Using AlphaFold to predict the impact of single mutations on protein stability and function

By M. Pak, Karina A. Markhieva, Mariia Novikova, D. Petrov, I. Vorobyev, E. S. Maksimova, Fyodor A. Kondrashov, Dmitry N. Ivankov. 2021

283 citations recorded by the source corpus.

AlphaFold changed the field of structural biology by achieving three-dimensional (3D) structure prediction from protein sequence at experimental quality. The astounding success even led to claims that the protein folding problem is “solved”. However, protein folding problem is more than just structure prediction from sequence. Presently, it is unknown if the AlphaFold-triggered revolution could help to solve other problems related to protein folding. Here we assay the ability of AlphaFold to predict the impact of single mutations on protein stability (ΔΔG) and function. To study the question we extracted metrics from AlphaFold predictions before and after single mutation in a protein and correlated the predicted change with the experimentally known ΔΔG values. Additionally, we correlated the AlphaFold predictions on the impact of a single mutation on structure with a large scale dataset of single mutations in GFP with the experimentally assayed levels of fluorescence. We found a very weak or no correlation between AlphaFold output metrics and change of protein stability or fluorescence. Our results imply that AlphaFold cannot be immediately applied to other problems or applications in protein folding.

Subfields: Materials Chemistry, Molecular Biology

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Materials Chemistry

1. David Baker
2. Sergey Ovchinnikov
3. Jens Meiler
4. Yang Zhang
5. Frank DiMaio
6. Ivan V. Anishchenko
7. J. Dauparas
8. Arne Elofsson

In Molecular Biology

1. David Baker
2. Sergey Ovchinnikov
3. Yang Zhang
4. Jens Meiler
5. Björn Wallner
6. Arne Elofsson
7. Frank DiMaio
8. Jeffrey J. Gray

Citations in the corpus

• Predicting multiple conformations via sequence clustering and AlphaFold2 (2023, positive): “Beyond predicting the predominant state of naturally occurring proteins, we wanted to test the ability of AF-Cluster to predict effects of point mutations, a task that AF2 in its default settings has not achieved 40 .”
• AlphaFold2 and its applications in the fields of biology and medicine (2023, background): “Although there are researches showing that AF2 could predict the phenotypic effect of missense mutations,(209,210) it was also observed that the performance of missense mutations prediction is not good in other studies,(211) and there were only weak or no correlations between the output metrics of A”
• Sampling alternative conformational states of transporters and receptors with AlphaFold2 (2022, background): “Several preprints have provided evidence that AF2, despite its accuracy, likely does not learn the energy landscapes underpinning protein folding and function ( Saldaño et al., 2021 ; Pak et al., 2021 ; Akdel et al., 2021 ).”
• Scaffolding protein functional sites using deep learning (2022, background): “For almost all problems, we obtained designs that are closely recapitulated by AF with overall and motif (functional site) RMSD typically <2 Å and <1 Å respectively, with high model confidence (pLDDT > 80; Table S2); such recapitulation suggests the designed sequences encode the designed structures ”
• The Statistical Trends of Protein Evolution: A Lesson from AlphaFold Database (2022, unknown): “Despite the limitations noted in previous research (Bagdonas et al. 2021; Pak et al. 2021; Ruff and Pappu 2021), AlphaFold 2 is acknowledged to provide high-accuracy predictions of protein structures, even for sequences with relatively fewer homologous sequences ( Jumper et al. 2021).”
• Masked Inverse Folding with Sequence Transfer for Protein Representation Learning (2023, background): “However, this would be computationally expensive for large datasets, and it is currently unclear how good AlphaFold is at predicting the effects of single mutations (Pak et al., 2021).”
• Harnessing protein folding neural networks for peptide–protein docking (2021, background): “Thus, while it has been discussed that AF2 is not to be used for the modeling of structural effects of point mutations in the query sequence 42 , models can still be the basis for alanine scanning and other structure-based characterizations.”
• The impact of AlphaFold2 one year on (2022, unknown): “These two concerns will create a huge dilemma.”
• AlphaFold2 can predict structural and phenotypic effects of single mutations (2022, background): “12 We see no clear a priori reason that pLDDT should correlate with either stability or fluorescence.”
• A structural biology community assessment of AlphaFold2 applications (2021, background): “However, some reports have indicated that AF2 does not appear to be well suited to predict the structures of mutated proteins 54,55 .”