2022-06-08 Thompson/Wagner Check-in meeting notes

General updates

• JW – In our previous meeting with external stakeholders, I mentioned a tradeoff between “JW needs to help the science team with benchmarking” and “JW can help with Interchange” - I’ll be mentioning this more in talks with our funders because frankly this is a tangible tradeoff for “money → progress” that we can present, and it does reflect the bottlenecks in our processes/resources.

Interchange regression testing

• Partial charge difference magnitudes?

  • The failures reflect differences of over 0.005.

  • MT – I suspect that we’d still see the run-to-run variation if we ran the same toolkit version twice

  • JW – My suspicion is that there’s some chemistries that the conformer generator thinks should be “planar”, but the AM1 thinks should be “puckered”, and the AM1 optimization starts randomly goes one way or the other on the first step.

    • MT – I see a lot of nitrogen-containing rings in these structures. So it could be components disagreeing about aromaticity.

  • 99.45% of dataset had no atoms with more than 0.005 e- different

  • Of remaining

  • 

PR+issue clearance

• JW – Is #1316 done? There are a bunch of “TODO: Move this into interchange” statements and I don't want to push for closure if those are on your mind right now, but I would approve for merge if it’s ready.

• JW – Approved 1310 with some suggestions

• JW – Suggested closing 1309 because of added difficulty

• JW – (not asked) MT, could you review #1213?