long nanopore generated reads alike, the webserver provides a wide range of supplementary annotations and visualizations. Half a decade after the initial publication, we revisited existing functionality, added comprehensive visualizations, and increased the number of data analysis customization options for further experimentation. The webserver now allows for visualization-supported differential analysis of samples, which is computationally expensive and typically only performed in coverage-based binning methods. Further, users may now optionally check their uploaded samples for plasmid sequences using PLSDB as a reference database. Lastly, a new application programming interface with a supporting python package was implemented, to allow power users fully automated access to the resource and integration into existing workflows. The webserver is freely available under: https://www.ccb.uni-saarland.de/busybee.

Challenges, Strategies, and Perspectives for Reference-Independent Longitudinal Multi-Omic Microbiome Studies

In recent years, multi-omic studies have enabled resolving community structure and interrogating community function of microbial communities. Simultaneous generation of metagenomic, metatranscriptomic, metaproteomic, and (meta) metabolomic data is more feasible than ever before, thus enabling in-depth assessment of community structure, function, and phenotype, thus resulting in a multitude of multi-omic microbiome datasets and the development of innovative methods to integrate and interrogate those multi-omic datasets. Specifically, the application of reference-independent approaches provides opportunities in identifying novel organisms and functions. At present, most of these large-scale multi-omic datasets stem from spatial sampling (e.g., water/soil microbiomes at several depths, microbiomes in/on different parts of the human anatomy) or case-control studies (e.g., cohorts of human microbiomes). We believe that longitudinal multi-omic microbiome datasets are the logical next step in microbiome studies due to their characteristic advantages in providing a better understanding of community dynamics, including: observation of trends, inference of causality, and ultimately, prediction of community behavior. Furthermore, the acquisition of complementary host-derived omics, environmental measurements, and suitable metadata will further enhance the aforementioned advantages of longitudinal data, which will serve as the basis to resolve drivers of community structure and function to understand the biotic and abiotic factors governing communities and specific populations. Carefully setup future experiments hold great potential to further unveil ecological mechanisms to evolution, microbe-microbe interactions, or microbe-host interactions. In this article, we discuss the challenges, emerging strategies, and best-practices applicable to longitudinal microbiome studies ranging from sampling, biomolecular extraction, systematic multi-omic measurements, reference-independent data integration, modeling, and validation.

Improving metagenomic binning results with overlapped bins using assembly graphs

Background

Metagenomic sequencing allows us to study the structure, diversity and ecology in microbial communities without the necessity of obtaining pure cultures. In many metagenomics studies, the reads obtained from metagenomics sequencing are first assembled into longer contigs and these contigs are then binned into clusters of contigs where contigs in a cluster are expected to come from the same species. As different species may share common sequences in their genomes, one assembled contig may belong to multiple species. However, existing tools for binning contigs only support non-overlapped binning, i.e., each contig is assigned to at most one bin (species).

Results

In this paper, we introduce GraphBin2 which refines the binning results obtained from existing tools and, more importantly, is able to assign contigs to multiple bins. GraphBin2 uses the connectivity and coverage information from assembly graphs to adjust existing binning results on contigs and to infer contigs shared by multiple species. Experimental results on both simulated and real datasets demonstrate that GraphBin2 not only improves binning results of existing tools but also supports to assign contigs to multiple bins.

Conclusion

GraphBin2 incorporates the coverage information into the assembly graph to refine the binning results obtained from existing binning tools. GraphBin2 also enables the detection of contigs that may belong to multiple species. We show that GraphBin2 outperforms its predecessor GraphBin on both simulated and real datasets. GraphBin2 is freely available at https://github.com/Vini2/GraphBin2.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13015-021-00185-6.