| |
|---|
Next week's meeting | |
Jang reports
| Message from Hyesu: I did some test this morning checking (1) the generation of interchange object from the spruced protein OEMols; and (2) from_polymer_pdb and to_file functions. I haven't seen any failure with the datasets I tested, but one thing I noticed was that when reading molecule from PDB and writing it back to PDB, it washes out the atom numbers (2nd column in pdb) and the b factors (11th column in pdb). I don't think they will matter much, but just report what I observed. JW – Spruce intercompatibility is of GREAT interest to us. OpenFF and OpenFE will be using Spruce for PL benchmarking HJ – Adding hydrogens is the big one, I didn’t notice sensitivity to other settings JW – I will document that we won’t preserve b-factors HJ – When I roundtrip structures with existing b-factors, it gets set to 20 for all atoms. MT – I don’t see us setting a b-factor, or particular mention of the number 20 in our codebase General – This is probably some default behavior in OpenMM
JW – The atom numbers all being set to 0 is a bug. I don’t know what that happens. Â
|
 | JW – I just pushed a change that will only temporarily only have PDB loading if RDKit is installed. I’ll fix this ASAP, but HJ – Timeline for OpenMM import to interchange? JW – Uncertain (shows WBS) – AMBER and GROMACS exports are higher priority than any imports, because that will be needed for protein-ligand free energy benchmarking, and that will be needed for Rosemary. If we can’t raise funding to get a new person on the science team, I’m going to need to divert my own effort to help the science team after the biopolymer release. But if we can raise funding for a person on the science team, then I can try to accelerate Interchange development. So the timine for an OpenMM importer is uncertain. MT – Agree that exporters are critical for current FF development plans, so importers will come after. HJ – That sounds good.
Â
|