General updates | JW – Working on multicomponent PDB loading. Will be RDKit-only. ETA before the annual meeting. JW – CCP thing in the UK – Looks like a big win to me. What are your thoughts? MT – Not super excited - Seems like teaching 1st year PhD students. Not super cheap in terms of my time. Will probably take two weeks of prep once we actually get to it. Not sure how much uptake we’ll get from the group since they’re so early in their educational pipeline. JW – Two weeks sounds right - Feel free to tap Josh M for revisions/reformatting into longstanding docs. MT – I’ve done workshops on Interchange before, would need to review material for a while to see how reusable it is. The time could also be valuable as docs refreshment. JW – Happy to help limit scope/responsiiblity if JH or DC ask you to do a lot of work past what seems reasonable.
JW – WWang thing - Thinking about it. I’m going to ask her to write her own first draft and then come to you for help. I’d instruct you to spend up to 4 hours getting this set up, but we want to make sure that correctness/maintenance is her job. MT – Actually, I attended the newcastle meeting today. Talked with JH and WW about this and decided that the best way forward was to have a working session next week, and that’s on the calendar. WW is using an openmm plugin as the implementation for the OpenMM Force, but is still using OpenMM XMLs to define parameter assignment. So I think we can update that to use SMIRNOFF and interchange. There are two ways to make interchange plugins - One uses the existing OpenMM Force creation code in interchange is utopian and open-ended and could in theory allow lots of flexibility, and the other one asks the user two write their own OpenMM Force population logic manually. JW – I think you handled this perfectly. Good job. It’d be good to emphasize that WW will “own” the plugin, but that you’ll help make it. So you can encourage her to run energy tests afterwards and debug any remaining issues.
MT – What are some good “totally vanilla” GROMACS and AMBER inputs that reflect what our users want?
|