partial hallucination with multiple chains

[mouung]

I would like to design a protein complex with partial hallucination. My input pdb has two chains, A and B. Chain A is the target, so the whole sequence is fixed. In chain B, I would like to do a partial hallucination on a loop (between B100-B116). My script is something like this:

af_model = mk_afdesign_model(protocol="partial", use_templates=False, # set True to constrain positions using template input data_dir="/home/n721/AFD_params")

old_pos = "A1-A286,B1-B100,B116-B130"

af_model.prep_inputs(pdb_filename="ranked_0_286_HC.pdb", chain="A,B", pos=old_pos, # define positions to contrain fix_seq=True, # set True to constrain the sequence use_sidechains=False) # set True to restrain the sidechains

af_model.rewire(order = [0,1,2], # set order of segments loops = [0,15], offset=0)

However, I got an output pdb with just one long chain that connects my chain A and chain B. I guess rewire will connect all fragments into one chain. How do I only connect the two fragments in chain B with partial hallucination while keeping chain A intact?

[sokrypton]

this is not currently supported... but you can easily hack it!

try: af_model._inputs["residue_index"][cutpoint:] += 50 This will offset the residue index creating essentially a seperate chain.

[blakely3dnm]

I gave this a try for a 3 chain complex and while the index offset created a gap at the desired locations, the gap stays at most ~1 residue in length and resulting models still behave as though there is a connection restraint there, which gives incorrect folding.

In the partial hallucination rewire notebook I added the following to the second cell before the print line

model._inputs["residue_index"][41] += 50
model._inputs["residue_index"][78] += 50

I apologize in advance if this is a simple python error as this is not my area of expertise.