CoChat
Sign Up
Cited by 177Protein Structure and DynamicsMachine Learning in BioinformaticsGenomics and Rare Diseases

The DOI matches its Crossref record; 65 of the 66 references it cites were found in a registry or an index.

Rapid protein stability prediction using deep learning representations

Open access
Blaabjerg, Lasse M, Kassem, Maher M, Good, Lydia L, Jonsson, Nicolas, Cagiada, Matteo, Johansson, Kristoffer E, Boomsma, Wouter, Stein, Amelie, Lindorff-Larsen, Kresten·2023·eLife
DOI 10.7554/elife.82593PMC PMC10266766

Published version of a preprint

Open this paper in CoChat — free account

Abstract

Predicting the thermodynamic stability of proteins is a common and widely used step in protein engineering, and when elucidating the molecular mechanisms behind evolution and disease. Here, we present RaSP, a method for making rapid and accurate predictions of changes in protein stability by leveraging deep learning representations. RaSP performs on-par with biophysics-based methods and enables saturation mutagenesis stability predictions in less than a second per residue. We use RaSP to calculate ∼ 230 million stability changes for nearly all single amino acid changes in the human proteome, and examine variants observed in the human population. We find that variants that are common in the population are substantially depleted for severe destabilization, and that there are substantial differences between benign and pathogenic variants, highlighting the role of protein stability in genetic diseases. RaSP is freely available—including via a Web interface—and enables large-scale analyses of stability in experimental and predicted protein structures.

Overview

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.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.7554/elife.82593

