Just a quick question. I am running a few (thousand) structures through score_jd2 using the command below, and I am occasionally seeing this message I've never seen before:
No structure mods allowed during scoring!
Any idea how this would be tripped? Some of these structures may have missing atoms and need some rebuilding of side chains during pose construction, but shouldn't that occur before scoring? The structures themselves look fine, and can be loaded into PyRosetta with no issues.
score_jd2.linuxclangrelease -database $ROSETTA3/database -no_output -l ../PyIgClassify/DBOUT/renumbered_pdbs/PDBLIST.txt -ignore_unrecognized_res -ignore_waters -ignore_zero_occupancy false
If it's just score_jd2, then you shouldn't be seeing anything like that. You're right that rebuilding of missing atoms will happen before the scoring step, so that shouldn't be an issue.
Do the same PDBs consistently give you the error? If they do, then with an example PDB we have a chance of debugging what exactly is going on. If there's no consistency as to which PDBs are giving you the error, then there could be something wrong with your Rosetta compile. Try clearing out your build directory and recompiling. (Also, which version of Rosetta are you using which displays this issue?)
Just checking - you said that they load into PyRosetta fine, but do they also *score* fine under PyRosetta?
I'm using the developers trunk, in a branch. Only thing I added was a FeaturesReporter. When I run the structures through rosetta_scripts for reporting features, the PDBs load fine. And yes, the PDBs score fine in PyRosetta. Running a single PDB that reported this error through score_jd2 with the same cmd-line options gives no errors. Running with the -l flag consistently gives this error with those PDBs.
I swear there is a ghost floating around in the JobDistributor with my Rosetta compile whenever I add classes. I'll try in a non-modified branch soon and if it runs ok I will laugh maniacally and grab a beer.