I've been running several rounds of RosettaMatch lately and I'd like them to go faster. I'm pretty sure writing matches is the part slowing it down. Is there anyway to limit the number of matches it outputs? For example, for the same matched sequence (say H133H135H432), I get anywhere from 50-400 cloudPDB files. I don't want more than the first one or two, because the rest are almost always just different ligand positions and I'll end up running EnzDesign anyways. I'm using these flags:
-match_grouper SameSequenceGrouper
-output_matches_per_group 100
-only_enumerate_non_match_redundant_ligand_rotamers true
Are there other/different flags that would alleviate writing huge numbers of nearly identical matches?? Or anything else that would speed it up?
Do you add the option "-match:output_format CloudPDB" to your runs? More or less the same ligand positions should be all in the same file with the normal CloudPDB output. IIRC, the default is CloudPDB, but I don't know if there's something else going on that is turning off CloudPDB output. (If you're getting a multi-model-esque output with disconnected sidechains in the subsequent models, then you are getting a CloudPDB output.) The other option you may want to try is "-match:consolidate_matches"
CloudPDB output is what you want to cut down on over-production of output files, especially if you're using the SameSequenceGrouper. Past that, you could possibly play around with the "-match:grouper_downstream_rmsd" The default is 1.5, but you could try increasing it, to combine more things together.
Failing that, you could try adjusting how much sampling you do. That is, try eliminating some of the -ex flags, if present, or changing the sampling level in your match constraints file, to resuce the number of redundant samples. You need to be careful about this, as if you reduce it too much, you'll lose all your matches.