Cites

65 of 65 references verified0 citation errors · 0 not checkable

Statistics

Figures

Figure 1. Overview of model training. We trained a self-supervised three-dimensional convolutional neural network (CNN) to learn internal representations of protein structures by predicting wild-type amino acid labels from protein structures. The representation model is trained to predict amino acid type based on the local atomic environment parameterized using a 3D sphere around the wild-type residue. Using the representations from the convolutional neural network as input, a second downstream and supervised fully connected neural network (FCNN) was trained to predict Rosetta Δ⁢Δ⁢G values.
Figure 1. Overview of model training. We trained a self-supervised three-dimensional convolutional neural network (CNN) to learn internal representations of protein structures by predicting wild-type amino acid labels from protein structures. The representation model is trained to predict amino acid type based on the local atomic environment parameterized using a 3D sphere around the wild-type residue. Using the representations from the convolutional neural network as input, a second downstream and supervised fully connected neural network (FCNN) was trained to predict Rosetta Δ⁢Δ⁢G values.CC BY · PubMed Central
Figure 2. Overview of RaSP downstream model training and testing. (A) Learning curve for training of the RaSP downstream model, with Pearson correlation coefficients (ρ) and mean absolute error (MAEF) of RaSP predictions. During training we transformed the target Δ⁢Δ⁢G data using a switching (Fermi) function, and MAEF refers to this transformed data (see Methods for further details). Error bars represent the standard deviation of 10 independently trained models, that were subsequently used in ensemble averaging. Val: validation set; Train: training set. (B) After training, we applied the RaSP model to an independent test set to predict Δ⁢Δ⁢G values for a full saturation mutagenesis of 10 proteins. Pearson correlation coefficients and mean absolute errors (MAE) were for this figure computed using only variants with Rosetta Δ⁢Δ⁢G values in the range [–1;7] kcal/mol.
Figure 2. Overview of RaSP downstream model training and testing. (A) Learning curve for training of the RaSP downstream model, with Pearson correlation coefficients (ρ) and mean absolute error (MAEF) of RaSP predictions. During training we transformed the target Δ⁢Δ⁢G data using a switching (Fermi) function, and MAEF refers to this transformed data (see Methods for further details). Error bars represent the standard deviation of 10 independently trained models, that were subsequently used in ensemble averaging. Val: validation set; Train: training set. (B) After training, we applied the RaSP model to an independent test set to predict Δ⁢Δ⁢G values for a full saturation mutagenesis of 10 proteins. Pearson correlation coefficients and mean absolute errors (MAE) were for this figure computed using only variants with Rosetta Δ⁢Δ⁢G values in the range [–1;7] kcal/mol.CC BY · PubMed Central
Figure 2—figure supplement 1. Learning curve for the self-supervised 3D convolutional neural network. The model obtained at epoch 15 achieves a classification accuracy of 63% on the validation set.
Figure 2—figure supplement 1. Learning curve for the self-supervised 3D convolutional neural network. The model obtained at epoch 15 achieves a classification accuracy of 63% on the validation set.CC BY · PubMed Central
Figure 2—figure supplement 2. Mean absolute prediction error for RaSP on the validation set, split by amino acid type of the wild-type and variant residue. Substitutions from glycine and cysteine as well as to proline generally have higher errors.
Figure 2—figure supplement 2. Mean absolute prediction error for RaSP on the validation set, split by amino acid type of the wild-type and variant residue. Substitutions from glycine and cysteine as well as to proline generally have higher errors.CC BY · PubMed Central
Figure 2—figure supplement 3. RaSP versus Rosetta Δ⁢Δ⁢G values for a full saturation mutagenesis of 10 test proteins separated into either exposed (A) or buried (B) residues. We speculate, that the RaSP prediction task is harder in the case of buried residues because Rosetta Δ⁢Δ⁢G values generally have higher variance in those regions. Pearson correlation coefficients and mean absolute errors (MAE) were for this figure computed using only variants with Rosetta Δ⁢Δ⁢G values in the range [–1;7] kcal/mol. Buried and exposed residue were classified based a relative surface accessible surface area (SASA) cut-off of 0.2.
Figure 2—figure supplement 3. RaSP versus Rosetta Δ⁢Δ⁢G values for a full saturation mutagenesis of 10 test proteins separated into either exposed (A) or buried (B) residues. We speculate, that the RaSP prediction task is harder in the case of buried residues because Rosetta Δ⁢Δ⁢G values generally have higher variance in those regions. Pearson correlation coefficients and mean absolute errors (MAE) were for this figure computed using only variants with Rosetta Δ⁢Δ⁢G values in the range [–1;7] kcal/mol. Buried and exposed residue were classified based a relative surface accessible surface area (SASA) cut-off of 0.2.CC BY · PubMed Central
