The origin of biased sequence depth in sequence-independent nucleic acid amplification and optimization for efficient massive parallel sequencing

CoCoNet: an efficient deep learning tool for viral metagenome binning

MOTIVATION

Metagenomic approaches hold the potential to characterize microbial communities and unravel the intricate link between the microbiome and biological processes. Assembly is one of the most critical steps in metagenomics experiments. It consists of transforming overlapping DNA sequencing reads into sufficiently accurate representations of the community's genomes. This process is computationally difficult and commonly results in genomes fragmented across many contigs. Computational binning methods are used to mitigate fragmentation by partitioning contigs based on their sequence composition, abundance or chromosome organization into bins representing the community's genomes. Existing binning methods have been principally tuned for bacterial genomes and do not perform favorably on viral metagenomes.

RESULTS

We propose Composition and Coverage Network (CoCoNet), a new binning method for viral metagenomes that leverages the flexibility and the effectiveness of deep learning to model the co-occurrence of contigs belonging to the same viral genome and provide a rigorous framework for binning viral contigs. Our results show that CoCoNet substantially outperforms existing binning methods on viral datasets.

AVAILABILITY AND IMPLEMENTATION

CoCoNet was implemented in Python and is available for download on PyPi (https://pypi.org/). The source code is hosted on GitHub at https://github.com/Puumanamana/CoCoNet and the documentation is available at https://coconet.readthedocs.io/en/latest/index.html. CoCoNet does not require extensive resources to run. For example, binning 100k contigs took about 4 h on 10 Intel CPU Cores (2.4 GHz), with a memory peak at 27 GB (see Supplementary Fig. S9). To process a large dataset, CoCoNet may need to be run on a high RAM capacity server. Such servers are typically available in high-performance or cloud computing settings.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Enrichment approach for unbiased sequencing of respiratory syncytial virus directly from clinical samples

Background: Nasopharyngeal samples contain higher quantities of bacterial and host nucleic acids relative to viruses; presenting challenges during virus metagenomics sequencing, which underpins agnostic sequencing protocols. We aimed to develop a viral enrichment protocol for unbiased whole-genome sequencing of respiratory syncytial virus (RSV) from nasopharyngeal samples using the Oxford Nanopore Technology (ONT) MinION platform. Methods: We assessed two protocols using RSV positive samples. Protocol 1 involved physical pre-treatment of samples by centrifugal processing before RNA extraction, while Protocol 2 entailed direct RNA extraction without prior enrichment. Concentrates from Protocol 1 and RNA extracts from Protocol 2 were each divided into two fractions; one was DNase treated while the other was not. RNA was then extracted from both concentrate fractions per sample and RNA from both protocols converted to cDNA, which was then amplified using the tagged Endoh primers through Sequence-Independent Single-Primer Amplification (SISPA) approach, a library prepared, and sequencing done. Statistical significance during analysis was tested using the Wilcoxon signed-rank test. Results: DNase-treated fractions from both protocols recorded significantly reduced host and bacterial contamination unlike the untreated fractions (in each protocol p<0.01). Additionally, DNase treatment after RNA extraction (Protocol 2) enhanced host and bacterial read reduction compared to when done before (Protocol 1). However, neither protocol yielded whole RSV genomes. Sequenced reads mapped to parts of the nucleoprotein (N gene) and polymerase complex (L gene) from Protocol 1 and 2, respectively. Conclusions: DNase treatment was most effective in reducing host and bacterial contamination, but its effectiveness improved if done after RNA extraction than before. We attribute the incomplete genome segments to amplification biases resulting from the use of short length random sequence (6 bases) in tagged Endoh primers. Increasing the length of the random nucleotides from six hexamers to nine or 12 in future studies may reduce the coverage biases.

A library preparation optimized for metagenomics of RNA viruses

Our understanding of the viral communities associated to animals has not yet reached the level attained on the bacteriome. This situation is due to, among others, technical challenges in adapting metagenomics using high-throughput sequencing to the study of RNA viromes in animals. Although important developments have been achieved in most steps of viral metagenomics, there is yet a key step that has received little attention: the library preparation. This situation differs from bacteriome studies in which developments in library preparation have largely contributed to the democratisation of metagenomics. Here, we present a library preparation optimized for metagenomics of RNA viruses from insect vectors of viral diseases. The library design allows a simple PCR-based preparation, such as those routinely used in bacterial metabarcoding, that is adapted to shotgun sequencing as required in viral metagenomics. We first optimized our library preparation using mock viral communities and then validated a full metagenomic approach incorporating our preparation in two pilot studies with field-caught insect vectors; one including a comparison with a published metagenomic protocol. Our approach provided a fold increase in virus-like sequences compared to other studies, and nearly-full genomes from new virus species. Moreover, our results suggested conserved trends in virome composition within a population of a mosquito species. Finally, the sensitivity of our approach was compared to a commercial diagnostic PCR for the detection of an arbovirus in field-caught insect vectors. Our approach could facilitate studies on viral communities from animals and the democratization of metagenomics in community ecology of viruses.

Iterative epigenomic analyses in the same single cell

Gene expression in individual cells is epigenetically regulated by DNA modifications, histone modifications, transcription factors, and other DNA-binding proteins. It has been shown that multiple histone modifications can predict gene expression and reflect future responses of bulk cells to extracellular cues. However, the predictive ability of epigenomic analysis is still limited for mechanistic research at a single cell level. To overcome this limitation, it would be useful to acquire reliable signals from multiple epigenetic marks in the same single cell. Here, we propose a new approach and a new method for analysis of several components of the epigenome in the same single cell. The new method allows reanalysis of the same single cell. We found that reanalysis of the same single cell is feasible, provides confirmation of the epigenetic signals, and allows application of statistical analysis to identify reproduced reads using data sets generated only from the single cell. Reanalysis of the same single cell is also useful to acquire multiple epigenetic marks from the same single cells. The method can acquire at least five epigenetic marks: H3K27ac, H3K27me3, mediator complex subunit 1, a DNA modification, and a DNA-interacting protein. We can predict active signaling pathways in K562 single cells using the epigenetic data and confirm that the predicted results strongly correlate with actual active signaling pathways identified by RNA-seq results. These results suggest that the new method provides mechanistic insights for cellular phenotypes through multilayered epigenome analysis in the same single cells.

