Personal transcriptome variation is poorly explained by current genomic deep learning models
Open accessPublished version of a preprint
Abstract
Genomic deep learning models can predict genome-wide epigenetic features and gene expression levels directly from DNA sequence. While current models perform well at predicting gene expression levels across genes in different cell types from the reference genome, their ability to explain expression variation between individuals due to cis -regulatory genetic variants remains largely unexplored. Here, we evaluate four state-of-the-art models on paired personal genome and transcriptome data and find limited performance when explaining variation in expression across individuals. In addition, models often fail to predict the correct direction of effect of cis -regulatory genetic variation on expression.
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
AlphaGenome has not yet been benchmarked on personal genome prediction, a known weakness of models in this space.
“We have not yet benchmarked the model on personal genome prediction, which is a known weakness of models in this space40,41.”
From Advancing regulatory variant effect prediction with AlphaGenome (2026) · cites this paper as [41]
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Cites
Base-resolution models of transcription-factor binding reveal soft motif syntax
Avsec, Ž., Weilert, M., Shrikumar, A., Krueger, S., Alexandari, A., Dalal, K., Fropf, R., McAnany, C., Gagneur, J., Kundaje, A., et al. (2021) · Nature Genetics · doi:10.1038/s41588-021-00782-6 · Cited by 610 · cited 1× · see citing works
“These genomic deep learning models are trained using genome-wide data from a variety of cell types and cellular contexts and have been shown to learn biologically relevant regulatory motifs within the input DNA sequence [8, 9].”
Sequence-based modeling of three-dimensional genome architecture from kilobase to chromosome scale
Zhou, J. (2022) · Nature Genetics · doi:10.1038/s41588-022-01065-4 · Cited by 177 · see citing works
Flagged first, then by citations
Research works printed but not linked: 18
- [1]Kelley DR Sequential regulatory activity prediction across chromosomes with convolutional neural networks Genome Res. 2018 28 739 750 10.1101/gr.227819.117 29588361 PMC5932613 · DOI · Paper page
- [2]Zhou J Deep learning sequence-based ab initio prediction of variant effects on expression and disease risk Nat. Genet. 2018 50 1171 1179 10.1038/s41588-018-0160-6 30013180 PMC6094955 · DOI · Paper page
- [3]Agarwal V Shendure J Predicting mRNA abundance directly from genomic sequence using deep convolutional neural networks Cell Rep. 2020 31 107663 10.1016/j.celrep.2020.107663 32433972 · DOI · Paper page
- [4]Avsec Ž Effective gene expression prediction from sequence by integrating long-range interactions Nat. Methods 2021 18 1196 1203 10.1038/s41592-021-01252-x 34608324 PMC8490152 · DOI · Paper page
- [5]Fudenberg, G., Kelley, D. R. & Pollard, K. S. Predicting 3D genome folding from DNA sequence with Akita. Nat. Methods17, 1111–1117 (2020). PMC8211359 33046897
- [7]Zhou, J. & Troyanskaya, O. G. Predicting effects of noncoding variants with deep learning-based sequence model. Nat. Methods12, 931–934 (2015). PMC4768299 26301843
- [8]Kelley DR Snoek J Rinn JL Basset: learning the regulatory code of the accessible genome with deep convolutional neural networks Genome Res. 2016 26 990 999 10.1101/gr.200535.115 27197224 PMC4937568 · DOI · Paper page
- [10]Chen KM Wong AK Troyanskaya OG Zhou J A sequence-based global map of regulatory activity for deciphering human genetics Nat. Genet. 2022 54 940 949 10.1038/s41588-022-01102-2 35817977 PMC9279145 · DOI · Paper page
- [11]Kelley DR Cross-species regulatory sequence activity prediction PLoS Comput. Biol. 2020 16 1008050 10.1371/journal.pcbi.1008050 PMC7392335 32687525 · DOI · Paper page
- [12]Lappalainen T Transcriptome and genome sequencing uncovers functional variation in humans Nature 2013 501 506 511 10.1038/nature12531 24037378 PMC3918453 · DOI · Paper page
- [13]Sasse, A. et al. Benchmarking of deep neural networks for predicting personal gene expression from DNA sequence highlights shortcomings. Nat. Genet.10.1038/s41588-023-01524-6 (2023). 38036778 · DOI · Paper page
