Predicting changes in protein thermodynamic stability upon point mutation with deep 3D convolutional neural networks
Open accessThe DOI matches its Crossref record; 19 of the 19 references it cites were found in a registry or an index.
Abstract
Predicting mutation-induced changes in protein thermodynamic stability (ΔΔG) is of great interest in protein engineering, variant interpretation, and protein biophysics. We introduce ThermoNet, a deep, 3D-convolutional neural network (3D-CNN) designed for structure-based prediction of ΔΔGs upon point mutation. To leverage the image-processing power inherent in CNNs, we treat protein structures as if they were multi-channel 3D images. In particular, the inputs to ThermoNet are uniformly constructed as multi-channel voxel grids based on biophysical properties derived from raw atom coordinates. We train and evaluate ThermoNet with a curated data set that accounts for protein homology and is balanced with direct and reverse mutations; this provides a framework for addressing biases that have likely influenced many previous ΔΔG prediction methods. ThermoNet demonstrates performance comparable to the best available methods on the widely used S sym test set. In addition, ThermoNet accurately predicts the effects of both stabilizing and destabilizing mutations, while most other methods exhibit a strong bias towards predicting destabilization. We further show that homology between S sym and widely used training sets like S2648 and VariBench has likely led to overestimated performance in previous studies. Finally, we demonstrate the practical utility of ThermoNet in predicting the ΔΔGs for two clinically relevant proteins, p53 and myoglobin, and for pathogenic and benign missense variants from ClinVar. Overall, our results suggest that 3D-CNNs can model the complex, non-linear interactions perturbed by mutations, directly from biophysical properties of atoms.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Adam: A Method for Stochastic Optimization
Kingma, D. P., & Ba, J. (2014) · arXiv · doi:10.48550/arxiv.1412.6980 · Cited by 98,946
LeCun, Y., Bengio, Y., & Hinton, G. (2015) · Nature · doi:10.1038/nature14539 · Cited by 84,126
A global reference for human genetic variation
The 1000 Genomes Project Consortium, authors, C., Auton, A., Abecasis, G. R., Steering committee, Altshuler, D. M., Durbin, R. M., Abecasis, G. R., Bentley, D. R., Chakravarti, A., et al. (2015) · Nature · doi:10.1038/nature15393 · Cited by 18,247 · cited 1×
“(Note that 69 the term “wild type” is not preferred because humans have substantial protein-coding genetic 70 diversity (2).”
BackgroundSelf-citationThe Ensembl Variant Effect Predictor
McLaren, W., Gil, L., Hunt, S. E., Riat, H. S., Ritchie, G. R. S., Thormann, A., Flicek, P., & Cunningham, F. (2016) · Genome Biology · doi:10.1186/s13059-016-0974-4 · Cited by 8,099
ClinVar: improving access to variant interpretations and supporting evidence
Landrum, M. J., Lee, J. M., Benson, M., Brown, G. R., Chao, C., Chitipiralla, S., Gu, B., Hart, J., Hoffman, D., Jang, W., et al. (2018) · Nucleic Acids Research · doi:10.1093/nar/gkx1153 · Cited by 4,196
Leaver-Fay (2011) · doi:10.1016/b978-0-12-381270-4.00019-6 · Cited by 1,904 · cited 1×
“6 macromolecular modeling suite (Methods) (9, 33).”
Guerois, R., Nielsen, J. E., & Serrano, L. (2002) · Journal of Molecular Biology · doi:10.1016/s0022-2836(02)00442-4 · Cited by 1,728 · cited 4×
“Experimentally measured ∆∆Gs generally fall within -5 kcal/mol to +5 kcal/mol (13, 46); 381 thus, ThermoNet’s predictions are consistent with the range of observed values.”
BackgroundMethodInfluentialThe Rosetta All-Atom Energy Function for Macromolecular Modeling and Design
Alford, R. F., Leaver-Fay, A., Jeliazkov, J. R., O’Meara, M. J., DiMaio, F. P., Park, H., Shapovalov, M. V., Renfrew, P. D., Mulligan, V. K., Kappel, K., et al. (2017) · Journal of Chemical Theory and Computation · doi:10.1021/acs.jctc.7b00125 · Cited by 1,532
The coming of age of de novo protein design
Huang, P.-S., Boyken, S. E., & Baker, D. (2016) · Nature · doi:10.1038/nature19946 · Cited by 1,478
Deep mutational scanning: a new style of protein science
Fowler, D. M., & Fields, S. (2014) · Nature Methods · doi:10.1038/nmeth.3027 · Cited by 1,287 · cited 1×
“Nonetheless, the increasing adoption of deep mutational scanning techniques for systematic study of the molecular effects of mutations is generating an unprecedented amount of data [52].”