Deep Impact of Random Amplification and Library Construction Methods on Viral Metagenomics Results

Clinical metagenomics is a broad-range agnostic detection method of pathogens, including novel microorganisms. A major limit is the low pathogen load compared to the high background of host nucleic acids. To overcome this issue, several solutions exist, such as applying a very high depth of sequencing, or performing a relative enrichment of viral genomes associated with capsids. At the end, the quantity of total nucleic acids is often below the concentrations recommended by the manufacturers of library kits, which necessitates to random amplify nucleic acids. Using a pool of 26 viruses representative of viral diversity, we observed a deep impact of the nature of sample (total nucleic acids versus RNA only), the reverse transcription, the random amplification and library construction method on virus recovery. We further optimized the two most promising methods and assessed their performance with fully characterized reference virus stocks. Good genome coverage and limit of detection lower than 100 or 1000 genome copies per mL of plasma, depending on the genome viral type, were obtained from a three million reads dataset. Our study reveals that optimized random amplification is a technique of choice when insufficient amounts of nucleic acid are available for direct libraries constructions.

Comparison of Illumina MiSeq and the Ion Torrent PGM and S5 platforms for whole-genome sequencing of picornaviruses and caliciviruses

Next-generation sequencing is a powerful tool for virological surveillance. While Illumina® and Ion Torrent® sequencing platforms are used extensively for generating viral RNA genome sequences, there is limited data comparing different platforms. The Illumina MiSeq, Ion Torrent PGM and Ion Torrent S5 platforms were evaluated using a panel of sixteen specimens containing picornaviruses and human caliciviruses (noroviruses and sapoviruses). The specimens were processed, using combinations of three library preparation and five sequencing kits, to assess the quality and completeness of assembled viral genomes, and an estimation of cost per sample to generate the data was calculated. The choice of library preparation kit and sequencing platform was found to impact the breadth of genome coverage and accuracy of consensus viral genomes. The Ion Torrent S5 510 chip runs produced more reads at a lower cost per sample than the highest output Ion Torrent PGM 318 chip run, and generated the highest proportion of reads for enterovirus D68 samples. However, indels at homopolymer regions impacted the accuracy of consensus genome sequences. For lower throughput sequencing runs (i.e., Ion Torrent 510 and Illumina MiSeq Nano V2), the cost per sample was lower on the MiSeq platform, whereas with higher throughput runs (Ion Torrent 530 and Illumina MiSeq V2) there is less of a difference in the cost per sample between the two sequencing platforms ($5.47-$10.25 more per sample for an Ion Torrent 530 chip run when multiplexing 24 samples). These findings suggest that the Ion Torrent S5 and Illumina MiSeq platforms are both viable options for genomic sequencing of RNA viruses, each with specific advantages and tradeoffs.

Laboratory Methods in Molecular Epidemiology: Viral Infections

Viruses, which are the most abundant biological entities on the planet, have been regarded as the "dark matter" of biology in the sense that despite their ubiquity and frequent presence in large numbers, their detection and analysis are not always straightforward. The majority of them are very small (falling under the limit of 0.5 μm), and collectively, they are extraordinarily diverse. In fact, the majority of the genetic diversity on the planet is found in the so-called virosphere, or the world of viruses. Furthermore, the most frequent viral agents of disease in humans display an RNA genome, and frequently evolve very fast, due to the fact that most of their polymerases are devoid of proofreading activity. Therefore, their detection, genetic characterization, and epidemiological surveillance are rather challenging. This review (part of the Curated Collection on Advances in Molecular Epidemiology of Infectious Diseases) describes many of the methods that, throughout the last few decades, have been used for viral detection and analysis. Despite the challenge of having to deal with high genetic diversity, the majority of these methods still depend on the amplification of viral genomic sequences, using sequence-specific or sequence-independent approaches, exploring thermal profiles or a single nucleic acid amplification temperature. Furthermore, viral populations, and especially those with RNA genomes, are not usually genetically uniform but encompass swarms of genetically related, though distinct, viral genomes known as viral quasispecies. Therefore, sequence analysis of viral amplicons needs to take this fact into consideration, as it constitutes a potential analytic problem. Possible technical approaches to deal with it are also described here. *This article is part of a curated collection.

A Versatile Sample Processing Workflow for Metagenomic Pathogen Detection

Metagenomics is currently the only generic method for pathogen detection. Starting from RNA allows the assessment of the whole sample community including RNA viruses. Here we present our modular concerted protocol for sample processing for diagnostic metagenomics analysis of human, animal, and food samples. The workflow does not rely on dedicated amplification steps at any stage in the process and, in contrast to published methods, libraries prepared accordingly will yield only minute amounts of unclassifiable reads. We confirmed the performance of the approach using a spectrum of pathogen/matrix-combinations showing it has the potential to become a commonly usable analytical framework.

Evaluation of bias induced by viral enrichment and random amplification protocols in metagenomic surveys of saliva DNA viruses

BACKGROUND

Viruses are key players regulating microbial ecosystems. Exploration of viral assemblages is now possible thanks to the development of metagenomics, the most powerful tool available for studying viral ecology and discovering new viruses. Unfortunately, several sources of bias lead to the misrepresentation of certain viruses within metagenomics workflows, hindering the shift from merely descriptive studies towards quantitative comparisons of communities. Therefore, benchmark studies on virus enrichment and random amplification protocols are required to better understand the sources of bias.

RESULTS

We assessed the bias introduced by viral enrichment on mock assemblages composed of seven DNA viruses, and the bias from random amplification methods on human saliva DNA viromes, using qPCR and deep sequencing, respectively. While iodixanol cushions and 0.45 μm filtration preserved the original composition of nuclease-protected viral genomes, low-force centrifugation and 0.22 μm filtration removed large viruses. Comparison of unamplified and randomly amplified saliva viromes revealed that multiple displacement amplification (MDA) induced stochastic bias from picograms of DNA template. However, the type of bias shifted to systematic using 1 ng, with only a marginal influence by amplification time. Systematic bias consisted of over-amplification of small circular genomes, and under-amplification of those with extreme GC content, a negative bias that was shared with the PCR-based sequence-independent, single-primer amplification (SISPA) method. MDA based on random priming provided by a DNA primase activity slightly outperformed those based on random hexamers and SISPA, which may reflect differences in ability to handle sequences with extreme GC content. SISPA viromes showed uneven coverage profiles, with high coverage peaks in regions with low linguistic sequence complexity. Despite misrepresentation of certain viruses after random amplification, ordination plots based on dissimilarities among contig profiles showed perfect overlapping of related amplified and unamplified saliva viromes and strong separation from unrelated saliva viromes. This result suggests that random amplification bias has a minor impact on beta diversity studies.

