Skip to main content
APS Plant Health 2025 Main banner
Icon Legend
This session is not in your schedule.
This session is in your schedule. Click again to remove it.

POSTERS: Disease Detection and Diagnosis I

P-187 - Improving Metagenomics-based Plant Virus Identification by Improving Reference Databases

Monday, August 4, 2025
3:45 PM - 4:30 PM HST
Location: Exhibit Hall 3

    Presenting Author(s)

  • MG

    Mitchell Gercken

    School of Plant and Environmental Sciences and Graduate Program in Genetics, Bioinformatics, and Computational Biology, Virginia Tech, Blacksburg, VA 24061, U.S.A
    Blacksburg, VA, USA

    • Author(s)

  • LR

    Lina Rodriguez Salamanca

    School of Plant and Environmental Sciences, Virginia Tech, Blacksburg, VA 24061, U.S.A.
    Blacksburg, Virginia, United States

  • SK

    Sehgeet Kaur

    School of Plant and Environmental Sciences and Graduate Program in Genetics, Bioinformatics, and Computational Biology, Virginia Tech, Blacksburg, VA 24061, U.S.A
    Blacksburg, Virginia, United States

  • RM

    Reza Mazloom

    Graduate Researcher
    Department of Computer Science, Virginia Tech, Blacksburg, VA 24061, U.S.A.
    Blacksburg, Virginia, United States

  • KB

    Kassaye Belay (he/him/his)

    School of Plant and Environmental Sciences and Graduate Program in Genetics, Bioinformatics, and Computational Biology, Virginia Tech, Blacksburg, VA 24061, U.S.A
    Blacksburg, VA, USA

  • LH

    Lenwood S. Heath

    Department of Computer Science, Virginia Tech, Blacksburg, VA 24061, U.S.A.
    Blacksburg, Virginia, United States

  • BV

    Boris A. Vinatzer

    Professor
    School of Plant and Environmental Sciences, Virginia Tech, Blacksburg, VA 24061
    Blacksburg, Virginia, United States

Abstract Text: High-throughput next-generation sequencing makes it possible to identify viruses directly from RNA of infected plants. Yet, genome reference databases remain limited, often suffering from unequal representation of viral diversity, causing new viruses to be misclassified or remain unclassified when using taxonomic classifiers, such as Kraken 2. The Life Identification Number (LIN) system provides an approach to organize bacterial genomes using hierarchical Average Nucleotide Identity (ANI) thresholds. Here, we used LINs to choose representative genomes from eleven plant viral families using the NCBI GenBank and RefSeq databases. Four Kraken 2 databases were built, selecting one representative genome for genome similarity groups at 95% (~3,600 genomes), 99% (~7,700), 99.975% (~12,600), and 99.999% (~12,800) ANI thresholds. Benchmarking against the default Kraken 2 database (~2,500 RefSeq genomes) using plant viral metagenomes showed up to a two-fold increase in identified reads with our custom databases, yielding more complete genome assemblies. For a viral metagenome containing Tobacco Rattle Virus (TRV), a custom database assembly yielded two major contigs (6,815nt and 2,470nt), similar to the TRV reference genome consisting of RNA 1 (6,781nt) and RNA 2 (3,855nt), while the default database produced a more fragmented and incomplete genome consisting of three contigs (3,297nt, 3,210nt, and 2,161nt). In conclusion, our plant viral databases, with increased representation of genomic diversity, provide a new resource for improved metagenomics-based plant virus identification.