Overcoming limitations to customize DeepVariant for domesticated animals with TrioTrain
Genome Research · 2025-08 · 35 (8) · 1859-1874
DOI: 10.1101/gr.279542.124 ↗Kurz gefasst
Fachartikel in der Zeitschrift Genome Research, online erschienen am 02.06.2025. Verfasst von Jenna Kalleberg und 2 weiteren Autor:innen. Beteiligte Einrichtung: Division of Animal Sciences, University of Missouri, Columbia, Missouri 65201, USA. Hauptthemen laut MEDLINE: Software, Animals, Domestic, Genomics. Die Arbeit wurde bisher 4-mal zitiert (Crossref) und nennt 83 Literaturangaben. Der Volltext ist frei in PubMed Central verfügbar. Gefördert von: NLM NIH HHS, National Institutes of Health, National Library of Medicine, U.S. Department of Agriculture, National Institute of Food and Agriculture.
Abstract (Originalsprache)
Generating high-quality variant callsets across diverse species remains challenging as most bioinformatic tools default to assumptions based on human genomes. DeepVariant (DV) excels without joint genotyping while offering fewer implementation barriers. However, the growing appeal of a “universal” algorithm has magnified the unknown impacts when used with non-human species. Here, we use bovine genomes to assess the limits of using human genome–trained variant callers, including the allele frequency channel (DV-AF) and joint-caller DeepTrio (DT). Our novel approach, TrioTrain, automates extending DV for diploid species lacking Genome-in-a-Bottle (GIAB) resources, using a region shuffling approach to mitigate barriers for SLURM-based clusters. Imperfect animal truth labels are curated to remove Mendelian discordant sites before training DV to genotype the offspring correctly. With TrioTrain, we use cattle, yak, and bison trios to create the first multispecies-trained DV-AF checkpoint. Although incomplete bovine truth sets constrain recall within challenging repetitive regions, we observe a mean SNV F1 score >0.990 across new checkpoints during GIAB benchmarking. With HG002, a bovine-trained checkpoint (28) decreased the Mendelian inheritance error (MIE) rate by a factor of two compared with the default (DV). Checkpoint 28 has a mean MIE rate of 0.03% in three bovine interspecies cross genomes. These results illustrate that a multispecies, trio-based training strategy reduces inheritance errors during single-sample variant calling. Although exclusively training with human genomes deters transferring deep-learning-based variant calling to new species, we use the diverse ancestry within bovids to illustrate the need for advanced tools designed for comparative genomics.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Autor:innen und Institutionen
- Division of Animal Sciences, University of Missouri, Columbia, Missouri 65201, USA.
- Division of Animal Sciences, University of Missouri, Columbia, Missouri 65201, USA; schnabelr@missouri.edu.
- Genetics Area Program, University of Missouri, Columbia, Missouri 65201, USA.
ORCID-Profile: Jenna Kalleberg, Robert D Schnabel
Förderung
- NLM NIH HHS
- National Institutes of Health, National Library of Medicine
- U.S. Department of Agriculture, National Institute of Food and Agriculture
- HATCH MO
Zitieren
APA
Kalleberg, J., Rissman, J., & Schnabel, R. D. (2025). Overcoming limitations to customize DeepVariant for domesticated animals with TrioTrain. Genome Research, 35(8), 1859-1874. https://doi.org/10.1101/gr.279542.124
Vancouver
Kalleberg J, Rissman J, Schnabel RD. Overcoming limitations to customize DeepVariant for domesticated animals with TrioTrain. Genome Research. 2025;35(8):1859-1874. doi:10.1101/gr.279542.124
BibTeX
@article{Kalleberg2025,
author = {Kalleberg, Jenna and Rissman, Jacob and Schnabel, Robert D},
title = {Overcoming limitations to customize DeepVariant for domesticated animals with TrioTrain},
journal = {Genome Research},
year = {2025},
volume = {35},
number = {8},
pages = {1859--1874},
doi = {10.1101/gr.279542.124}
}
EUVIMED