Critical Assessment of Metagenome Interpretation-a benchmark of metagenomics software

Assignment of virus and antimicrobial resistance genes to microbial hosts in a complex microbial community by combined long-read assembly and proximity ligation

We describe a method that adds long-read sequencing to a mix of technologies used to assemble a highly complex cattle rumen microbial community, and provide a comparison to short read-based methods. Long-read alignments and Hi-C linkage between contigs support the identification of 188 novel virus-host associations and the determination of phage life cycle states in the rumen microbial community. The long-read assembly also identifies 94 antimicrobial resistance genes, compared to only seven alleles in the short-read assembly. We demonstrate novel techniques that work synergistically to improve characterization of biological features in a highly complex rumen microbial community.

ReprDB and panDB: minimalist databases with maximal microbial representation

BACKGROUND

Profiling of shotgun metagenomic samples is hindered by a lack of unified microbial reference genome databases that (i) assemble genomic information from all open access microbial genomes, (ii) have relatively small sizes, and (iii) are compatible to various metagenomic read mapping tools. Moreover, computational tools to rapidly compile and update such databases to accommodate the rapid increase in new reference genomes do not exist. As a result, database-guided analyses often fail to profile a substantial fraction of metagenomic shotgun sequencing reads from complex microbiomes.

RESULTS

We report pipelines that efficiently traverse all open access microbial genomes and assemble non-redundant genomic information. The pipelines result in two species-resolution microbial reference databases of relatively small sizes: reprDB, which assembles microbial representative or reference genomes, and panDB, for which we developed a novel iterative alignment algorithm to identify and assemble non-redundant genomic regions in multiple sequenced strains. With the databases, we managed to assign taxonomic labels and genome positions to the majority of metagenomic reads from human skin and gut microbiomes, demonstrating a significant improvement over a previous database-guided analysis on the same datasets.

CONCLUSIONS

reprDB and panDB leverage the rapid increases in the number of open access microbial genomes to more fully profile metagenomic samples. Additionally, the databases exclude redundant sequence information to avoid inflated storage or memory space and indexing or analyzing time. Finally, the novel iterative alignment algorithm significantly increases efficiency in pan-genome identification and can be useful in comparative genomic analyses.

Metagenome Assembly and Contig Assignment

The recent development of metagenomic assembly has revolutionized metagenomic data analysis, thanks to the improvement of sequencing techniques, more powerful computational infrastructure and the development of novel algorithms and methods. Using longer assembled contigs rather than raw reads improves the process of metagenomic binning and annotation significantly, ultimately resulting in a deeper understanding of the microbial dynamics of the metagenomic samples being analyzed. In this chapter, we demonstrate a typical metagenomic analysis pipeline including raw read quality evaluation and trimming, assembly and contig binning. Alternative tools that can be used for each step are also discussed.

CoMet: a workflow using contig coverage and composition for binning a metagenomic sample with high precision

Background

In metagenomics, the separation of nucleotide sequences belonging to an individual or closely matched populations is termed binning. Binning helps the evaluation of underlying microbial population structure as well as the recovery of individual genomes from a sample of uncultivable microbial organisms. Both supervised and unsupervised learning methods have been employed in binning; however, characterizing a metagenomic sample containing multiple strains remains a significant challenge.

In this study, we designed and implemented a new workflow, Coverage and composition based binning of Metagenomes (CoMet), for binning contigs in a single metagenomic sample. CoMet utilizes coverage values and the compositional features of metagenomic contigs. The binning strategy in CoMet includes the initial grouping of contigs in guanine-cytosine (GC) content-coverage space and refinement of bins in tetranucleotide frequencies space in a purely unsupervised manner. With CoMet, the clustering algorithm DBSCAN is employed for binning contigs. The performances of CoMet were compared against four existing approaches for binning a single metagenomic sample, including MaxBin, Metawatt, MyCC (default) and MyCC (coverage) using multiple datasets including a sample comprised of multiple strains.

