[AMBER] Thermodynamic integration questions questions and questions

From: Debarati DasGupta <debarati_dasgupta.hotmail.com>
Date: Tue, 29 Oct 2019 15:59:12 +0000

Sent from Mail<https://go.microsoft.com/fwlink/?LinkId=550986> for Windows 10

From: Debarati DasGupta <debarati_dasgupta.hotmail.com>
Sent: Tuesday, October 29, 2019 11:57:11 AM
To: amber.ambermd.org <amber.ambermd.org>; David A Case <david.case.rutgers.edu>

I have a few queries about the Thermodynamic setup , just to make sure I am understanding it correctly:

My system is a kinase with a ethanol molecule as a weak binder. I have different ligands to try out with, but my first case is ethanol.

For an absolute free energy calculation method, we have to breakup the problem into 3 steps: 1. Ligand ( full charge + full interactions) ---> 2. Ligand (no charge, full interaction)----> 3. Nothing (No charge, no interaction)

I am trying to setup the entire thing using pmemd with soft-core potentials…

The steps should be decharging the ligand; then vdw+bonded terms are switched off...The ligand disappears. I hope I do not have to do the recharge as I am trying to annihilate ethanol from the binding pocket.

Q1. DO I have to do the decharging, vdw+bonded steps separately or they can done in 1 single step?

Will the decharge steps, the vdw steps be done in different lamda values? If yes, how many?

Q2. For my TI runs, will the lambda values be 0, 0.1,0.2,0.3,0.4,0.5...till 1.0. The tutorial does it in 10 steps, but we have to see if we need more lambdas, more is computationally exhaustive, but less can mean not so accurate calculations. How to choose lamdas, I have no solid publication which can help. Different people have tried with different values. Prof Simmerling has a very recent publication on TI calculations on the D3R targets, and he has a set of lambdas varying from 0.001 to 0.995 etc etc.. How he did choose those lamdas is not clear from reading the paper.

Q3. When my ligand actually translates to being nothing, do I have to keep restraints on it else with no interactions, no charges, it can float about anywhere in the box...I am bit worried as to how to tether a "nothing molecule" to the binding hotspot? There is no dummy atoms involved in the pmemd setup, so how to restrain disappearing molecules in AMBER?

Q4. Also for my runs, do I use the restart files from the previous lambda folder, like for example, when I do TI runs for lambda=0.4, do I use the .rst files from my 0.3 lambda simulation and so on and so forth?

I hope I could ask my questions properly, and it makes sense. Please do answer the questions if you will, that will help me a lot.

AMBER mailing list
Received on Tue Oct 29 2019 - 09:00:03 PDT
Custom Search