CONCLUSIONS

Benchmark analyses of mock and natural communities of viruses improve understanding and mitigate bias in metagenomics surveys. Bias induced by random amplification methods has only a minor impact on beta diversity studies of human saliva viromes.

An optimized methodology for whole genome sequencing of RNA respiratory viruses from nasopharyngeal aspirates

Over the last decade, the number of viral genome sequences deposited in available databases has grown exponentially. However, sequencing methodology vary widely and many published works have relied on viral enrichment by viral culture or nucleic acid amplification with specific primers rather than through unbiased techniques such as metagenomics. The genome of RNA viruses is highly variable and these enrichment methodologies may be difficult to achieve or may bias the results. In order to obtain genomic sequences of human respiratory syncytial virus (HRSV) from positive nasopharyngeal aspirates diverse methodologies were evaluated and compared. A total of 29 nearly complete and complete viral genomes were obtained. The best performance was achieved with a DNase I treatment to the RNA directly extracted from the nasopharyngeal aspirate (NPA), sequence-independent single-primer amplification (SISPA) and library preparation performed with Nextera XT DNA Library Prep Kit with manual normalization. An average of 633,789 and 1,674,845 filtered reads per library were obtained with MiSeq and NextSeq 500 platforms, respectively. The higher output of NextSeq 500 was accompanied by the increasing of duplicated reads percentage generated during SISPA (from an average of 1.5% duplicated viral reads in MiSeq to an average of 74% in NextSeq 500). HRSV genome recovery was not affected by the presence or absence of duplicated reads but the computational demand during the analysis was increased. Considering that only samples with viral load ≥ E+06 copies/ml NPA were tested, no correlation between sample viral loads and number of total filtered reads was observed, nor with the mapped viral reads. The HRSV genomes showed a mean coverage of 98.46% with the best methodology. In addition, genomes of human metapneumovirus (HMPV), human rhinovirus (HRV) and human parainfluenza virus types 1–3 (HPIV1-3) were also obtained with the selected optimal methodology.

Characterization of the fecal virome in dogs with chronic enteropathy

The fecal virome has been investigated in humans and various animal species using next generation sequencing. However, limited information is available about the fecal virome of dogs with chronic enteropathy (CE). We aimed to characterize the canine fecal virome of dogs with CE and compare it with the virome of previously analyzed healthy dogs.A total of 16 adult dogs; 8 healthy dogs (data from a parallel study) and 8 dogs with CE had fecal samples assessed by viral shotgun sequencing. Fecal samples were subjected to enrichment of viral nucleic acids prior to sequencing and metagenomic analyses. Characterization of the complete genome of a canine kobuvirus was performed by Sanger sequencing. An additional 21 healthy dogs and 14 dogs with CE were further analyzed for the prevalence of canine kobuvirus.Three fecal samples from dogs with CE contained in total 3 eukaryotic viral families. In contrast, 4/8 fecal samples previously identified from healthy dogs, contained 5 eukaryotic viral families with 2 families exclusive to this group. Bacteriophages were identified in all fecal samples from CE and healthy dogs. Canine kobuvirus was identified in one dog with CE, by shotgun sequencing, and the complete genome was then characterized. This kobuvirus was classified within canine kobuvirus group, being similar to strains from Korea and China. The larger prevalence study did not detect additional samples positive for canine kobuvirus. The fecal virome of dogs with CE differs in number and type of viral families from healthy dogs. The first Australian canine kobuvirus sequence was identified and characterized from a dog with CE.

Methods for Enrichment and Sequencing of Oral Viral Assemblages: Saliva, Oral Mucosa, and Dental Plaque Viromes

The oral cavity is a major portal of entry for human pathogens including viruses. However, metagenomics has revealed that highly personalized and time-persistent bacteriophage assemblages dominate this habitat. Most oral bacteriophages follow lysogenic life cycles, deploying complex strategies to manage bacterial homeostasis. Although bacterial dysbiosis underlies common oral pathologies such as caries and periodontitis, the cause of these bacteria replacements remains obscure, and it is theorized that bacteriophages play an important role. The enormous sensitivity of metagenomics coupled with next-generation sequencing has made technically feasible to address the putative role of bacteriophages in oral dysbiosis and represents a valuable tool to discover new human viruses.This chapter proposes a workflow that consists of a simple viral enrichment protocol, two alternative random amplification methods, and next-generation sequencing to access virome composition in three oral environments: supragingival plaque, saliva, and mucosa. These protocols circumvent some well-known sources of bias, providing genomic information about DNA and RNA viral communities with minimal contamination from human and bacterial sources.

Single primer isothermal amplification (SPIA) combined with next generation sequencing provides complete bovine coronavirus genome coverage and higher sequence depth compared to sequence-independent single primer amplification (SISPA)

Coronaviruses are of major importance for both animal and human health. With the emergence of novel coronaviruses such as SARS and MERS, the need for fast genome characterisation is ever so important. Further, in order to understand the influence of quasispecies of these viruses in relation to biology, techniques for deep-sequence and full-length viral genome analysis are needed. In the present study, we compared the efficiency of two sequence-independent approaches [sequence-independent single primer amplification (SISPA) and single primer isothermal amplification (SPIA, represented by the Ovation kit)] coupled with high-throughput sequencing to generate the full-length genome of bovine coronavirus (BCoV) from a nasal swab. Both methods achieved high genome coverage (100% for SPIA and 99% for SISPA), however, there was a clear difference in the percentage of reads that mapped to BCoV. While approximately 45% of the Ovation reads mapped to BCoV (sequence depth of 169-284 944), only 0.07% of the SISPA reads (sequence depth of 0-249) mapped to the reference genome. Although BCoV was the focus of the study we also identified a bovine rhinitis B virus (BRBV) in the data sets. The trend for this virus was similar to that observed for BCoV regarding Ovation vs. SISPA, but with fewer sequences mapping to BRBV due to a lower amount of this virus. In summary, the SPIA approach used in this study produced coverage of the entire BCoV (high copy number) and BRBV (low copy number) and a high sequence/genome depth compared to SISPA. Although this is a limited study, the results indicate that the Ovation method could be a preferred approach for full genome sequencing if a low copy number of viral RNA is expected and if high sequence depth is desired.