The IARC TP53 database: new online mutation analysis and recommendations to users
Olivier, M., Eeles, R., Hollstein, M., Khan, M. A., Harris, C. C., & Hainaut, P. (2002) · Human Mutation · doi:10.1002/humu.10081 · Cited by 1,248
Distance-based protein folding powered by deep learning
Xu, J. (2019) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.1821309116 · Cited by 401
Greener, J. G., Kandathil, S. M., & Jones, D. T. (2019) · Nature Communications · doi:10.1038/s41467-019-11994-0 · Cited by 178
Assessment of contact predictions in CASP12: Co‐evolution and deep learning coming of age
Schaarschmidt, J., Monastyrskyy, B., Kryshtafovych, A., & Bonvin, A. M. J. J. (2018) · Proteins: Structure, Function, and Bioinformatics · doi:10.1002/prot.25407 · Cited by 177
Quantification of biases in predictions of protein stability changes upon mutations
Pucci, F., Bernaerts, K. V., Kwasigroch, J. M., & Rooman, M. (2018) · Bioinformatics · doi:10.1093/bioinformatics/bty348 · Cited by 169 · cited 4×
“This training set was obtained by excluding S variants from the data set used to train STRUM (Quan et al., 2016).”
BackgroundMethodInfluential3D deep convolutional neural networks for amino acid environment similarity analysis
Torng, W., & Altman, R. B. (2017) · BMC Bioinformatics · doi:10.1186/s12859-017-1702-0 · Cited by 142
Assessing predictors of changes in protein stability upon mutation using self-consistency
Thiltgen, G., & Goldstein, R. A. (2012) · PLoS ONE · doi:10.1371/journal.pone.0046084 · Cited by 110 · cited 4×
“On the other hand, 92 several studies have highlighted significant bias in the predictions of machine learning-based 93 methods; they tend to predict mutations as destabilizing more often than stabilizing (19, 23-25).”
BackgroundInfluentialUsmanova, D. R., Bogatyreva, N. S., Ariño Bernad, J., Eremina, A. A., Gorshkova, A. A., Kanevskiy, G. M., Lonishin, L. R., Meister, A. V., Yakupova, A. G., Kondrashov, F. A., et al. (2018) · Bioinformatics · doi:10.1093/bioinformatics/bty340 · Cited by 93
On the biases in predictions of protein stability changes upon variations: the INPS test case
Montanucci, L., Savojardo, C., Martelli, P. L., Casadio, R., & Fariselli, P. (2019) · Bioinformatics · doi:10.1093/bioinformatics/bty979 · Cited by 33
Cited by
Learning inverse folding from millions of predicted structures
Hsu, C., Verkuil, R., Liu, J., Lin, Z., Hie, B., Sercu, T., Lerer, A., & Rives, A. (2022) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2022.04.10.487779 · Cited by 416
Transfer learning to leverage larger datasets for improved prediction of protein stability changes
Dieckhaus, H., Brocidiacono, M., Randolph, N. Z., & Kuhlman, B. (2024) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.2314853121 · Cited by 205
“Convolutional neural networks (CNNs) such as ThermoNet (11) and RaSP (12) require voxelization and computationally expensive convolutions, while methods using more efficient graph neural networks (GNNs) such as ProS - GNN (13) and BayeStab (14) require generation of a modeled mutant structure for inference, adding sub…”
Rapid protein stability prediction using deep learning representations
Blaabjerg, L. M., Kassem, M. M., Good, L. L., Jonsson, N., Cagiada, M., Johansson, K. E., Boomsma, W., Stein, A., & Lindorff-Larsen, K. (2023) · eLife · doi:10.7554/elife.82593 · Cited by 177
Metal3D: a general deep learning framework for accurate metal ion location prediction in proteins
Dürr, S. L., Levy, A., & Rothlisberger, U. (2023) · Nature Communications · doi:10.1038/s41467-023-37870-6 · Cited by 88
Wang, T., Xiang, G., He, S., Su, L., Wang, Y., Yan, X., & Lu, H. (2024) · Briefings in Bioinformatics · doi:10.1093/bib/bbae409 · Cited by 57
Metadata
- Authors
- Li, B., Yang, Y. T., Capra, J. A., & Gerstein, M. B.
- Venue
- PLOS Computational Biology
- Publisher
- Public Library of Science (PLoS)
- Published
- 2020-11-30
- Volume
- 16
- Issue
- 11
- Pages
- e1008291
- ISSN
- 1553-734X (ISSN-L)
- 1553-7358
- OpenAlex
- W3109312535
- Cited by
- 140
- References
- 19
- FWCI
- 4.15
- Type
- journal article
- Language
- en
- Open access
- gold
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
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.