Figure 3. Comparing RaSP and Rosetta predictions to experimental stability measurements. Predictions of changes in stability obtained using (A) RaSP and (B) Rosetta are compared to experimental data on five test proteins; myoglobin (1BVC), lysozyme (1LZ1), chymotrypsin inhibitor (2CI2), RNAse H (2RN2) and Protein G (1PGA) (Kumar, 2006; Ó Conchúir et al., 2015; Nisthal et al., 2019). Metrics used are Pearson correlation coefficient (ρ), mean absolute error (MAE) and mean error (ME). In the experimental study of Protein G, 105 variants were assigned a Δ⁢Δ⁢G value of at least 4 kcal/mol due to low stability, presence of a folding intermediate, or lack expression (Nisthal et al., 2019).
Figure 3. Comparing RaSP and Rosetta predictions to experimental stability measurements. Predictions of changes in stability obtained using (A) RaSP and (B) Rosetta are compared to experimental data on five test proteins; myoglobin (1BVC), lysozyme (1LZ1), chymotrypsin inhibitor (2CI2), RNAse H (2RN2) and Protein G (1PGA) (Kumar, 2006; Ó Conchúir et al., 2015; Nisthal et al., 2019). Metrics used are Pearson correlation coefficient (ρ), mean absolute error (MAE) and mean error (ME). In the experimental study of Protein G, 105 variants were assigned a Δ⁢Δ⁢G value of at least 4 kcal/mol due to low stability, presence of a folding intermediate, or lack expression (Nisthal et al., 2019).CC BY · PubMed Central
Figure 3—figure supplement 1. Comparing RaSP and Rosetta predictions to experimental stability measurements. Stability predictions obtained using (A–E) RaSP and (F–J) Rosetta are compared to experimental data for the five test proteins; myoglobin (1BVC), lysozyme (1LZ1), chymotrypsin inhibitor (2CI2), RNAse H (2RN2) and Protein G (1PGA) (Kumar, 2006; Ó Ó Conchúir et al., 2015; Nisthal et al., 2019). In the experimental study of Protein G, 105 variants were assigned a Δ⁢Δ⁢G value of at least 4 kcal/mol due to low stability, presence of a folding intermediate, or lack expression (Nisthal et al., 2019).
Figure 3—figure supplement 1. Comparing RaSP and Rosetta predictions to experimental stability measurements. Stability predictions obtained using (A–E) RaSP and (F–J) Rosetta are compared to experimental data for the five test proteins; myoglobin (1BVC), lysozyme (1LZ1), chymotrypsin inhibitor (2CI2), RNAse H (2RN2) and Protein G (1PGA) (Kumar, 2006; Ó Ó Conchúir et al., 2015; Nisthal et al., 2019). In the experimental study of Protein G, 105 variants were assigned a Δ⁢Δ⁢G value of at least 4 kcal/mol due to low stability, presence of a folding intermediate, or lack expression (Nisthal et al., 2019).CC BY · PubMed Central
Figure 3—figure supplement 2. RaSP performance on three recently published data sets (Pancotti et al., 2022): (A) The S669 data set, (B) The Ssym+ direct data set, (C) The Ssym+ reverse data set.
Figure 3—figure supplement 2. RaSP performance on three recently published data sets (Pancotti et al., 2022): (A) The S669 data set, (B) The Ssym+ direct data set, (C) The Ssym+ reverse data set.CC BY · PubMed Central
Figure 3—figure supplement 3. RaSP performance on the recently published mega-scale experiments (Tsuboyama et al., 2022). The experimental data has been filtered to include only well-defined experimental Δ⁢Δ⁢G values from single substitution mutations in natural protein domains (Tsuboyama et al., 2022). This filtered data set contains a total of 164,524 variants across 164 protein domain structures.
Figure 3—figure supplement 3. RaSP performance on the recently published mega-scale experiments (Tsuboyama et al., 2022). The experimental data has been filtered to include only well-defined experimental Δ⁢Δ⁢G values from single substitution mutations in natural protein domains (Tsuboyama et al., 2022). This filtered data set contains a total of 164,524 variants across 164 protein domain structures.CC BY · PubMed Central
Figure 3—figure supplement 4. Benchmarking RaSP and Rosetta using VAMP-seq data. We compare stability predictions with VAMP-seq scores for three test proteins (A) TPMT (PDB: 2H11) (Matreyek et al., 2018), (B) PTEN (PDB: 1D5R) (Matreyek et al., 2018) and (C) NUDT15 (PDB: 5BON) (Suiter et al., 2020).
Figure 3—figure supplement 4. Benchmarking RaSP and Rosetta using VAMP-seq data. We compare stability predictions with VAMP-seq scores for three test proteins (A) TPMT (PDB: 2H11) (Matreyek et al., 2018), (B) PTEN (PDB: 1D5R) (Matreyek et al., 2018) and (C) NUDT15 (PDB: 5BON) (Suiter et al., 2020).CC BY · PubMed Central
Figure 4. Stability predictions from structures created by template-based modelling. Pearson correlation coefficients (ρ) between experimental stability measurements and predictions using protein homology models with decreasing sequence identity to the target sequence. Pearson correlation coefficients were computed in the range of [–1;7] kcal/mol.
Figure 4. Stability predictions from structures created by template-based modelling. Pearson correlation coefficients (ρ) between experimental stability measurements and predictions using protein homology models with decreasing sequence identity to the target sequence. Pearson correlation coefficients were computed in the range of [–1;7] kcal/mol.CC BY · PubMed Central
