EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Rethinking DeepVariant: Efficient Neural Architectures for Intelligent Variant Calling

Anastasiia Gurianova, Anastasiia Pestruilova, Aleksandra Beliaeva iD, Artem Kasianov iD, Liudmila Mikhailova iD, Egor Guguchkin iD, Evgeny Karpulevich iD

International Journal of Molecular Sciences · 2026

DOI: 10.3390/ijms27010513 ↗

Kurz gefasst

Fachartikel in der Zeitschrift International Journal of Molecular Sciences, 2026. Verfasst von Anastasiia Gurianova und 6 weiteren Autor:innen. Beteiligte Einrichtung: Ivannikov Institute for System Programming of the Russian Academy of Science, 109004 Moscow, Russia. Hauptthemen laut MEDLINE: Neural Networks, Computer, Computational Biology, Genetic Variation, Software. Die Arbeit wurde bisher 3-mal zitiert (Crossref) und nennt 38 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Ministry of Economic Development of the Russian Federation, Ivannikov Institute for System Programming of the Russian Academy of Sciences.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

DeepVariant has revolutionized the field of genetic variant identification by reframing variant detection as an image classification problem. However, despite its wide adoption in bioinformatics workflows, the tool continues to evolve mainly through the expansion of training datasets, while its core neural network architecture—Inception V3—has remained unchanged. In this study, we revisited the DeepVariant design and presented a prototype of a modernized version that supports alternative neural network backbones. As a proof of concept, we replaced the legacy Inception V3 model with a mid-sized EfficientNet model and evaluated its performance using the benchmark dataset from the Genome in a Bottle (GIAB) project. Alternative architecture demonstrated faster convergence, a twofold reduction in the number of parameters, and improved accuracy in variant identification. On the test dataset, updated workflow achieved consistent improvements of +0.1% in SNP F1-score, enabling the detection of up to several hundred additional true variants per genome. These results show that optimizing the neural architecture alone can enhance the accuracy, robustness, and efficiency of variant calling, thereby improving the overall quality of sequencing data analysis.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Neural Networks, ComputerComputational BiologyGenetic VariationSoftwarePolymorphism, Single NucleotideHumansConvolutional Neural NetworksGenomicsAlgorithmsHigh-Throughput Nucleotide Sequencing

Stichwörter der Autor:innen

DeepVariantGIABNGSSNPWESWGSconvolutional neural networksgenomicsvariant calling

Autor:innen und Institutionen

  • Ivannikov Institute for System Programming of the Russian Academy of Science, 109004 Moscow, Russia.
  • Center for Applied AI, Skolkovo Institute of Science and Technology, 121205 Moscow, Russia.
  • BIOPOLIS Program in Genomics, Biodiversity and Land Planning, CIBIO, 4485-684 Vairão, Portugal.
  • Economic Faculty, Lomonosov Moscow State University, 119991 Moscow, Russia.
  • Higher School of Management, Financial University Under the Government of the Russian Federation, 125167 Moscow, Russia.
  • Research Center for Trusted Artificial Intelligence, Ivannikov Institute for System Programming of the Russian Academy of Science, 109004 Moscow, Russia.

ORCID-Profile: Aleksandra Beliaeva, Artem Kasianov, Liudmila Mikhailova, Egor Guguchkin, Evgeny Karpulevich

Förderung

  • Ministry of Economic Development of the Russian Federation
  • Ivannikov Institute for System Programming of the Russian Academy of Sciences

Zitieren

APA

Gurianova, A., Pestruilova, A., Beliaeva, A., Kasianov, A., Mikhailova, L., Guguchkin, E., & Karpulevich, E. (2026). Rethinking DeepVariant: Efficient Neural Architectures for Intelligent Variant Calling. International Journal of Molecular Sciences. https://doi.org/10.3390/ijms27010513

Vancouver

Gurianova A, Pestruilova A, Beliaeva A, Kasianov A, Mikhailova L, Guguchkin E, et al. Rethinking DeepVariant: Efficient Neural Architectures for Intelligent Variant Calling. International Journal of Molecular Sciences. 2026. doi:10.3390/ijms27010513

BibTeX
@article{Gurianova2026,
  author = {Gurianova, Anastasiia and Pestruilova, Anastasiia and Beliaeva, Aleksandra and Kasianov, Artem and Mikhailova, Liudmila and Guguchkin, Egor and Karpulevich, Evgeny},
  title = {Rethinking DeepVariant: Efficient Neural Architectures for Intelligent Variant Calling},
  journal = {International Journal of Molecular Sciences},
  year = {2026},
  doi = {10.3390/ijms27010513}
}
Kostenlos starten →