Structural pre-training improves physical accuracy of antibody structure prediction using deep learning.
Open accessPublished version of a preprint
Abstract
Abstract Protein folding problem obtained a practical solution recently, owing to advances in deep learning. There are classes of proteins though, such as antibodies, that are structurally unique, where the general solution still lacks. In particular, the prediction of the CDR-H3 loop, which is an instrumental part of an antibody in its antigen recognition abilities, remains a challenge. Antibody-specific deep learning frameworks were proposed to tackle this problem noting great progress, both on accuracy and speed fronts. Oftentimes though, the original networks produce physically implausible bond geometries that then need to undergo a time-consuming energy minimization process. Here we hypothesized that pre-training the network on a large, augmented set of models with correct physical geometries, rather than a small set of real antibody X-ray structures, would allow the network to learn better bond geometries. We show that fine-tuning such a pre-trained network on a task of shape prediction on real X-ray structures improves the number of correct peptide bond distances. We further demonstrate that pre-training allows the network to produce physically plausible shapes on an artificial set of CDR-H3s, showing the ability to generalize to the vast antibody sequence space. We hope that our strategy will benefit the development of deep learning antibody models that rapidly generate physically plausible geometries, without the burden of time-consuming energy minimization.
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.
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
Highly accurate protein structure prediction with AlphaFold
Jumper, J., Evans, R., Pritzel, A., Green, T., Figurnov, M., Ronneberger, O., Tunyasuvunakool, K., Bates, R., Žídek, A., Potapenko, A., et al. (2021) · Nature · doi:10.1038/s41586-021-03819-2 · Cited by 43,764 · see citing works
Evolutionary-scale prediction of atomic level protein structure with a language model
Lin, Z., Akin, H., Rao, R., Hie, B., Zhu, Z., Lu, W., Smetanin, N., Verkuil, R., Kabeli, O., Shmueli, Y., et al. (2022) · doi:10.1101/2022.07.20.500902 · Cited by 5,523 · see citing works
Commonality despite exceptional diversity in the baseline human antibody repertoire
Briney, B., Inderbitzin, A., Joyce, C., & Burton, D. R. (2019) · Nature · doi:10.1038/s41586-019-0879-y · Cited by 544 · see citing works
Ahdritz, G., Bouatta, N., Floristean, C., Kadyan, S., Xia, Q., Gerecke, W., O’Donnell, T. J., Berenberg, D., Fisk, I., Zanichelli, N., et al. (2024) · Nature Methods · doi:10.1038/s41592-024-02272-z · Cited by 439 · see citing works
Kovaltsuk, A., Leem, J., Kelm, S., Snowden, J., Deane, C. M., & Krawczyk, K. (2018) · The Journal of Immunology · doi:10.4049/jimmunol.1800708 · Cited by 247 · see citing works
Self-citationAntibody structure prediction using interpretable deep learning
Ruffolo, J. A., Sulam, J., & Gray, J. J. (2022) · Patterns · doi:10.1016/j.patter.2021.100406 · Cited by 216 · see citing works
Second antibody modeling assessment (AMA-II)
Almagro, J. C., Teplyakov, A., Luo, J., Sweet, R. W., Kodangattil, S., Hernandez‐Guzman, F., & Gilliland, G. L. (2014) · Proteins Structure Function and Bioinformatics · doi:10.1002/prot.24567 · Cited by 144 · see citing works
Ig-VAE: Generative modeling of protein structure by direct 3D coordinate generation
Eguchi, R. R., Choe, C. A., & Huang, P.-S. (2022) · PLOS Computational Biology · doi:10.1371/journal.pcbi.1010271 · Cited by 141 · see citing works
The H3 loop of antibodies shows unique structural characteristics
Regep, C., Georges, G., Shi, J., Popovic, B., & Deane, C. M. (2017) · Proteins Structure Function and Bioinformatics · doi:10.1002/prot.25291 · Cited by 99 · see citing works
Structurally mapping antibody repertoires
Krawczyk, K., Kelm, S., Kovaltsuk, A., Galson, J. D., Kelly, D., Trück, J., Regep, C., Leem, J., Wong, W. K., Nowak, J., et al. (2018) · Frontiers in Immunology · doi:10.3389/fimmu.2018.01698 · Cited by 42 · see citing works
Antibody side chain conformations are position-dependent
Leem, J., Georges, G., Shi, J., & Deane, C. M. (2018) · Proteins Structure Function and Bioinformatics · doi:10.1002/prot.25453 · Cited by 22 · see citing works
Flagged first, then by citations
- Authors
- Kończak, J., Janusz, B., Młokosiewicz, J., Satława, T., Wróbel, S., Dudzic, P., & Krawczyk, K.
- Corresponding author
- Krawczyk, Konrad
- Venue
- ImmunoInformatics
- Publisher
- Elsevier BV
- Published
- 2023-09
- Volume
- 11
- Pages
- 100028
- ISSN
- 2667-1190 (ISSN-L)
- OpenAlex
- W4381332577
- Cited by
- 5
- References
- 11
- Citations by year
- 2025: 3 · 2024: 2
- FWCI
- 0.44
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Licence
- Locations
- ImmunoInformatics (publishedVersion, cc-by-nc-nd)
- bioRxiv (Cold Spring Harbor Laboratory) (acceptedVersion, cc-by)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
- Topics
- Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Protein Structure and Dynamics
- Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Machine Learning in Bioinformatics
- Physical Sciences › Computer Science › Computational Theory and Mathematics › Computational Drug Discovery Methods
- Keywords
- antibody structure prediction, fine-tuning, energy minimization, deep learning
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
