Possible to output ligand docked poses in mmCIF, SDF, etc?

[greg-rezo]

I'm running GALigandDock via RosettaScripts and found that a small but significant fraction of the time, the PDB records of the ligand docked pose (HETATM, CONECT) aren't sufficient to fully reconstruct the ligand geometry in RDKit. I get a lot of valence errors and kekulization errors. I have tried a step that calls RDKit's AssignBondOrderFromTemplate to map the docked pose onto the pre-rosetta compound input template which helps, but still fails on some molecules.

AFAICT, this is a limitation of the PDB format for small molecules and am wondering if it's possible to output the docked poses in a more suitable format such as mmCIF, SDF, etc.

This is my command:

rosetta_scripts -overwrite -s=target.pdb -parser:protocol=vsx.xml -gen_potential -beta_cart -no_autogen_cart_improper -multi_cool_annealer 10 -score::hb_don_strength hbdon_GENERIC_SC:1.45 -score::hb_acc_strength hbacc_GENERIC_SP2SC:1.19 -score::hb_acc_strength hbacc_GENERIC_SP3SC:1.19 -score::hb_acc_strength hbacc_GENERIC_RINGSC:1.19 -parser:script_vars liglist=codes_for_2mols.txt -constant_seed -jran 2046530742 -extra_res_fa=L00.params -extra_res_fa=L01.params  -out:prefix=/tmp/dock_output/ -out:file:scorefile /tmp/score.json -out:file:scorefile_format json -out:levels all:150 -missing_density_to_jump -ignore_unrecognized_res -no_his_his_pairE -flip_HNQ -ex1 -ex2 -ex1aro -no_optH=false'

Appreciate the assistance.

Best,
Greg

[roccomoretti]

You should be able to generate mmCIF outputs with the option -out:mmCIF. Note that this flag typically only affects the "main" outputs of the protocol, and some intermediately dumped structures may be hard-coded to be output as PDBs.

A caveat for Rosetta mmCIF output is that it may not fully contain all the information which your downstream application is expecting. It's primarily focused on outputting the coordinate table, and all the additional annotational tables may not be fully represented.

[greg-rezo]

Thanks. I added that flag to the command above but get a SIGBART. This is using the rosettacommons/rosetta:serial-405 docker image (although I also see it with the mpi-386 tagged image.

I'm not tied to the mmCIF approach. If you have any other suggestions for getting the compound out of rosetta's internal state with higher fidelity (silent file?), that'd be appreciated.

********  (C) Copyright Rosetta Commons Member Institutions.  ***************
* Use of Rosetta for commercial purposes may require purchase of a license. *
********  See LICENSE.md or email license@uw.edu for more details. **********
 ------------------------ Begin developer's backtrace ------------------------- 
BACKTRACE:
/usr/local/bin/libutility.so(utility::save_crash_report(char const*, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&, int)+0x38) [0x7ffffcea8728]
/usr/local/bin/libutility.so(+0x198a5b) [0x7ffffce8da5b]
/lib/x86_64-linux-gnu/libstdc++.so.6(+0xaa37c) [0x7ffffcbb937c]
/lib/x86_64-linux-gnu/libstdc++.so.6(+0xaa3e7) [0x7ffffcbb93e7]
/lib/x86_64-linux-gnu/libstdc++.so.6(+0xaa699) [0x7ffffcbb9699]
/usr/local/bin/libcifparse.so(+0x31465) [0x7fffeec23465]
/usr/local/bin/libcore.3.so(core::io::mmcif::dump_cif(std::ostream&, std::shared_ptr<core::io::StructFileRep>, core::io::StructFileRepOptions const&)+0x69d0) [0x7ffff2648b60]
/usr/local/bin/libcore.3.so(core::io::mmcif::dump_cif(core::pose::Pose const&, std::ostream&, std::shared_ptr<core::io::StructFileRepOptions const>)+0x8d) [0x7ffff264b91d]
/usr/local/bin/libprotocols.1.so(protocols::jd2::mmCIFJobOutputter::dump_pose(std::shared_ptr<protocols::jd2::Job const>, core::pose::Pose const&, utility::io::ozstream&, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&)+0x17e) [0x7ffffdd0787e]
/usr/local/bin/libprotocols.1.so(protocols::jd2::wwPDBJobOutputter::final_pose(std::shared_ptr<protocols::jd2::Job>, core::pose::Pose const&, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&)+0x737) [0x7ffffdd8e387]
/usr/local/bin/libprotocols.1.so(protocols::jd2::JobDistributor::job_succeeded(core::pose::Pose&, double, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&)+0x142) [0x7ffffdcdde42]
/usr/local/bin/libprotocols.1.so(protocols::jd2::FileSystemJobDistributor::job_succeeded(core::pose::Pose&, double, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&)+0xd) [0x7ffffdcd0efd]
/usr/local/bin/libprotocols.1.so(protocols::jd2::JobDistributor::write_output_from_job(core::pose::Pose&, std::shared_ptr<protocols::moves::Mover>, protocols::moves::MoverStatus, unsigned long, unsigned long&)+0x65b) [0x7ffffdcdfb5b]
/usr/local/bin/libprotocols.1.so(protocols::jd2::JobDistributor::run_one_job(std::shared_ptr<protocols::moves::Mover>&, long, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> >&, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> >&, unsigned long&, unsigned long&, bool)+0x25ee) [0x7ffffdce361e]
/usr/local/bin/libprotocols.1.so(protocols::jd2::JobDistributor::go_main(std::shared_ptr<protocols::moves::Mover>)+0x238) [0x7ffffdce4a48]
/usr/local/bin/libprotocols.1.so(protocols::jd2::FileSystemJobDistributor::go(std::shared_ptr<protocols::moves::Mover>)+0x57) [0x7ffffdccbde7]
rosetta_scripts(+0x4a96) [0x555555558a96]
/lib/x86_64-linux-gnu/libc.so.6(__libc_start_main+0xf3) [0x7ffffc924083]
rosetta_scripts(+0x4d3e) [0x555555558d3e]
 ------------------------- End developer's backtrace -------------------------- 


AN INTERNAL ERROR HAS OCCURED. PLEASE SEE THE CONTENTS OF ROSETTA_CRASH.log FOR DETAILS.


terminate called after throwing an instance of 'EmptyContainerException'
  what():  Message: "No columns in table" generated at: ISTable::InsertRow

Got some signal... It is:6
Signal 6 (SIGABRT) means that the process was aborted.  This usually means an internal Rosetta error caused by (often) bad inputs, (sometimes) developer error, or (rarely) hardware problems.

[roccomoretti]

We recently updated the mmCIF handling. Ironically, it's the commit that's right after rosettacommons/rosetta:serial-405
I'm not sure if we've put out a more recent Docker container yet.

If there's any way to update to a more recent version (e.g. the Rosetta 3.15 release), that will likely get you better results when working with mmCIF files.

Failing that, GALigandDocking won't change the chemical identity of the ligand, just the atom coordinates. So if you can somehow merge the coordinates of the PDB file with the bonding information from the original Mol2 file, that would likely work. (Though I don't know of a script/application to do that off-hand.)

[greg-rezo]

Thanks, that makes sense. I'll try the newer version and fallback to improving on RDKit's matching logic.