Figure 5. Large-scale analysis of disease-causing variants and variants observed in the population. The grey distribution shown in the background of all plots represents the distribution of Δ⁢Δ⁢G values calculated using RaSP for all single amino acid changes in the 1,366 proteins that we analysed (15 of the 1381 proteins that we calculated Δ⁢Δ⁢G for did not have variants in ClinVar or gnomAD and were therefore not included in this analysis). Each plot is also labelled with the median Δ⁢Δ⁢G of the subset analysed as well as a range of Δ⁢Δ⁢G values that cover 95% of the data in that subset (box plot shows median, quartiles and outliers). The plots only show values between –1 and 7 kcal/mol (for the full range see Figure 5—figure supplement 2). (A) Distribution of RaSP Δ⁢Δ⁢G values for benign (blue) and pathogenic (tan) variants extracted from the ClinVar database (Landrum et al., 2018). We observe that the median RaSP Δ⁢Δ⁢G value is significantly higher for pathogenic variants compared to benign variants using bootstrapping. (B) Distribution of RaSP Δ⁢Δ⁢G values for variants with different allele frequencies (AF) extracted from the gnomAD database Karczewski et al., 2020 in the ranges (i) AF>10-2 (green), (ii) 10-2 > AF>10-4 (orange), and (iii) AF<10-4 (purple). We observe a gradual shift in the median RaSP Δ⁢Δ⁢G going from common variants (AF>10-2) towards rarer ones (AF<10-4).
Figure 5. Large-scale analysis of disease-causing variants and variants observed in the population. The grey distribution shown in the background of all plots represents the distribution of Δ⁢Δ⁢G values calculated using RaSP for all single amino acid changes in the 1,366 proteins that we analysed (15 of the 1381 proteins that we calculated Δ⁢Δ⁢G for did not have variants in ClinVar or gnomAD and were therefore not included in this analysis). Each plot is also labelled with the median Δ⁢Δ⁢G of the subset analysed as well as a range of Δ⁢Δ⁢G values that cover 95% of the data in that subset (box plot shows median, quartiles and outliers). The plots only show values between –1 and 7 kcal/mol (for the full range see Figure 5—figure supplement 2). (A) Distribution of RaSP Δ⁢Δ⁢G values for benign (blue) and pathogenic (tan) variants extracted from the ClinVar database (Landrum et al., 2018). We observe that the median RaSP Δ⁢Δ⁢G value is significantly higher for pathogenic variants compared to benign variants using bootstrapping. (B) Distribution of RaSP Δ⁢Δ⁢G values for variants with different allele frequencies (AF) extracted from the gnomAD database Karczewski et al., 2020 in the ranges (i) AF>10-2 (green), (ii) 10-2 > AF>10-4 (orange), and (iii) AF<10-4 (purple). We observe a gradual shift in the median RaSP Δ⁢Δ⁢G going from common variants (AF>10-2) towards rarer ones (AF<10-4).CC BY · PubMed Central
Figure 5—figure supplement 1. Histogram of Δ⁢Δ⁢G values from saturation mutagenesis using RaSP on 1,366 PDB structures corresponding to ∼8.8 million predicted Δ⁢Δ⁢G values.
Figure 5—figure supplement 1. Histogram of Δ⁢Δ⁢G values from saturation mutagenesis using RaSP on 1,366 PDB structures corresponding to ∼8.8 million predicted Δ⁢Δ⁢G values.CC BY · PubMed Central
Figure 5—figure supplement 2. Large-scale analysis of disease-causing variants and variants observed in the population using the RaSP model. The grey distribution shown in the background of all plots represents the distribution of Δ⁢Δ⁢G for all single amino acid changes in the 1366 proteins that we analysed. Each plot is also labelled with the median Δ⁢Δ⁢G of the subset analysed as well as a range of Δ⁢Δ⁢G values that cover 95% of the data in that subset (box plot shows median, quartiles and outliers). (A) Distribution of RaSP Δ⁢Δ⁢G values for benign (blue) and pathogenic (tan) variants extracted from the ClinVar database (Landrum et al., 2018). We observe that the median RaSP Δ⁢Δ⁢G value is higher for pathogenic variants compared to benign variants. (B) Distribution of RaSP Δ⁢Δ⁢G values for variants with different allele frequencies (AF) extracted from the gnomAD database Karczewski et al., 2020 in the ranges (i) AF > 10-2 (green), (ii) 10-2 > AF > 10-4 (orange), (iii) AF < 10-4 (purple). We observe a gradual shift in the median RaSP Δ⁢Δ⁢G going from common variants (AF> 10-2) towards rarer ones (AF< 10-4).
Figure 5—figure supplement 2. Large-scale analysis of disease-causing variants and variants observed in the population using the RaSP model. The grey distribution shown in the background of all plots represents the distribution of Δ⁢Δ⁢G for all single amino acid changes in the 1366 proteins that we analysed. Each plot is also labelled with the median Δ⁢Δ⁢G of the subset analysed as well as a range of Δ⁢Δ⁢G values that cover 95% of the data in that subset (box plot shows median, quartiles and outliers). (A) Distribution of RaSP Δ⁢Δ⁢G values for benign (blue) and pathogenic (tan) variants extracted from the ClinVar database (Landrum et al., 2018). We observe that the median RaSP Δ⁢Δ⁢G value is higher for pathogenic variants compared to benign variants. (B) Distribution of RaSP Δ⁢Δ⁢G values for variants with different allele frequencies (AF) extracted from the gnomAD database Karczewski et al., 2020 in the ranges (i) AF > 10-2 (green), (ii) 10-2 > AF > 10-4 (orange), (iii) AF < 10-4 (purple). We observe a gradual shift in the median RaSP Δ⁢Δ⁢G going from common variants (AF> 10-2) towards rarer ones (AF< 10-4).CC BY · PubMed Central
Figure 5—figure supplement 3. Histogram of Δ⁢Δ⁢G values from saturation mutagenesis using RaSP on predicted structures of the entire human proteome corresponding to ∼300 million predicted Δ⁢Δ⁢G values predicted from 23,391 protein structures.
Figure 5—figure supplement 3. Histogram of Δ⁢Δ⁢G values from saturation mutagenesis using RaSP on predicted structures of the entire human proteome corresponding to ∼300 million predicted Δ⁢Δ⁢G values predicted from 23,391 protein structures.CC BY · PubMed Central

