If those are calcium atoms, you should have 3 jumps total: A to X, A to calcium 1, A to calcium 2. Probably some intervening code has changed and is no longer properly assigning a jump; probably the two calciums are connected improperly by bond; I’m guessing that’s what is going on with edge EDGE 736 737 -1.
I don’t know how to trace where code in the middle has changed between PyRosetta versions, so let’s look at some of the other options?
A) If the CA atoms are not at your interface, just delete them, and add them back in later.
It appears the between-proteins jump is now numbered 1; try setting your movers to work on jump number 1.
C) If neither of those work, we can try repairing the fold tree. I think it can be repaired by deleting EDGE 736 737 -1 and replacing it with an EDGE 735 737 2. (So long as the protein-to-CA jumps aren’t flexible, it doesn’t really matter where in protein 1 they originate). (Also, I don’t mean literally editing the string here, I mean modifying the FoldTree. I think there may be a text-based FoldTree read in method though). If that’s too vague I can describe it in C++ better.