Hi Thomas,
Thank you very much for your precious suggestion. Yes, MMGBSA without considering entropy is quick compared to MD. But it takes very long (couple of days) to calculate entropy contribution using normal-mode analyis for a single frame of the whole complex with 90 amino acids and 7 nucleic base pairs. Since entropy calculation requires large number of frames, I have been thinking about a way to accelerate the calculation meanwhile keep the accuracy.
Thanks
________________________________
From: Thomas Cheatham <tec3.utah.edu>
Sent: Friday, 11 March 2022 4:33 PM
To: amber.ambermd.org <amber.ambermd.org>
Subject: Re: [AMBER] calculate entropy using MMPBSA.py
I do not fully understand the error, but I am trying to understand the premise of reducing the system size for MM-PBSA because of the cost (even if publications did this). MM-PBSA is not computationally costly and how do you know what you "cut" from the system does not actually contributed to the dG? Do the whole system. Yes, doing the whole system could add more noise that a reduced system may shield, but the whole system is more correct. [Brings up issues of sampling, convergence, ... which I would be happy to chat about]
Post-processing the MM-PBSA/GBSA is relatively trivial (and I mean by procedure and computational cost and not necessarily in terms of interpretation which requires experience) and much less costly than the production MD. Of course, I hope you have multiple replicates of the system with different initial conditions (at least 3 replicates) to test for similarity and reproducibility.
--tec3
p.s. entropy much harder to estimate than potential energy (enthalpy). Requires "convergence" which is elusive. I'll re-emphasize multiple replicates...
____________________________________
From: Ming Tang <m21.tang.qut.edu.au>
Sent: Thursday, March 10, 2022 11:11 PM
To: amber.ambermd.org
Subject: [AMBER] calculate entropy using MMPBSA.py
Dear list,
I am calculating binding entropy of dna to protein using MMPBSA. In the literature, researchers truncate the system by deleting the protein residues further than 8 Å away from the ligand to reduce the computational cost. I tried to extract the corresponding residues' trajectory and parmtop, and they match well in chimera visualisation. But gave me the below error:
File "/home/tm/Downloads/amber16/bin/MMPBSA.py", line 99, in <module>
app.file_setup()
File "/home/tm/Downloads/amber16/lib/python2.7/site-packages/MMPBSA_mods/main.py", line 156, in file_setup
self.mpi_size, str(external_progs['cpptraj']), self.pre)
File "/home/tm/Downloads/amber16/lib/python2.7/site-packages/MMPBSA_mods/make_trajs.py", line 104, in make_trajectories
traj.Strip(INPUT['ligand_mask'])
File "/home/tm/Downloads/amber16/lib/python2.7/site-packages/MMPBSA_mods/make_trajs.py", line 646, in Strip
raise InternalError('Cannot pass no mask to Strip!')
InternalError: Cannot pass no mask to Strip!
The command and control file work well for the full complex. Do I have to use tleap to process the truncated complex and get the coordinates and parmtop file that can be recognised by amber?
Thanks,
Tammy
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
https://urldefense.com/v3/__http://lists.ambermd.org/mailman/listinfo/amber__;!!NVzLfOphnbDXSw!WsQ8gI9DWxrmel3qvO2bzRD6oueiDatzlBYPPQ_sQA3syWXc5U6fOLytBMu2oQxiDvHO$
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
https://urldefense.com/v3/__http://lists.ambermd.org/mailman/listinfo/amber__;!!NVzLfOphnbDXSw!WsQ8gI9DWxrmel3qvO2bzRD6oueiDatzlBYPPQ_sQA3syWXc5U6fOLytBMu2oQxiDvHO$
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Fri Mar 11 2022 - 16:30:03 PST
