OTF-General 0.1.0

Alchemical binding free‐energy workflow

OTF General

Implementation of On-The-Fly (OTF) Optimization for alchemical binding free energy simulations using thermodynamic integration in AMBER20. See 10.26434/chemrxiv-2023-rtpsz for details.

Setup

1. Create a directory, clone otf_general.
2. Install Anaconda.
3. Install dependencies:
   • pymbar==4.0.3
   • alchemlyb==2.3.1
   • scipy==1.8.1
   • AMBER20
   • GAUSSIAN 09 or GAUSSIAN 16 (for ligand parameterization)

Algorithm

On-The-Fly Resource Optimization for Binding Free Energy Simulations

1. Perform an initial run of the simulation.
2. Apply Automatic Equilibration Detection (AED) and decorrelation:
   • Split gradient data in half and bin it into histograms.
   • Calculate the Jensen-Shannon (JS) distance between the histograms.
3. Evaluate the JS distance:
   • If JS distance < 0.1:
     • Terminate simulation (converged).
   • Else:
     • Check if simulation time is less than 6.5 nanoseconds (ns):
       • If yes:
         • Proceed to Step 4 (short additional simulation).
       • Else:
         • Check if there are more than 50 decorrelated samples:
           • If yes:
             • Terminate simulation (sufficient decorrelated samples).
           • Else:
             • Proceed to Step 4 (short additional simulation).
4. Run a short additional simulation and increment the simulation time.
5. Evaluate total computed simulation time:
   • If total simulation time < 10.5 ns:
     • Repeat from Step 2.
   • Else:
     • Terminate simulation (time limit reached).