Towards a Universal Molecular Microbiological Test

The standard paradigm for microbiological testing is dependent on the presentation of a patient to a clinician. Tests are then requested based on differential diagnoses using the patient's symptoms as a guide. The era of high-throughput genomic methods has the potential to replace this model for the first time with what could be referred to as "hypothesis-free testing." This approach utilizes one of a variety of methodologies to obtain a sequence from potentially any nucleic acid in a clinical sample, without prior knowledge of its content. We discuss the advantages of such an approach and the challenges in making this a reality.

Use of Sequence-Independent, Single-Primer-Amplification (SISPA) for rapid detection, identification, and characterization of avian RNA viruses

Current technologies with next generation sequencing have revolutionized metagenomics analysis of clinical samples. To achieve the non-selective amplification and recovery of low abundance genetic sequences, a simplified Sequence-Independent, Single-Primer Amplification (SISPA) technique in combination with MiSeq platform was applied to target negative- and positive-sense single-stranded RNA viral sequences. This method allowed successful sequence assembly of full or near full length avian influenza virus (AIV), infectious bronchitis virus (IBV), and Newcastle disease virus (NDV) viral genome. Moreover, SISPA analysis applied to unknown clinical cases of mixed viral infections produced genome assemblies comprising 98% NDV and 99% of IBV genomes. Complete or near complete virus genome sequence was obtained with titers at or above 10^4.5 EID₅₀/ml (50% embryo infectious dose), and virus identification could be detected with titers at or above 10³ EID₅₀/ml. Taken together, these studies demonstrate a simple template enrichment protocol for rapid detection and accurate characterization of avian RNA viruses.

•

A simple, random priming technique was optimized to target viral RNA genomes.

•

This technique allows characterization of multiple viruses in single reaction.

•

Complete or near complete genome sequence with titers at or above 10^4.5 EID₅₀/ml.

•

The detection limit with viral titers at or above 10³ EID₅₀/ml.

Characterisation of the canine faecal virome in healthy dogs and dogs with acute diarrhoea using shotgun metagenomics

The virome has been increasingly investigated in numerous animal species and in different sites of the body, facilitating the identification and discovery of a variety of viruses. In spite of this, the faecal virome of healthy dogs has not been investigated. In this study we describe the faecal virome of healthy dogs and dogs with acute diarrhoea in Australia, using a shotgun metagenomic approach. Viral sequences from a range of different virus families, including both RNA and DNA families, and known pathogens implicated in enteric disease were documented. Twelve viral families were identified, of which four were bacteriophages. Eight eukaryotic viral families were detected: Astroviridae, Coronaviridae, Reoviridae, Picornaviridae, Caliciviridae, Parvoviridae, Adenoviridae and Papillomaviridae. Families Astroviridae, Picornaviridae and Caliciviridae were found only in dogs with acute diarrhoea, with Astroviridae being the most common family identified in this group. Due to its prevalence, characterisation the complete genome of a canine astrovirus was performed. These studies indicate that metagenomic analyses are useful for the investigation of viral populations in the faeces of dogs. Further studies to elucidate the epidemiological and biological relevance of these findings are warranted.

The role of viruses in coral health and disease

Metagenomic and electron microscopy studies confirm that the coral microbiome contains a rich diversity and abundance of viruses. While there have been no definitive tests of disease causation by viruses in corals, viruses have been implicated as coral pathogens in a number of studies. Growing evidence also indicates that latent viral infections can compromise the algal symbionts under environmental stress and may be involved in the coral bleaching response. Conversely, bacteriophages and archaeal phage viruses are abundant in the microbiome of healthy corals and are likely to be involved in complex ecological networks, genetic material transfer and selective co-evolution within the surface mucus layers and tissues. The relative importance of viral control of bacterial and archaeal populations is unknown, but they are almost certain to be exerting some level of control on the composition and maintenance of the coral microbiome. While rapid leaps in the capability to detect viruses have been made due to advances in metagenomics and bioinformatics, these approaches need now to be integrated with in vitro culture and challenge experiments to assess the functional roles of viruses in health and disease, and it is imperative that interactions with other members of the coral microbiome are taken into account when assessing disease causation.

Aquatic viral metagenomics: Lights and shadows

Viruses are the most abundant biological entities on Earth, exceeding bacteria in most of the ecosystems. Specially in oceans, viruses are thought to be the major planktonic predators shaping microorganism communities and controlling ocean biological capacity. Plankton lysis by viruses plays an important role in ocean nutrient and energy cycles. Viral metagenomics has emerged as a powerful tool to uncover viral diversity in aquatic ecosystems through the use of Next Generation Sequencing. However, many of the commonly used viral sample preparation steps have several important biases that must be considered to avoid a misinterpretation of the results. In addition to biases caused by the purification of virus particles, viral DNA/RNA amplification and the preparation of genomic libraries could also introduce biases, and a detailed knowledge about such protocols is required. In this review, the main steps in the viral metagenomic workflow are described paying special attention to the potential biases introduced by each one.

Whole Genome Sequencing of Enterovirus species C Isolates by High-Throughput Sequencing: Development of Generic Primers

Enteroviruses are among the most common viruses infecting humans and can cause diverse clinical syndromes ranging from minor febrile illness to severe and potentially fatal diseases. Enterovirus species C (EV-C) consists of more than 20 types, among which the three serotypes of polioviruses, the etiological agents of poliomyelitis, are included. Biodiversity and evolution of EV-C genomes are shaped by frequent recombination events. Therefore, identification and characterization of circulating EV-C strains require the sequencing of different genomic regions. A simple method was developed to quickly sequence the entire genome of EV-C isolates. Four overlapping fragments were produced separately by RT-PCR performed with generic primers. The four amplicons were then pooled and purified prior to being sequenced by a high-throughput technique. The method was assessed on a panel of EV-Cs belonging to a wide-range of types. It can be used to determine full-length genome sequences through de novo assembly of thousands of reads. It was also able to discriminate reads from closely related viruses in mixtures. By decreasing the workload compared to classical Sanger-based techniques, this method will serve as a precious tool for sequencing large panels of EV-Cs isolated in cell cultures during environmental surveillance or from patients, including vaccine-derived polioviruses.

