touchpoint_29 Aug 2024 - omsf/OpenPharmMDFlow GitHub Wiki

Agenda

  1. Housekeeping and review TODO (15 mins)
  2. Review project board (30 mins)
  3. Plan next week (15 mins)

Attendees

  1. Mike
  2. Stephen
  3. Steven
  4. Amjad

Meeting notes

  • Next week’s meeting canceled
  • The budget has an allocation for allowing Mike to travel to BMS two times. Stephen proposed that we use the allocation to get Mike to come here and the team to do a hackathon. The initial suggestion is to do a 3 day hackathon where the first day can be dedicated to hardware and installation. This led to a discussion about whether BMS access to conda is restored and which led to the discussion about whether OpenPharmMDFlow supports windows. Mike suggested we could have do a light installation package. Stephen will check with his manager to make sure a 3 day hackthon is reasonable to do
  • In the last meeting we had started a longer Humira CG model MD run. Mike was able to run 633 ns simulation on one GPU. It took <12 hrs. However, the molecule collapsed to a ball quite fast. We also noticed that it doesn’t quite turn into a ball because of the longer bonds present in the system (added by Martini) to keep the secondary structure in place. We checked the RDF b/w water and not-water and found that the first neighbor distance is about 5 units (maybe A). This makes sense since the Martini water molecule radius is about 4 angstroms. Amjad found two papers that used Martini for proteins:
  1. Exploring the dynamics and interaction of a full ErbB2 receptor and Trastuzumab-Fab antibody in a lipid bilayer model using Martini coarse-grained force field
  2. Discovering Novel Small Molecule Compound for Prevention of Monoclonal Antibody Self-Association
  • Things to try:
    • Check if the temperature setpoint 310 K set in openMM is actually 310K. We will try rerunning it in gromacs because most people use gromacs to run Martini simulations
    • Check with the Martini email group if we are reinventing the wheel

Action Items

  1. Check with management if a 3 day hackathon is doable in BMS - Stephen
  2. Follow up on temperature scale from openMM+Martini - Mike
  3. Post question on Martini email group about modeling mABs - Mike
  4. Send name of library for antibody structure prediction - Stephen
  5. Create a new user story for Bespoke FF integration with OpenFE workflow - Yilin, Kaushik, Stephen (to follow up)
  6. Work on the slides - Team
  7. Lit review on Martini CG and charge models - Mike