Does anyone know a method for generating a disordered protein ensemble with AlphaFold-Multimer? The protein I am interested in is disordered and is bound to a protein that is mostly folded. I know of some methods that have been described in the literature (for instance, varying the random seed), but if someone has implementation details on such an approach I would be extremely grateful.
Hi,
I haven't heard of any implementation for the exact application you have in mind, but besides asking AF-Multimer to generate a bunch of models, and repeating this with different random seeds, the following 2 papers may also inspire you to [1] do some in silico mutagenesis (e.g. Ala-screening) on the conserved resideus of the IDP; and [2] re-run of AF-Multimer with subsampled MSAs:
https://pubmed.ncbi.nlm.nih.gov/35994486/
https://pubmed.ncbi.nlm.nih.gov/35238773/
Alternatively, you could do MD or Monte Carlo simulations. These are believed to produce conformers more representative for the Boltzmann ensemble.
Best of luck,
Tamas
- Badges
- Science topic