Results

Binning methods based on both compositional features and coverages of contigs had higher performances than the method which is based only on compositional features of contigs. CoMet yielded higher or comparable precision in comparison to the existing binning methods on benchmark datasets of varying complexities. MyCC (coverage) had the highest ranking score in F1-score. However, the performances of CoMet were higher than MyCC (coverage) on the dataset containing multiple strains. Furthermore, CoMet recovered contigs of more species and was 18 - 39% higher in precision than the compared existing methods in discriminating species from the sample of multiple strains. CoMet resulted in higher precision than MyCC (default) and MyCC (coverage) on a real metagenome.

Conclusions

The approach proposed with CoMet for binning contigs, improves the precision of binning while characterizing more species in a single metagenomic sample and in a sample containing multiple strains. The F1-scores obtained from different binning strategies vary with different datasets; however, CoMet yields the highest F1-score with a sample comprised of multiple strains.

Electronic supplementary material

The online version of this article (doi:10.1186/s12859-017-1967-3) contains supplementary material, which is available to authorized users.

Optimizing and evaluating the reconstruction of Metagenome-assembled microbial genomes

Background

Microbiome/host interactions describe characteristics that affect the host's health. Shotgun metagenomics includes sequencing a random subset of the microbiome to analyze its taxonomic and metabolic potential. Reconstruction of DNA fragments into genomes from metagenomes (called metagenome-assembled genomes) assigns unknown fragments to taxa/function and facilitates discovery of novel organisms. Genome reconstruction incorporates sequence assembly and sorting of assembled sequences into bins, characteristic of a genome. However, the microbial community composition, including taxonomic and phylogenetic diversity may influence genome reconstruction. We determine the optimal reconstruction method for four microbiome projects that had variable sequencing platforms (IonTorrent and Illumina), diversity (high or low), and environment (coral reefs and kelp forests), using a set of parameters to select for optimal assembly and binning tools.

Methods

We tested the effects of the assembly and binning processes on population genome reconstruction using 105 marine metagenomes from 4 projects. Reconstructed genomes were obtained from each project using 3 assemblers (IDBA, MetaVelvet, and SPAdes) and 2 binning tools (GroopM and MetaBat). We assessed the efficiency of assemblers using statistics that including contig continuity and contig chimerism and the effectiveness of binning tools using genome completeness and taxonomic identification.

Results

We concluded that SPAdes, assembled more contigs (143,718 ± 124 contigs) of longer length (N50 = 1632 ± 108 bp), and incorporated the most sequences (sequences-assembled = 19.65%). The microbial richness and evenness were maintained across the assembly, suggesting low contig chimeras. SPAdes assembly was responsive to the biological and technological variations within the project, compared with other assemblers. Among binning tools, we conclude that MetaBat produced bins with less variation in GC content (average standard deviation: 1.49), low species richness (4.91 ± 0.66), and higher genome completeness (40.92 ± 1.75) across all projects. MetaBat extracted 115 bins from the 4 projects of which 66 bins were identified as reconstructed metagenome-assembled genomes with sequences belonging to a specific genus. We identified 13 novel genomes, some of which were 100% complete, but show low similarity to genomes within databases.

Conclusions

In conclusion, we present a set of biologically relevant parameters for evaluation to select for optimal assembly and binning tools. For the tools we tested, SPAdes assembler and MetaBat binning tools reconstructed quality metagenome-assembled genomes for the four projects. We also conclude that metagenomes from microbial communities that have high coverage of phylogenetically distinct, and low taxonomic diversity results in highest quality metagenome-assembled genomes.

Electronic supplementary material

The online version of this article (10.1186/s12864-017-4294-1) contains supplementary material, which is available to authorized users.

Rokubacteria: Genomic Giants among the Uncultured Bacterial Phyla

Recent advances in single-cell genomic and metagenomic techniques have facilitated the discovery of numerous previously unknown, deep branches of the tree of life that lack cultured representatives. Many of these candidate phyla are composed of microorganisms with minimalistic, streamlined genomes lacking some core metabolic pathways, which may contribute to their resistance to growth in pure culture. Here we analyzed single-cell genomes and metagenome bins to show that the "Candidate phylum Rokubacteria," formerly known as SPAM, represents an interesting exception, by having large genomes (6-8 Mbps), high GC content (66-71%), and the potential for a versatile, mixotrophic metabolism. We also observed an unusually high genomic heterogeneity among individual Rokubacteria cells in the studied samples. These features may have contributed to the limited recovery of sequences of this candidate phylum in prior cultivation and metagenomic studies. Our analyses suggest that Rokubacteria are distributed globally in diverse terrestrial ecosystems, including soils, the rhizosphere, volcanic mud, oil wells, aquifers, and the deep subsurface, with no reports from marine environments to date.