- [14]GTEx Consortium et al. A global reference for human genetic variation. Nature526, 68–74 (2015). PMC4750478 26432245
- [15]Gamazon ER A gene-based association method for mapping traits using reference transcriptome data Nat. Genet. 2015 47 1091 1098 10.1038/ng.3367 26258848 PMC4552594 · DOI
- [16]Karollus, A., Mauermeier, T. & Gagneur, J. Current sequence-based models capture gene expression determinants in promoters but mostly ignore distal enhancers. Genome Biol.24, 56 (2023). PMC10045630 36973806
- [17]Danecek, P. et al. Twelve years of SAMtools and BCFtools. GigaScience10, giab008 (2021). PMC7931819 33590861
- [19]Wang G Sarkar A Carbonetto P Stephens M A simple new approach to variable selection in regression, with application to genetic fine mapping J. R. Stat. Soc. Ser. B Stat. Methodol. 2020 82 1273 1300 10.1111/rssb.12388 PMC10201948 37220626 · DOI · Paper page
- [20]Wang QS Leveraging supervised learning for functionally informed fine-mapping of cis-eqtls identifies an additional 20,913 putative causal eqtls Nat. Commun. 2021 12 3394 10.1038/s41467-021-23134-8 34099641 PMC8184741 · DOI
- [21]Tehranchi A Fine-mapping cis-regulatory variants in diverse human populations eLife 2019 8 39595 10.7554/eLife.39595 PMC6335058 30650056 · DOI
Cited by
Sorted by citations
AlphaGenome: advancing regulatory variant effect prediction with a unified DNA sequence model
Avsec, Ž., Latysheva, N., Cheng, J., Novati, G., Taylor, K. R., Ward, T., Bycroft, C., Nicolaisen, L., Arvaniti, E., Pan, J., et al. (2025) · doi:10.1101/2025.06.25.661532 · Cited by 395 · see citing works
Advancing regulatory variant effect prediction with AlphaGenome
Avsec, Ž., Latysheva, N., Cheng, J., Novati, G., Taylor, K. R., Ward, T., Bycroft, C., Nicolaisen, L., Arvaniti, E., Pan, J., et al. (2026) · Nature · doi:10.1038/s41586-025-10014-0 · Cited by 373 · see citing works
“We have not yet benchmarked the model on personal genome prediction, which is a known weakness of models in this space 40,41 .”
Predicting RNA-seq coverage from DNA sequence as a unifying model of gene regulation
Linder, J., Srivastava, D., Yuan, H., Agarwal, V., & Kelley, D. R. (2025) · Nature Genetics · doi:10.1038/s41588-024-02053-6 · Cited by 265 · see citing works
TFXcan reveals transcription factor programs driving GWAS loci
Adeluwa, T., Salazar-Magaña, S., Sumner, S., Gona, S., Nyasimi, F., Carbonetto, P., Baca, S. C., Freedman, M., Gusev, A., Liu, B., et al. (2026) · Cell Genomics · doi:10.1016/j.xgen.2026.101398
- Authors
- Huang, C., Shuai, R. W., Baokar, P., Chung, R., Rastogi, R., Kathail, P., & Ioannidis, N. M.
- Affiliations
- University of California, Berkeley (US)
- Chan Zuckerberg Biohub San Francisco (US)
- Corresponding author
- Ioannidis, Nilah M.
- Venue
- Nature Genetics
- Published
- 2023-11-30
- Volume
- 55
- Issue
- 12
- Pages
- 2056-2059
- ISSN
- 1061-4036 (ISSN-L)
- 1546-1718
- PMC
- PMC10703684
- OpenAlex
- W4389174423
- Cited by
- 127
- References
- 2
- Citations by year
- 2026: 50 · 2025: 39 · 2024: 29 · 2023: 9
- FWCI
- 21.54
- Type
- journal article
- Language
- en
- Open access
- hybrid
- doi.org
- Locations
- Nature Genetics (publishedVersion, cc-by)
- bioRxiv (Cold Spring Harbor Laboratory) (acceptedVersion, cc-by-nc-nd)
- PubMed Central (publishedVersion, cc-by)
- PubMed (publishedVersion)
- eScholarship (California Digital Library) (submittedVersion, cc-by)
- Funders
- U.S. Department of Health & Human Services | NIH | National Human Genome Research Institute (R00HG00967)
- Topics
- Life Sciences › Biochemistry, Genetics and Molecular Biology › Genetics › Genetic Associations and Epidemiology
- Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Genomics and Chromatin Dynamics
- Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Genomics and Phylogenetic Studies
- Keywords
- gene expression prediction, cis-regulatory variation, transcriptome, epigenetic features, model evaluation, single-cell transcriptomics
- MeSH terms
- Deep Learning, Humans, Genetic Variation, Genome, Genomics, Transcriptome
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
