Deep learning shows promise for predicting gene expression levels from DNA sequences. However, recent studies show that current state-of-the-art models struggle to accurately characterize expression variation from personal genomes, limiting their usefulness in personalized medicine.
References
Leslie, R. et al. Bioinformatics 30, i185–i194 (2014).
Avsec, Ž. et al. Nat. Methods. 18, 1196–1203 (2021).
Toneyan, S. et al. Nat. Mach. Intell. 4, 1088–1100 (2022).
Karollus, A. et al. Genome Biol. 24, 56 (2023).
Sasse, A. et al. Nat. Genet. https://doi.org/10.1038/s41588-023-01524-6 (2023).
Huang, C. et al. Nat. Genet. https://doi.org/10.1038/s41588-023-01574-w (2023).
Bennett, D. A. et al. J. Alzheimers Dis. 64, S161–S189 (2018).
Kelley, D. R. et al. Genome Res. 28, 739–750 (2018).
Zhou, J. et al. Nat. Genet. 50, 1171–1179 (2018).
Agarwal, V. & Shendure, J. Cell Rep. 31, 107663 (2020).
Lappalainen, T. et al. Nature 501, 506–511 (2013).
Amazon, E. R. et al. Nat. Genet. 47, 1091–1098 (2015).
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Competing interest
The authors declare no competing interests.
Rights and permissions
About this article
Cite this article
Tang, Z., Toneyan, S. & Koo, P.K. Current approaches to genomic deep learning struggle to fully capture human genetic variation. Nat Genet 55, 2021–2022 (2023). https://doi.org/10.1038/s41588-023-01517-5
Published:
Issue Date:
DOI: https://doi.org/10.1038/s41588-023-01517-5
- Springer Nature America, Inc.