Benchmarking viromics: an in silico evaluation of metagenome-enabled estimates of viral community composition and diversity

Background

Viral metagenomics (viromics) is increasingly used to obtain uncultivated viral genomes, evaluate community diversity, and assess ecological hypotheses. While viromic experimental methods are relatively mature and widely accepted by the research community, robust bioinformatics standards remain to be established. Here we used in silico mock viral communities to evaluate the viromic sequence-to-ecological-inference pipeline, including (i) read pre-processing and metagenome assembly, (ii) thresholds applied to estimate viral relative abundances based on read mapping to assembled contigs, and (iii) normalization methods applied to the matrix of viral relative abundances for alpha and beta diversity estimates.

Results

Tools specifically designed for metagenomes, specifically metaSPAdes, MEGAHIT, and IDBA-UD, were the most effective at assembling viromes. Read pre-processing, such as partitioning, had virtually no impact on assembly output, but may be useful when hardware is limited. Viral populations with 2–5 × coverage typically assembled well, whereas lesser coverage led to fragmented assembly. Strain heterogeneity within populations hampered assembly, especially when strains were closely related (average nucleotide identity, or ANI ≥97%) and when the most abundant strain represented <50% of the population. Viral community composition assessments based on read recruitment were generally accurate when the following thresholds for detection were applied: (i) ≥10 kb contig lengths to define populations, (ii) coverage defined from reads mapping at ≥90% identity, and (iii) ≥75% of contig length with ≥1 × coverage. Finally, although data are limited to the most abundant viruses in a community, alpha and beta diversity patterns were robustly estimated (±10%) when comparing samples of similar sequencing depth, but more divergent (up to 80%) when sequencing depth was uneven across the dataset. In the latter cases, the use of normalization methods specifically developed for metagenomes provided the best estimates.

Conclusions

These simulations provide benchmarks for selecting analysis cut-offs and establish that an optimized sample-to-ecological-inference viromics pipeline is robust for making ecological inferences from natural viral communities. Continued development to better accessing RNA, rare, and/or diverse viral populations and improved reference viral genome availability will alleviate many of viromics remaining limitations.

Challenging a bioinformatic tool's ability to detect microbial contaminants using in silico whole genome sequencing data

High sensitivity methods such as next generation sequencing and polymerase chain reaction (PCR) are adversely impacted by organismal and DNA contaminants. Current methods for detecting contaminants in microbial materials (genomic DNA and cultures) are not sensitive enough and require either a known or culturable contaminant. Whole genome sequencing (WGS) is a promising approach for detecting contaminants due to its sensitivity and lack of need for a priori assumptions about the contaminant. Prior to applying WGS, we must first understand its limitations for detecting contaminants and potential for false positives. Herein we demonstrate and characterize a WGS-based approach to detect organismal contaminants using an existing metagenomic taxonomic classification algorithm. Simulated WGS datasets from ten genera as individuals and binary mixtures of eight organisms at varying ratios were analyzed to evaluate the role of contaminant concentration and taxonomy on detection. For the individual genomes the false positive contaminants reported depended on the genus, with Staphylococcus, Escherichia, and Shigella having the highest proportion of false positives. For nearly all binary mixtures the contaminant was detected in the in-silico datasets at the equivalent of 1 in 1,000 cells, though F. tularensis was not detected in any of the simulated contaminant mixtures and Y. pestis was only detected at the equivalent of one in 10 cells. Once a WGS method for detecting contaminants is characterized, it can be applied to evaluate microbial material purity, in efforts to ensure that contaminants are characterized in microbial materials used to validate pathogen detection assays, generate genome assemblies for database submission, and benchmark sequencing methods.

