If we are using alpha-fold protein structures for studying protein-protein interaction, do these structures need to be stabilized through MD simulation before actually studying their interaction pattern through MD simulations ?
There are quite amount of MD simulation using initial model from AlphaFold, but there are some head-ups before you conduct MD simulation using AlphaFold model.
For a great example: you need to consider the reliability of your AlphaFold model by looking at the "per-residue measure of local confidence".
It can be done by the predicted local distance difference test (pLDDT) from AlphaFold. For those regions with low pLDDT score. You shouldn't simply rely on the AlphaFold model.
Please see this practical guide:
https://www.ebi.ac.uk/training/online/courses/alphafold/inputs-and-outputs/evaluating-alphafolds-predicted-structures-using-confidence-scores/plddt-understanding-local-confidence/
Hi Shrikirti,
It is good idea to run a short MD simulation before studying protein-protein interactions with docking and dynamic simulation. Alphafold predictions may have artifacts like steric clashes or unrealistic conformations, especially in low confidence regions. An energy minimization or equilibration help to ensure that the modeled structures is more realistic, allowing for more reliable docking and interaction studies.
- Badges
- Science topic
- Similar topics
- Physical Sciences
- Materials Science
- Materials Chemistry
- Polymer Chemistry