MetLab: An In Silico Experimental Design, Simulation and Analysis Tool for Viral Metagenomics Studies

Metagenomics, the sequence characterization of all genomes within a sample, is widely used as a virus discovery tool as well as a tool to study viral diversity of animals. Metagenomics can be considered to have three main steps; sample collection and preparation, sequencing and finally bioinformatics. Bioinformatic analysis of metagenomic datasets is in itself a complex process, involving few standardized methodologies, thereby hampering comparison of metagenomics studies between research groups. In this publication the new bioinformatics framework MetLab is presented, aimed at providing scientists with an integrated tool for experimental design and analysis of viral metagenomes. MetLab provides support in designing the metagenomics experiment by estimating the sequencing depth needed for the complete coverage of a species. This is achieved by applying a methodology to calculate the probability of coverage using an adaptation of Stevens’ theorem. It also provides scientists with several pipelines aimed at simplifying the analysis of viral metagenomes, including; quality control, assembly and taxonomic binning. We also implement a tool for simulating metagenomics datasets from several sequencing platforms. The overall aim is to provide virologists with an easy to use tool for designing, simulating and analyzing viral metagenomes. The results presented here include a benchmark towards other existing software, with emphasis on detection of viruses as well as speed of applications. This is packaged, as comprehensive software, readily available for Linux and OSX users at https://github.com/norling/metlab.

Measurements of Intrahost Viral Diversity Are Extremely Sensitive to Systematic Errors in Variant Calling

With next-generation sequencing technologies, it is now feasible to efficiently sequence patient-derived virus populations at a depth of coverage sufficient to detect rare variants. However, each sequencing platform has characteristic error profiles, and sample collection, target amplification, and library preparation are additional processes whereby errors are introduced and propagated. Many studies account for these errors by using ad hoc quality thresholds and/or previously published statistical algorithms. Despite common usage, the majority of these approaches have not been validated under conditions that characterize many studies of intrahost diversity. Here, we use defined populations of influenza virus to mimic the diversity and titer typically found in patient-derived samples. We identified single-nucleotide variants using two commonly employed variant callers, DeepSNV and LoFreq. We found that the accuracy of these variant callers was lower than expected and exquisitely sensitive to the input titer. Small reductions in specificity had a significant impact on the number of minority variants identified and subsequent measures of diversity. We were able to increase the specificity of DeepSNV to >99.95% by applying an empirically validated set of quality thresholds. When applied to a set of influenza virus samples from a household-based cohort study, these changes resulted in a 10-fold reduction in measurements of viral diversity. We have made our sequence data and analysis code available so that others may improve on our work and use our data set to benchmark their own bioinformatics pipelines. Our work demonstrates that inadequate quality control and validation can lead to significant overestimation of intrahost diversity.

IMPORTANCE

Advances in sequencing technology have made it feasible to sequence patient-derived viral samples at a level sufficient for detection of rare mutations. These high-throughput, cost-effective methods are revolutionizing the study of within-host viral diversity. However, the techniques are error prone, and the methods commonly used to control for these errors have not been validated under the conditions that characterize patient-derived samples. Here, we show that these conditions affect measurements of viral diversity. We found that the accuracy of previously benchmarked analysis pipelines was greatly reduced under patient-derived conditions. By carefully validating our sequencing analysis using known control samples, we were able to identify biases in our method and to improve our accuracy to acceptable levels. Application of our modified pipeline to a set of influenza virus samples from a cohort study provided a realistic picture of intrahost diversity and suggested the need for rigorous quality control in such studies.

Development and evaluation of a non-ribosomal random PCR and next-generation sequencing based assay for detection and sequencing of hand, foot and mouth disease pathogens

Background

Hand, foot and mouth disease (HFMD) has become a major public health problem across the Asia-Pacific region, and is commonly caused by enterovirus A71 (EV-A71) and coxsackievirus A6 (CV-A6), CV-A10 and CV-A16. Generating pathogen whole-genome sequences is essential for understanding their evolutionary biology. The frequent replacements among EV serotypes and a limited numbers of available whole-genome sequences hinder the development of overlapping PCRs for whole-genome sequencing.

We developed and evaluated a non-ribosomal random PCR (rPCR) and next-generation sequencing based assay for sequence-independent whole-genome amplification and sequencing of HFMD pathogens. A total of 16 EV-A71/CV-A6/CV-A10/CV-A16 PCR positive rectal/throat swabs (Cp values: 20.9–33.3) were used for assay evaluation.

Results

Our assay evidently outperformed the conventional rPCR in terms of the total number of EV-A71 reads and the percentage of EV-A71 reads: 2.6 % (1275/50,000 reads) vs. 0.1 % (31/50,000) and 6 % (3008/50,000) vs. 0.9 % (433/50,000) for two samples with Cp values of 30 and 26, respectively. Additionally the assay could generate genome sequences with the percentages of coverage of 94–100 % of 4 different enterovirus serotypes in 73 % of the tested samples, representing the first whole-genome sequences of CV-A6/10/16 from Vietnam, and could assign correctly serotyping results in 100 % of 24 tested specimens. In all but three the obtained consensuses of two replicates from the same sample were 100 % identical, suggesting that our assay is highly reproducible.

Conclusions

In conclusion, we have successfully developed a non-ribosomal rPCR and next-generation sequencing based assay for sensitive detection and direct whole-genome sequencing of HFMD pathogens from clinical samples.

Electronic supplementary material

The online version of this article (doi:10.1186/s12985-016-0580-9) contains supplementary material, which is available to authorized users.

