I want to model a hetero 10-mer protein with 5 unique chains. can i do it with alphafold? or if there are other tools please let know how can i do it?
It is unlikely that you can model this in one single step. If each of the unique chains appears twice, your protein most likely is a homodimer of two heteropentamers. I would first try to build the best possible model of each of the individual chains. In looking at the pdb (www.rcsb.org) for potential modelling templates, I would pay careful attention to the template structures to see whether any of them could also serve as templates for the assembly of part or all of the complex. If this is not sufficient to build up the complex, you might have to do pairwise docking. However, try to incorporate as much experimental information as possible into your model.
Hi Annemarie, Indeed you can model it with alphafold. just dump the sequences and AF will work on them. In native AF, there is virtually no limit of monomer number (I used up to 30, a pentamer of hexamers). In colab, the limit is 20. However, the memory limit is around 4000 aa on a A100 GPU with 40 GB memory. Pay uch attention to the PAE to check the validity of the model.
Best, Chris