Verification

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.

A few free checks a day · papers already verified open instantly · results are public

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

Show all 66 references

Research works printed but not linked: 1

  1. Linderstrøm-Lang Centre for Protein Science, University of Copenhagen. 2023. Rapid protein stability prediction using deep learning representations. Software Heritage. swh:1:rev:fb9438fa4724171e752fdf70e270db0cffc0f73c. https://archive.softwareheritage.org/swh:1:dir:3b5a5bf941013f41b9f7333e4190f4c9c4924cba;origin=https://github.com/KULL-Centre/_2022_ML-ddG-Blaabjerg;visit=swh:1:snp:2763a4dfde8edc3ee726a2aeb5897a90fc421c7c;anchor=swh:1:rev:fb9438fa4724171e752fdf70e270db0cffc0f73c.

Cited by

Sorted by citations

Related

See all

Metadata

Authors
Blaabjerg, L. M., Kassem, M. M., Good, L. L., Jonsson, N., Cagiada, M., Johansson, K. E., Boomsma, W., Stein, A., & Lindorff-Larsen, K.
Affiliations
  • University of Copenhagen (DK)
  • Linderstrøm-Lang Centre for Protein Science, Department of Biology, University of Copenhagen
Corresponding author
  • Boomsma, Wouter
  • Stein, Amelie
ORCID
Venue
eLife
Publisher
eLife Sciences Publications, Ltd
Published
2023-05-15
Volume
12
ISSN
2050-084X (ISSN-L)
DOI
10.7554/elife.82593
PMC
PMC10266766
OpenAlex
W4376612841
Cited by
177
References
66
Citations by year
2026: 48 · 2025: 59 · 2024: 48 · 2023: 16
FWCI
15.19
Type
journal article
Language
en
Open access
Licence
Locations
  • eLife (publishedVersion, cc-by)
  • bioRxiv (Cold Spring Harbor Laboratory) (acceptedVersion, cc-by-nc-nd)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • Research at the University of Copenhagen (University of Copenhagen) (publishedVersion, cc-by)
  • Apollo (University of Cambridge) (submittedVersion, cc-by)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
Funders
  • Novo Nordisk Fonden (NNF18OC0033950, NNF20OC0062606, NNF18OC0052719)
  • Lundbeckfonden (R272-2017-4528)
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
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Genetics › Genomics and Rare Diseases
Keywords
protein stability prediction, thermodynamic stability, RASP, saturation mutagenesis, single amino acid variants, pathogenic variants, genetic disorders, human proteome, protein engineering, protein evolution
MeSH terms
Deep Learning, Amino Acids, Humans, Proteins, Mutagenesis, Computational Biology, Protein Stability

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