RNA viruses as major contributors to Antarctic virioplankton

Early work on marine algal viruses focused exclusively on those having DNA genomes, but recent studies suggest that RNA viruses, especially those with positive-sense, single-stranded RNA (+ssRNA) genomes, are abundant in tropical and temperate coastal seawater. To test whether this was also true of polar waters, we estimated the relative abundances of RNA and DNA viruses using a mass ratio approach and conducted shotgun metagenomics on purified viral samples collected from a coastal site near Palmer Station, Antarctica on six occasions throughout a summer phytoplankton bloom (November-March). Our data suggest that RNA viruses contributed up to 65% of the total virioplankton (8-65%), and that, as observed previously in warmer waters, the majority of RNA viruses in these Antarctic RNA virus metagenomes had +ssRNA genomes most closely related to viruses in the order Picornavirales. Assembly of the metagenomic reads resulted in five novel, nearly complete genomes, three of which had features similar to diatom-infecting viruses. Our data are consistent with the hypothesis that RNA viruses influence diatom bloom dynamics in Antarctic waters.

The Intestinal Eukaryotic Virome in Healthy and Diarrhoeic Neonatal Piglets

Neonatal porcine diarrhoea of uncertain aetiology has been reported from a number of European countries. The aim of the present study was to use viral metagenomics to examine a potential viral involvement in this diarrhoea and to describe the intestinal virome with focus on eukaryotic viruses. Samples from the distal jejunum of 50 diarrhoeic and 19 healthy piglets from 10 affected herds were analysed. The viral fraction of the samples was isolated and nucleic acids (RNA and DNA fractions) were subjected to sequence independent amplification. Samples from diarrhoeic piglets from the same herds were pooled whereas samples from healthy piglets were analysed individually. In total, 29 clinical samples, plus two negative controls and one positive control consisting of a mock metagenome were sequenced using the Ion Torrent platform. The resulting sequence data was subjected to taxonomic classification using Kraken, Diamond and HMMER. In the healthy specimens, eight different mammalian virus families were detected (Adenoviridae, Anelloviridae, Astroviridae, Caliciviridae, Circoviridae, Parvoviridae, Picornaviridae, and Reoviridae) compared to four in the pooled diarrhoeic samples (Anelloviridae, Circoviridae, Picornaviridae, and Reoviridae). It was not possible to associate a particular virus family with the investigated diarrhoea. In conclusion, this study does not support the hypothesis that the investigated diarrhoea was caused by known mammalian viruses. The results do, however, indicate that known mammalian viruses were present in the intestine as early as 24–48 hours after birth, indicating immediate infection post-partum or possibly transplacental infection.

Diversity and Ecology of Viruses in Hyperarid Desert Soils

In recent years, remarkable progress has been made in the field of virus environmental ecology. In marine ecosystems, for example, viruses are now thought to play pivotal roles in the biogeochemical cycling of nutrients and to be mediators of microbial evolution through horizontal gene transfer. The diversity and ecology of viruses in soils are poorly understood, but evidence supports the view that the diversity and ecology of viruses in soils differ substantially from those in aquatic systems. Desert biomes cover ∼ 33% of global land masses, and yet the diversity and roles of viruses in these dominant ecosystems remain poorly understood. There is evidence that hot hyperarid desert soils are characterized by high levels of bacterial lysogens and low extracellular virus counts. In contrast, cold desert soils contain high extracellular virus titers. We suggest that the prevalence of microbial biofilms in hyperarid soils, combined with extreme thermal regimens, exerts strong selection pressures on both temperate and virulent viruses. Many desert soil virus sequences show low values of identity to virus genomes in public databases, suggesting the existence of distinct and as-yet-uncharacterized soil phylogenetic lineages (e.g., cyanophages). We strongly advocate for amplification-free metavirome analyses while encouraging the classical isolation of phages from dominant and culturable microbial isolates in order to populate sequence databases. This review provides an overview of recent advances in the study of viruses in hyperarid soils and of the factors that contribute to viral abundance and diversity in hot and cold deserts and offers technical recommendations for future studies.

Host-Associated Metagenomics: A Guide to Generating Infectious RNA Viromes

BACKGROUND

Metagenomic analyses have been widely used in the last decade to describe viral communities in various environments or to identify the etiology of human, animal, and plant pathologies. Here, we present a simple and standardized protocol that allows for the purification and sequencing of RNA viromes from complex biological samples with an important reduction of host DNA and RNA contaminants, while preserving the infectivity of viral particles.

PRINCIPAL FINDINGS

We evaluated different viral purification steps, random reverse transcriptions and sequence-independent amplifications of a pool of representative RNA viruses. Viruses remained infectious after the purification process. We then validated the protocol by sequencing the RNA virome of human body lice engorged in vitro with artificially contaminated human blood. The full genomes of the most abundant viruses absorbed by the lice during the blood meal were successfully sequenced. Interestingly, random amplifications differed in the genome coverage of segmented RNA viruses. Moreover, the majority of reads were taxonomically identified, and only 7-15% of all reads were classified as "unknown", depending on the random amplification method.

CONCLUSION

The protocol reported here could easily be applied to generate RNA viral metagenomes from complex biological samples of different origins. Our protocol allows further virological characterizations of the described viral communities because it preserves the infectivity of viral particles and allows for the isolation of viruses.

Recovering full-length viral genomes from metagenomes

Infectious disease metagenomics is driven by the question: “what is causing the disease?” in contrast to classical metagenome studies which are guided by “what is out there?” In case of a novel virus, a first step to eventually establishing etiology can be to recover a full-length viral genome from a metagenomic sample. However, retrieval of a full-length genome of a divergent virus is technically challenging and can be time-consuming and costly. Here we discuss different assembly and fragment linkage strategies such as iterative assembly, motif searches, k-mer frequency profiling, coverage profile binning, and other strategies used to recover genomes of potential viral pathogens in a timely and cost-effective manner.

