Bespoke Optimization

Objective

• Develop and improve tools for automated bespoke parameterization.

Objectives for 2020:

Team

Project leader:

Team members: @Joshua Horton @Chris Ringrose @Daniel Cole @David Mobley @Jeffrey Wagner @Pavan Behara @Jamshed Anwar (Deactivated)

Scientific questions

• Can we use optimization trajectory data to improve FF fitting and bespoke fitting?

• Can we use fragmentation during bespoke fitting to speed up the process and how well do the parameters transfer back to the parent molecule.

• Can we ML models like ani to speed up the derivation process of bespoke parameters and how compatible is this with our current default QC specification.

• Can we reuse data in QCArchive to speed up bespoke parameterization?

Possible solutions?

@Joshua Horton

OpenFF Bespoke workflow