Create custom equilibration scheme for Lipid21-POPC membrane runs

[David-Araripe]

It's not going to make it to the QligFEPv2 paper, but I have a new package converting PyMemDyn-generated united-atom (UA) POPC lipids to all-atom (AA) lipids, including hydrogens. The geometries aren't fully optimized, so they crash on eq1. I'm writing this issue to follow up in the near future.

Changing eq1.inp to a shorter step size with double the amount of steps without shake has enabled the equilibration to run:

[MD]
steps                     20000 ! This number was doubled
stepsize                  0.01 ! This used to be 0.1
temperature               1
bath_coupling             0.2
random_seed               SEED_VAR
initial_temperature       1
shake_solvent             off ! this used to be on
shake_hydrogens           off
shake_solute              off
lrf                       on
separate_scaling          on

⚠️ can probably be optimized e.g.: introduce eq0 with the shorter step size

FEP is still crashing on eq5, with shake failing on HOH hydrogens. Current input:

[MD]
steps                     50000
stepsize                  2.0
temperature               T_VAR
bath_coupling             10.0
shake_hydrogens           on
shake_solute              on
shake_solvent             on
lrf                       on
separate_scaling          on

Based on this, tasks for following up:

• Package the solution for converting UA to Lipid21 AA POPC lipids.
• Optimize our workflow for membrane proteins using these membrane molecules.
• Compare performance.