Bioinformatics for NGS-based metagenomics and the application to biogas research

Metagenomics has proven to be one of the most important research fields for microbial ecology during the last decade. Starting from 16S rRNA marker gene analysis for the characterization of community compositions to whole metagenome shotgun sequencing which additionally allows for functional analysis, metagenomics has been applied in a wide spectrum of research areas. The cost reduction paired with the increase in the amount of data due to the advent of next-generation sequencing led to a rapidly growing demand for bioinformatic software in metagenomics. By now, a large number of tools that can be used to analyze metagenomic datasets has been developed. The Bielefeld-Gießen center for microbial bioinformatics as part of the German Network for Bioinformatics Infrastructure bundles and imparts expert knowledge in the analysis of metagenomic datasets, especially in research on microbial communities involved in anaerobic digestion residing in biogas reactors. In this review, we give an overview of the field of metagenomics, introduce into important bioinformatic tools and possible workflows, accompanied by application examples of biogas surveys successfully conducted at the Center for Biotechnology of Bielefeld University.

MetaMeta: integrating metagenome analysis tools to improve taxonomic profiling

BACKGROUND

Many metagenome analysis tools are presently available to classify sequences and profile environmental samples. In particular, taxonomic profiling and binning methods are commonly used for such tasks. Tools available among these two categories make use of several techniques, e.g., read mapping, k-mer alignment, and composition analysis. Variations on the construction of the corresponding reference sequence databases are also common. In addition, different tools provide good results in different datasets and configurations. All this variation creates a complicated scenario to researchers to decide which methods to use. Installation, configuration and execution can also be difficult especially when dealing with multiple datasets and tools.

RESULTS

We propose MetaMeta: a pipeline to execute and integrate results from metagenome analysis tools. MetaMeta provides an easy workflow to run multiple tools with multiple samples, producing a single enhanced output profile for each sample. MetaMeta includes a database generation, pre-processing, execution, and integration steps, allowing easy execution and parallelization. The integration relies on the co-occurrence of organisms from different methods as the main feature to improve community profiling while accounting for differences in their databases.

CONCLUSIONS

In a controlled case with simulated and real data, we show that the integrated profiles of MetaMeta overcome the best single profile. Using the same input data, it provides more sensitive and reliable results with the presence of each organism being supported by several methods. MetaMeta uses Snakemake and has six pre-configured tools, all available at BioConda channel for easy installation (conda install -c bioconda metameta). The MetaMeta pipeline is open-source and can be downloaded at: https://gitlab.com/rki_bioinformatics .

Interpreting Microbial Biosynthesis in the Genomic Age: Biological and Practical Considerations

Genome mining has become an increasingly powerful, scalable, and economically accessible tool for the study of natural product biosynthesis and drug discovery. However, there remain important biological and practical problems that can complicate or obscure biosynthetic analysis in genomic and metagenomic sequencing projects. Here, we focus on limitations of available technology as well as computational and experimental strategies to overcome them. We review the unique challenges and approaches in the study of symbiotic and uncultured systems, as well as those associated with biosynthetic gene cluster (BGC) assembly and product prediction. Finally, to explore sequencing parameters that affect the recovery and contiguity of large and repetitive BGCs assembled de novo, we simulate Illumina and PacBio sequencing of the Salinispora tropica genome focusing on assembly of the salinilactam (slm) BGC.

Metagenomics and CAZyme Discovery

Microorganisms play a primary role in regulating biogeochemical cycles and are a valuable source of enzymes that have biotechnological applications, such as carbohydrate-active enzymes (CAZymes). However, the inability to culture the majority of microorganisms that exist in natural ecosystems using common culture-dependent techniques restricts access to potentially novel cellulolytic bacteria and beneficial enzymes. The development of molecular-based culture-independent methods such as metagenomics enables researchers to study microbial communities directly from environmental samples, and presents a platform from which enzymes of interest can be sourced. We outline key methodological stages that are required as well as describe specific protocols that are currently used for metagenomic projects dedicated to CAZyme discovery.