Molecular characterization of Tasmanian aquabirnaviruses from 1998 to 2013

Tasmanian aquabirnaviruses (TABVs) have been isolated intermittently since 1998 from healthy Atlantic salmon Salmo salar and rainbow trout Oncorhynchus mykiss farmed in Macquarie Harbour, Tasmania, Australia. However, beginning in 2011, TABVs have been isolated from rainbow trout in association with mortality events. To determine if recent molecular changes in TABV were contributing to increased mortalities, next generation sequencing was undertaken on 14 TABVs isolated from 1998 to 2013. Sequencing of both genome segments and analysis of the 5 viral proteins they encode revealed that minimal changes had occurred in the past 15 yr. Of the amino acid changes detected only 1, alanine to aspartic acid at position 139 of the minor structural VP3 protein, was unique to the recent disease events. The most dramatic changes observed were in the length of the non-structural VP5 protein varying from 43 to 133 amino acids. However, the amino acid substitution in VP3 and variable VP5 length were unlikely to have resulted in increased TABV pathogenicity. The genome of a novel Australian aquabirnavirus, Victorian trout aquabirnavirus (VTAB) was also sequenced and compared to TABV isolates.

Near full length hepatitis C virus genome reconstruction by next generation sequencing based on genotype-independent amplification

BACKGROUND

Deep sequencing has a deep impact on the study of rapidly mutating RNA viruses, such as hepatitis C virus, proving to be an invaluable tool for analyzing virus diversity and evolution.

AIM

Genotype-independent high-throughput pyrosequencing was used to obtain near full length hepatitis C virus genome sequence reconstruction directly from clinical samples.

METHODS

Samples from hepatitis C virus infected subjects harbouring different subtypes (1a, 1b, 2c) were analyzed (viral load range: 1.2-20.8 × 10(6)IU/ml). Data were generated with a modified sequence-independent single primer amplification method followed by 454 sequencing.

RESULTS

the extent of reconstructed hepatitis C virus genome varied from 79.95% to 99.64%. No correlation between extent of genome reconstruction and either viral load (r=0.4857, p=0.3556) or number of HCV reads (r=0.08571, p=0.9194) was observed.

CONCLUSION

This study describes a protocol for obtaining whole genome sequences from different hepatitis C virus patients with different genotypes in a single sequencing run.

Harnessing NGS and Big Data Optimally: Comparison of miRNA Prediction from Assembled versus Non-assembled Sequencing Data--The Case of the Grass Aegilops tauschii Complex Genome

MicroRNAs (miRNAs) are small, endogenous, non-coding RNA molecules that regulate gene expression at the post-transcriptional level. As high-throughput next generation sequencing (NGS) and Big Data rapidly accumulate for various species, efforts for in silico identification of miRNAs intensify. Surprisingly, the effect of the input genomics sequence on the robustness of miRNA prediction was not evaluated in detail to date. In the present study, we performed a homology-based miRNA and isomiRNA prediction of the 5D chromosome of bread wheat progenitor, Aegilops tauschii, using two distinct sequence data sets as input: (1) raw sequence reads obtained from 454-GS FLX Titanium sequencing platform and (2) an assembly constructed from these reads. We also compared this method with a number of available plant sequence datasets. We report here the identification of 62 and 22 miRNAs from raw reads and the assembly, respectively, of which 16 were predicted with high confidence from both datasets. While raw reads promoted sensitivity with the high number of miRNAs predicted, 55% (12 out of 22) of the assembly-based predictions were supported by previous observations, bringing specificity forward compared to the read-based predictions, of which only 37% were supported. Importantly, raw reads could identify several repeat-related miRNAs that could not be detected with the assembly. However, raw reads could not capture 6 miRNAs, for which the stem-loops could only be covered by the relatively longer sequences from the assembly. In summary, the comparison of miRNA datasets obtained by these two strategies revealed that utilization of raw reads, as well as assemblies for in silico prediction, have distinct advantages and disadvantages. Consideration of these important nuances can benefit future miRNA identification efforts in the current age of NGS and Big Data driven life sciences innovation.

Evaluation of convenient pretreatment protocols for RNA virus metagenomics in serum and tissue samples

Viral metagenomic approaches are increasingly being used for viral discovery. Various strategies are applied to enrich viral sequences, but there is often a lack of knowledge about their effective influence on the viral discovery sensitivity. We evaluate some convenient and widely used approaches for RNA virus discovery in clinical samples in order to reveal their sensitivity and potential bias introduced by the enrichment or amplifications steps. An RNA virus was artificially spiked at a fixed titer in serum and lung tissue, respectively, low and high nucleic acid content matrices. For serum, a simple DNase treatment on the RNA extract gave the maximum gain in proportion of viral sequences (83×), and a subsequent ribosomal RNA removal nearly doubled once more the proportion of viral sequences. For lung tissue, a ribosomal RNA depletion step on the RNA extract had the biggest gain in proportion of viral sequences (32×). We show also that direct sequencing of cDNA is recommended above an extra random PCR amplification step, and a that the virion enrichment strategy (filtration and nuclease treatment) has a beneficial effect for sequencing-based virus discovery. Our findings provide sample-dependent guidelines for targeted virus discovery strategies.

Accurate read-based metagenome characterization using a hierarchical suite of unique signatures

A major challenge in the field of shotgun metagenomics is the accurate identification of organisms present within a microbial community, based on classification of short sequence reads. Though existing microbial community profiling methods have attempted to rapidly classify the millions of reads output from modern sequencers, the combination of incomplete databases, similarity among otherwise divergent genomes, errors and biases in sequencing technologies, and the large volumes of sequencing data required for metagenome sequencing has led to unacceptably high false discovery rates (FDR). Here, we present the application of a novel, gene-independent and signature-based metagenomic taxonomic profiling method with significantly and consistently smaller FDR than any other available method. Our algorithm circumvents false positives using a series of non-redundant signature databases and examines Genomic Origins Through Taxonomic CHAllenge (GOTTCHA). GOTTCHA was tested and validated on 20 synthetic and mock datasets ranging in community composition and complexity, was applied successfully to data generated from spiked environmental and clinical samples, and robustly demonstrates superior performance compared with other available tools.

