MARVEL, a Tool for Prediction of Bacteriophage Sequences in Metagenomic Bins

Amgarten, Deyvid; Braga, Lucas Palma Perez; Silva, da; Setúbal, João Carlos

doi:10.3389/fgene.2018.00304

Cited by 135 publications

(112 citation statements)

References 53 publications

Supporting

Mentioning

112

Contrasting

Order By: Relevance

“…We chose VirSorter over virfinder (Ren, Ahlgren, Lu, Fuhrman, & Sun, ) because it has been shown that the later may misclassify eukaryotic sequences as viral. Some other approaches have been recently developed to retrieve viral signals from (meta‐)genomic data, such as marvel (Amgarten, Braga, da Silva, & Setubal, ) and virminer (Zheng et al, ), but they have been developed to detect viral genomes in prokaryotes. From the 64 viral contigs retrieved in the protist cells, the narrow host range of these viruses was remarkable given that >95% of the detected viral sequences ( n = 61) were specific to one stramenopile lineage and just a few were shared between lineages ( n = 3).…”

Section: Discussionmentioning

confidence: 99%

Assessing the viral content of uncultured picoeukaryotes in the global‐ocean by single cell genomics

et al. 2019

View full text Add to dashboard Cite

Viruses are the most abundant biological entities on Earth and have fundamental ecological roles in controlling microbial communities. Yet, although their diversity is being increasingly explored, little is known about the extent of viral interactions with their protist hosts as most studies are limited to a few cultivated species. Here, we exploit the potential of single‐cell genomics to unveil viral associations in 65 individual cells of 11 essentially uncultured stramenopiles lineages sampled during the Tara Oceans expedition. We identified viral signals in 57% of the cells, covering nearly every lineage and with narrow host specificity signal. Only seven out of the 64 detected viruses displayed homologies to known viral sequences. A search for our viral sequences in global ocean metagenomes showed that they were preferentially found at the DCM and within the 0.2–3 µm size fraction. Some of the viral signals were widely distributed, while others geographically constrained. Among the viral signals we detected an endogenous mavirus virophage potentially integrated within the nuclear genome of two distant uncultured stramenopiles. Virophages have been previously reported as a cell's defence mechanism against other viruses, and may therefore play an important ecological role in regulating protist populations. Our results point to single‐cell genomics as a powerful tool to investigate viral associations in uncultured protists, suggesting a wide distribution of these relationships, and providing new insights into the global viral diversity.

show abstract

Section: Discussionmentioning

confidence: 99%

Assessing the viral content of uncultured picoeukaryotes in the global‐ocean by single cell genomics

et al. 2019

View full text Add to dashboard Cite

show abstract

“…K-mer similarity score can discriminate viruses within other prokaryotic genomes even the viral reads is short(500 bp). Furthermore, MAR-VEL Amgarten et al (2018) tool is used for identifying bacteriophage sequences in metagenomic bins. Random forest approach is applied in MARVEL and model input is various engineered features, such as gene density, strand shifts, and fractions of significant hits to a viral protein database Grazziotin et al (2016).…”

Section: Machine Learning Toolsmentioning

confidence: 99%

Viral Sequence Identification in Metagenomes using Natural Language Processing Techniques

Abdelkareem

Khalil

Elbehery

et al. 2020

Preprint

View full text Add to dashboard Cite

A B S T R A C TViral reads identification is one of the important steps in metagenomic data analysis. It shows up the diversity of the microbial communities and the functional characteristics of microorganisms. There are various tools that can identify viral reads in mixed metagenomic data using similarity and statistical tools. However, the lack of available genome diversity is a serious limitation to the existing techniques. In this work, we applied natural language processing approaches for document classification in analyzing metagenomic sequences. Text featurization is presented by treating DNA similar to natural language. These techniques reveal the importance of using the text feature extraction pipeline in sequence identification by transforming DNA base pairs into a set of characters with a term frequency and inverse document frequency techniques. Various machine learning classification algorithms are applied to viral identification tasks such as logistic regression and multi-layer perceptron. Moreover, we compared classical machine learning algorithms with VirFinder and VirNet, our deep attention model for viral reads identification on generated fragments of viruses and bacteria for benchmarking viral reads identification tools. Then, as a verification of our tool, It was applied to a simulated microbiome and virome data for tool verification and real metagenomic data of Roche 454 and Illumina for a case study.

show abstract

“…A different approach consists of using machine learning techniques to learn from examples to classify viral genomes and to generalize to novel samples. In particular, several machine learning models for detecting viruses in metagenomic data have been already published [23,24,25], but none of them were trained nor tested to identify viruses in different human biospecimens.…”

Section: Introductionmentioning

confidence: 99%

“…Currently, the detection of potential viral genomes in human biospecimens is usually 19 performed by NCBI BLAST, which implements alignment-based classification where 20 sequences are aligned to known genomes from public databases and then estimates how 21 much percentage similarity they share. However, metagenomic samples might contain a 22 large number of highly divergent viruses that have no homologs at all among known 23 genomes. As a consequence, many sequences generated from NGS technologies are 24 classified as "unknown" by BLAST [5,18].…”

Section: Introductionmentioning

confidence: 99%

“…A different approach consists of using machine 32 learning techniques to learn from examples to classify viral genomes and to generalize to 33 novel samples. In particular, several machine learning models for detecting viruses in 34 metagenomic data have been already published [23][24][25], but none of them were trained 35 nor tested to identify viruses in different human biospecimens. 36 In this study, we have developed ViraMiner, a methodology which uses Convolutional 37 Neural Networks (CNN) on raw metagenomic contigs to identify potential viral 38 sequences across different human samples.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

ViraMiner: Deep Learning on Raw DNA Sequences for Identifying Viral Genomes in Human Samples

Tampuu

Bzhalava

Dillner

et al. 2019

Preprint

View full text Add to dashboard Cite

Despite its clinical importance, detection of highly divergent or yet unknown viruses is a major challenge. When human samples are sequenced, conventional alignments classify many assembled contigs as "unknown" since many of the sequences are not similar to known genomes. In this work, we developed ViraMiner, a deep learning-based method to identify viruses in various human biospecimens. ViraMiner contains two branches of Convolutional Neural Networks designed to detect both patterns and pattern-frequencies on raw metagenomics contigs. The training dataset included sequences obtained from 19 metagenomic experiments which were analyzed and labeled by BLAST. The model achieves significantly improved accuracy compared to other machine learning methods for viral genome classification. Using 300 bp contigs ViraMiner achieves 0.923 area under the ROC curve. To our knowledge, this is the first machine learning methodology that can detect the presence of viral sequences among raw metagenomic contigs from diverse human samples. We suggest that the proposed model captures different types of information of genome composition, and can be used as a recommendation system to further investigate sequences labeled as "unknown" by conventional alignment methods. Exploring these highly-divergent viruses, in turn, can enhance our knowledge of infectious causes of diseases.

show abstract

MARVEL, a Tool for Prediction of Bacteriophage Sequences in Metagenomic Bins

Cited by 135 publications

References 53 publications

Assessing the viral content of uncultured picoeukaryotes in the global‐ocean by single cell genomics

Assessing the viral content of uncultured picoeukaryotes in the global‐ocean by single cell genomics

Viral Sequence Identification in Metagenomes using Natural Language Processing Techniques

ViraMiner: Deep Learning on Raw DNA Sequences for Identifying Viral Genomes in Human Samples

Contact Info

Product

Resources

About