Hello,
I've been trying replicate the protein design methods from this paper, which uses RosettaScripts. I have been able to replicate most of the data, but I have been struggling to make the last design part t owork. The design_monomer_interface.xml file is used to design the residues of the generated three helical bundle protein (that has desired backbone structure, but not the desired residues) so that the bundles could bind to the target. I wanted to test the design_monomer_interface.xml protocol on a pdb generated through the paper methods, which includes the target protein and the three helix bundle protein. The original paper did not provide any input flags, so we attempted to use some input flags previously used in the paper for a different section, but we still keep running into errors. I uploaded the crash log, and the relevant files are in the Google Drive link below. Can someone please help me navigate through this?
https://drive.google.com/drive/folders/1Fn1wASs55gIkub83JqEJKOQ7kaFAPz6C?usp=share_link
This is the link to the original files used by the paper (also at the bottom of the paper): http://files.ipd.uw.edu/pub/SARS-CoV-2_binder_2020/scripts_models.zip
| Attachment | Size |
|---|---|
 ROSETTA_CRASH.log | 57.85 KB |
The issue looks to be with your fragment store. You're giving it a file with a tabular layout, but that's not what it wants.
Best I can tell, it looks like you actually want to provide it with a path name. This path name should at least contain a 'fragments/<store_name>/metadata.json' file, which will list a bunch of other files in that directory.
I think the directory tests/integration/tests/threefold_symm_peptide_design/backbone_profiler_database_06032014/ in the Rosetta distribution should contain an example of the sort of thing you're looking for. Obviously, this is a cut-down version used for testing. You would have to find/generate a more extensive version which would be used for the actual design process. Unfortunately I don't have details of which library was used, or how it was generated. My guess is that Longxing simply used a pre-existing database which was floating around the Baker lab (hence the lack of details on the generation).
The database is the IndexedStructureStore, used specifically for the Worst9mer/LeastNativeLike9mer filter. Publications by TJ Brunette and Alex Ford (with David Baker) are likely to have more details about it.
Worse comes to worse, you can simply comment out the use of the worst9mer in the XML. You'll lack that bit of quality control, but it should surmount the need for the database.
Hello,
I tried commenting out the lines that have worst9mer in the XML (specifically lines 194-195 and 255-256), but the program still spits out an error. This is a part of the output that I got, and I also uploaded the new crash log (all codes have added line breaks for readability, but is not in the original code):
From what I see, it seems like Rosetta is still trying to load something from the IndexedStructureStore database (even without the Worst9mer/LeastNativeLike9mer filter), but doesn't work, since I do not have the proper files.
I tried to provide a path name by including this in my input.flags (although I do not think this is the correct path name):
When you said I need to provide the fragment store, do you mean that i have to find the database IndexedStructureStore, and provide the path to that database? I am really new to Rosetta, and I am slightly confused by what I need to do in terms of setting this program up. Thank you so much for your help.
The StructProfileMover also uses the IndexedStructureStore, and that's what's crashing on you. (Specifically, it's the setup/definition of StructProfileMover which is triggering the error.)
If you want to use these movers and filters, you'll need to provide a properly generated and properly formatted fragment store database. I do not believe that the database used with these publications are provided with the regular Rosetta download. The version in the integration test directory is a cut-down version, and I mentioned it just so you have a sense of what sort of format you might be looking for. (Though the error message may mean that it's not 100% compatible with the settings you're using.)
It may be that you need to contact the authors of the paper you're interested in, and ask them about where the fragment store database they used for the paper came from. (That is, which other paper you need to reference for generating your own version, if the authors aren't able to provide you with a copy of the version they used.)