SISPA-Seq for rapid whole genome surveys of bacterial isolates

Whole genome sequencing (WGS) of large isolate collections has many applications, yet sequencing costs are still significant. We sought to develop a rapid and cost efficient WGS method to address fundamental questions in clinical microbiology. We evaluated the performance of SISPA (Sequence-Independent, Single-Primer Amplification) combined with next-generation sequencing (SISPA-Seq) of 75 clinical isolates of Acinetobacter baumannii to establish whether SISPA-Seq resulted in sufficient coverage and quality to (1) determine strain phylogenetic placement and (2) and carriage of known antibiotic resistance (AbR) genes. Strains for which whole genome sequences were available were included for validation. Two libraries for each strain were constructed from separate SISPA reactions with different barcoded primers, using genomic DNA prepared from either high quality or rapid heat-lysis preparations. SISPA-Seq resulted in a median of 65× genome coverage when reads from both primer sets were combined. Coverage and quality were sufficient for detection of AbR genes by comparison of reads to the ARG-ANNOT database and were often sufficient to distinguish between different allelic variants of the same gene. kSNP and RAxML were used to construct a robust phylogeny based on single-nucleotide variants (SNVs) that showed that the SISPA-Seq data was sufficient for sensitive and accurate phylogenetic placement. Advantages of the SISPA-Seq method include inexpensive and rapid DNA preparation and a typical total cost less than one-half that of standard genome sequencing. In summary, SISPA-Seq can be used to survey whole genomes of a large strain collection and identify strains that should be targeted for additional sequencing.

Assembly of viral genomes from metagenomes

Viral infections remain a serious global health issue. Metagenomic approaches are increasingly used in the detection of novel viral pathogens but also to generate complete genomes of uncultivated viruses. In silico identification of complete viral genomes from sequence data would allow rapid phylogenetic characterization of these new viruses. Often, however, complete viral genomes are not recovered, but rather several distinct contigs derived from a single entity are, some of which have no sequence homology to any known proteins. De novo assembly of single viruses from a metagenome is challenging, not only because of the lack of a reference genome, but also because of intrapopulation variation and uneven or insufficient coverage. Here we explored different assembly algorithms, remote homology searches, genome-specific sequence motifs, k-mer frequency ranking, and coverage profile binning to detect and obtain viral target genomes from metagenomes. All methods were tested on 454-generated sequencing datasets containing three recently described RNA viruses with a relatively large genome which were divergent to previously known viruses from the viral families Rhabdoviridae and Coronaviridae. Depending on specific characteristics of the target virus and the metagenomic community, different assembly and in silico gap closure strategies were successful in obtaining near complete viral genomes.

Metagenomic survey for viruses in Western Arctic caribou, Alaska, through iterative assembly of taxonomic units

Pathogen surveillance in animals does not provide a sufficient level of vigilance because it is generally confined to surveillance of pathogens with known economic impact in domestic animals and practically nonexistent in wildlife species. As most (re-)emerging viral infections originate from animal sources, it is important to obtain insight into viral pathogens present in the wildlife reservoir from a public health perspective. When monitoring living, free-ranging wildlife for viruses, sample collection can be challenging and availability of nucleic acids isolated from samples is often limited. The development of viral metagenomics platforms allows a more comprehensive inventory of viruses present in wildlife. We report a metagenomic viral survey of the Western Arctic herd of barren ground caribou (Rangifer tarandus granti) in Alaska, USA. The presence of mammalian viruses in eye and nose swabs of 39 free-ranging caribou was investigated by random amplification combined with a metagenomic analysis approach that applied exhaustive iterative assembly of sequencing results to define taxonomic units of each metagenome. Through homology search methods we identified the presence of several mammalian viruses, including different papillomaviruses, a novel parvovirus, polyomavirus, and a virus that potentially represents a member of a novel genus in the family Coronaviridae.

Next-generation sequencing in veterinary medicine: how can the massive amount of information arising from high-throughput technologies improve diagnosis, control, and management of infectious diseases?

The development of high-throughput molecular technologies and associated bioinformatics has dramatically changed the capacities of scientists to produce, handle, and analyze large amounts of genomic, transcriptomic, and proteomic data. A clear example of this step-change is represented by the amount of DNA sequence data that can be now produced using next-generation sequencing (NGS) platforms. Similarly, recent improvements in protein and peptide separation efficiencies and highly accurate mass spectrometry have promoted the identification and quantification of proteins in a given sample. These advancements in biotechnology have increasingly been applied to the study of animal infectious diseases and are beginning to revolutionize the way that biological and evolutionary processes can be studied at the molecular level. Studies have demonstrated the value of NGS technologies for molecular characterization, ranging from metagenomic characterization of unknown pathogens or microbial communities to molecular epidemiology and evolution of viral quasispecies. Moreover, high-throughput technologies now allow detailed studies of host-pathogen interactions at the level of their genomes (genomics), transcriptomes (transcriptomics), or proteomes (proteomics). Ultimately, the interaction between pathogen and host biological networks can be questioned by analytically integrating these levels (integrative OMICS and systems biology). The application of high-throughput biotechnology platforms in these fields and their typical low-cost per information content has revolutionized the resolution with which these processes can now be studied. The aim of this chapter is to provide a current and prospective view on the opportunities and challenges associated with the application of massive parallel sequencing technologies to veterinary medicine, with particular focus on applications that have a potential impact on disease control and management.

Generating viral metagenomes from the coral holobiont

Reef-building corals comprise multipartite symbioses where the cnidarian animal is host to an array of eukaryotic and prokaryotic organisms, and the viruses that infect them. These viruses are critical elements of the coral holobiont, serving not only as agents of mortality, but also as potential vectors for lateral gene flow, and as elements encoding a variety of auxiliary metabolic functions. Consequently, understanding the functioning and health of the coral holobiont requires detailed knowledge of the associated viral assemblage and its function. Currently, the most tractable way of uncovering viral diversity and function is through metagenomic approaches, which is inherently difficult in corals because of the complex holobiont community, an extracellular mucus layer that all corals secrete, and the variety of sizes and structures of nucleic acids found in viruses. Here we present the first protocol for isolating, purifying and amplifying viral nucleic acids from corals based on mechanical disruption of cells. This method produces at least 50% higher yields of viral nucleic acids, has very low levels of cellular sequence contamination and captures wider viral diversity than previously used chemical-based extraction methods. We demonstrate that our mechanical-based method profiles a greater diversity of DNA and RNA genomes, including virus groups such as Retro-transcribing and ssRNA viruses, which are absent from metagenomes generated via chemical-based methods. In addition, we briefly present (and make publically available) the first paired DNA and RNA viral metagenomes from the coral Acropora tenuis.