Page Comparison

Date

25 Feb 2020

Participants

• David Mobley

• Lee-Ping Wang

• Hyesu Jang

• Jessica Maat (Deactivated)

Goals

• Determine new procedure for selecting QM datasets for fitting (potentially for May meeting release, if ready in time)

• Divide up work to accomplish that

Notes:

Coverage of current QM dataset (Su)

...

Lee-Ping: Possible way to deal with coverage: Start with very big list of molecules. Build a list of molecules which use each parameter (being careful that for torsions they are exocyclic torsions) then you do clustering within that list.  Chemical similarity clustering for all moelcules molecules which use each parameter, then pick diverse molecules which use that parameter (e.g. five most diverse). 

...

Talk with Jeff: Class structure? script? Where does it live? 

Action items

•  Hyesu Jang Create Confluence page listing all available datasets (with Jessica Maat (Deactivated) enlisting David Mobley as needed)
•  Jessica Maat (Deactivated) develop prototype notebook which takes a FF and a set of molecules and a target parameter (ID) and picks the five most diverse molecules using that parameter. Should also take an optional argument which is a list of molecules to exclude (so that molecules which have already been used in other sets can be skipped)
•  Jessica Maat (Deactivated) and Hyesu Jang reach out to Jeffrey Wagner to discuss architecture of tools to be constructed, plan for sustainability and for where they should live. [Scheduled this meeting for Wednesday March 4 10 am -JM]
•  Hyesu Jang to determine how to enumerate protonation states and tautomers without doing semiempirical calculations (to speed set prep) talking to Chaya Stern (Deactivated) if needed, or if it can’t be done via that route, getting back to David Mobley for help with ideas. [ Update from Chaya: “@Hyesu Jang, the states module in fragmenter generates reasonable protonation / tautomer states. It uses quacpac and does not need AM1 calculations so is fast.
  https://github.com/openforcefield/fragmenter/blob/master/fragmenter/states.py “
•  Jessica Maat (Deactivated) and Hyesu Jang to come up with their goal timeline
  

Decisions

• Decided to make systematic approach for selecting molecules for QM data generation & fitting given a target dataset; this will be applied dataset-by-dataset to select new molecules for use in fitting
• Will attempt to select/redesign a new QM dataset for fitting rather than simply extending our prior QM dataset
• Decided on tentative algorithm for molecule selection approach