Impact of protein conformational diversity on AlphaFold predictions

By Tadeo E. Saldaño, Nahuel Escobedo, Julia Marchetti, Diego J. Zea, Juan Mac Donagh, Ana Julia Velez Rueda, Eduardo Gonik, Agustina García Melani, Julieta Novomisky Nechcoff, Martin Salas and 6 more. 2021

188 citations recorded by the source corpus.

After the outstanding breakthrough of AlphaFold in predicting protein 3D models, new questions appeared and remain unanswered. The ensemble nature of proteins, for example, challenges the structural prediction methods because the models should represent a set of conformers instead of single structures. The evolutionary and structural features captured by effective deep learning techniques may unveil the information to generate several diverse conformations from a single sequence. Here we address the performance of AlphaFold2 predictions under this ensemble paradigm. Using a curated collection of apo-holo conformations, we found that AlphaFold2 predicts the holo form of a protein in 70% of the cases, being unable to reproduce the observed conformational diversity with an equivalent error than in the estimation of a single conformation. More importantly, we found that AlphaFold2’s performance worsens with the increasing conformational diversity of the studied protein. This impairment is related to the heterogeneity in the degree of conformational diversity found between different members of the homologous family of the protein under study. Finally, we found that main-chain flexibility associated with apo-holo pairs of conformers negatively correlates with the predicted local model quality score plDDT, indicating that plDDT values in a single 3D model could be used to infer local conformational changes linked to ligand binding transitions.

Subfields: Computational Theory and Mathematics, 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 Computational Theory and Mathematics

1. Stefano Forli
2. David S. Goodsell
3. Matthew Holcomb
4. Ya-Ting Chang
5. Richard A. Stein
6. Hassane S. Mchaourab

In Materials Chemistry

1. Stefano Forli
2. Richard A. Stein
3. David S. Goodsell
4. Hassane S. Mchaourab

In Molecular Biology

1. Stefano Forli
2. David S. Goodsell
3. Richard A. Stein
4. Hassane S. Mchaourab

Citations in the corpus

• Predicting multiple conformations via sequence clustering and AlphaFold2 (2023, critical): “…has garnered interest in its ability to predict multiple conformations of proteins, yet AF2 has been demonstrated to fail in predicting multiple structures of metamorphic proteins 11 , proteins with apo/holo conformational changes 12 and other multi-state proteins 13 using its default settings.”
• Modeling Alternate Conformations with Alphafold2 via Modification of the Multiple Sequence Alignment (2021, background): “A recent report examining a curated subset of apo and ligand bound structures 302 suggested that pLDDT values infer local conformational changes.(31) The results presented here do not 303 suggest that these regions of low confidence are incorrect in their two- or three-dimensional structure, 304 but”
• Evaluation of AlphaFold2 structures as docking targets (2022, background): “…power of AF2-predicted structures in the context of a smaller set of 38 proteins (Zhang et al., 2022), and with a study that suggests AF2 has a tendency to predict holo structures (Saldaño et al., 2022), and in the case of GPCRs can be biased to produce the desired conformation (Heo & Feig, 2022).”