PhD Projects » ESR 12: A machine learning pipeline to identify and characterise novel viruses in metagenome data

@article{Wijesekara2024,

title = {Jaeger: an accurate and fast deep-learning tool to detect bacteriophage sequences},

author = {Yasas Wijesekara and Ling-Yi Wu and Rick Beeloo and Piotr Rozwalak and Ernestina Hauptfeld and Swapnil P. Doijad and Bas E. Dutilh and Lars Kaderali},

year  = {2024},

date = {2024-09-24},

journal = {bioRxiv},

abstract = {Viruses are integral to every biome on Earth, yet we still need a more comprehensive picture of their identity and global distribution. Global metagenomics sequencing efforts revealed the genomic content of tens of thousands of environmental samples, however identifying the viral sequences in these datasets remains challenging due to their vast genomic diversity. Here, we address identifying bacteriophage sequences in unlabeled sequencing data. In a recent benchmarking paper, we observed that existing deep-learning tools show a high true positive rate, but may also produce many false positives when confronted with divergent sequences. To tackle this challenge, we introduce Jaeger, a novel deep-learning method designed specifically for identifying bacteriophage genome fragments. Extensive benchmarking on the IMG/VR database and real-world metagenomes reveals Jaeger’s consistent high sensitivity (0.87) and precision (0.92). Applying Jaeger to over 16,000 metagenomic assemblies from the MGnify database yielded over five million putative phage contigs. On average, Jaeger is around 20 times faster than the other state-of-the-art methods. Jaeger is available at https://github.com/MGXlab/Jaeger.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{nokey,

title = {Ultraconserved bacteriophage genome sequence identified in 1300-year-old human palaeofaeces},

author = {Piotr Rozwalak and Jakub Barylski and Yasas Wijesekara and Bas E. Dutilh and Andrzej Zielezinski

},

doi = {10.1038/s41467-023-44370-0},

year  = {2024},

date = {2024-01-23},

urldate = {2024-01-23},

journal = {Nat Commun},

volume = {15},

issue = {1},

pages = {495},

abstract = {Bacteriophages are widely recognised as rapidly evolving biological entities. However, knowledge about ancient bacteriophages is limited. Here, we analyse DNA sequence datasets previously generated from ancient palaeofaeces and human gut-content samples, and identify an ancient phage genome nearly identical to present-day Mushuvirus mushu, a virus that infects gut commensal bacteria. The DNA damage patterns of the genome are consistent with its ancient origin and, despite 1300 years of evolution, the ancient Mushuvirus genome shares 97.7% nucleotide identity with its modern counterpart, indicating a long-term relationship between the prophage and its host. In addition, we reconstruct and authenticate 297 other phage genomes from the last 5300 years, including those belonging to unknown families. Our findings demonstrate the feasibility of reconstructing ancient phage genome sequences, thus expanding the known virosphere and offering insights into phage-bacteria interactions spanning several millennia.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{nokey,

title = {The International Virus Bioinformatics Meeting 2023.},

author = {Franziska Hufsky and Ana B. Abecasis and Artem Babaian and Sebastian Beck and Liam Brierley and Simon Dellicour and Christian Eggeling and Santiago F. Elena and Udo Gieraths and Anh D. Ha and Will Harvey and Terry C. Jones and Kevin Lamkiewicz and Gabriel Lencioni Lovate and Dominik Lücking and Martin Machyna and Luca Nishimura and Maximilian K. Nocke and Bernard Y. Renard and Shoichi Sakaguchi and Lygeri Sakellaridi and Jannes Spangenberg and Maria Tarradas-Alemany and Sandra Triebel and Yulia Vakulenko and Yasas Wijesekara and Fernando González-Candelas and Sarah Krautwurst and Alba Pérez-Cataluña and Walter Randazzo and Gloria Sánchez and Manja Marz },

doi = {10.3390/v15102031},

year  = {2023},

date = {2023-09-30},

urldate = {2023-09-30},

journal = {Viruses},

volume = {15},

issue = {10},

pages = {2031},

abstract = {The 2023 International Virus Bioinformatics Meeting was held in Valencia, Spain, from 24-26 May 2023, attracting approximately 180 participants worldwide. The primary objective of the conference was to establish a dynamic scientific environment conducive to discussion, collaboration, and the generation of novel research ideas. As the first in-person event following the SARS-CoV-2 pandemic, the meeting facilitated highly interactive exchanges among attendees. It served as a pivotal gathering for gaining insights into the current status of virus bioinformatics research and engaging with leading researchers and emerging scientists. The event comprised eight invited talks, 19 contributed talks, and 74 poster presentations across eleven sessions spanning three days. Topics covered included machine learning, bacteriophages, virus discovery, virus classification, virus visualization, viral infection, viromics, molecular epidemiology, phylodynamic analysis, RNA viruses, viral sequence analysis, viral surveillance, and metagenomics. This report provides rewritten abstracts of the presentations, a summary of the key research findings, and highlights shared during the meeting.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{nokey,

title = {Benchmarking Bioinformatic Virus Identification Tools Using Real-World Metagenomic Data across Biomes},

author = {Ling-Yi Wu and Nikolaos Pappas and Yasas Wijesekara and Gonçalo J. Piedade and Corina P.D. Brussaard and Bas E. Dutilh},

doi = {10.1101/2023.04.26.538077},

year  = {2023},

date = {2023-04-28},

journal = {bioRxiv},

abstract = {As most viruses remain uncultivated, metagenomics is currently the main method for virus discovery. Detecting viruses in metagenomic data is not trivial. In the past few years, many bioinformatic virus identification tools have been developed for this task, making it challenging to choose the right tools, parameters, and cutoffs. As all these tools measure different biological signals, and use different algorithms and training/reference databases, it is imperative to conduct an independent benchmarking to give users objective guidance. We compared the performance of ten state-of-the-art virus identification tools in thirteen modes on eight paired viral and microbial datasets from three distinct biomes, including a new complex dataset from Antarctic coastal waters. The tools had highly variable true positive rates (0 – 68%) and false positive rates (0 – 15%). PPR-Meta best distinguished viral from microbial contigs, followed by DeepVirFinder, VirSorter2, and VIBRANT. Different tools identified different subsets of the benchmarking data and all tools, except for Sourmash, found unique viral contigs. Tools performance could be improved with adjusted parameter cutoffs, indicating that adjustment of parameter cutoffs before usage should be considered. Together, our independent benchmarking provides guidance on choices of bioinformatic virus identification tools and gives suggestions for parameter adjustments for viromics researchers.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}