Previous Thread
Next Thread
Print Thread
computational alanine scanning
#17399 03/10/08 02:26 PM
Joined: Mar 2007
Posts: 148
J
jeffrey Offline OP
Forum Member
OP Offline
Forum Member
J
Joined: Mar 2007
Posts: 148
Dear all,

I am wondering if the compuational alanine scanning method has been implemented in CHARMM and in which version.

If not, is there an alternative way in CHARMM to analyze the per-residue contribution to the protein-protein binding free
energy?

I am now using the c31b2.

Thanks for the time.

----
Jeffry

Re: computational alanine scanning
jeffrey #17400 03/10/08 03:15 PM
Joined: Nov 2003
Posts: 36
Forum Member
Offline
Forum Member
Joined: Nov 2003
Posts: 36
Have you looked into free energy perturbation (FEP)? pert.doc


Joshua Ward Graduate Student Purdue University Department of Medicinal Chemistry and Pharmacology
Re: computational alanine scanning
jmwict13 #17401 03/11/08 02:01 PM
Joined: Mar 2007
Posts: 148
J
jeffrey Offline OP
Forum Member
OP Offline
Forum Member
J
Joined: Mar 2007
Posts: 148
Thanks very much, Josh.

I read the pert.doc, but still have some confusions about the command MKPRES:

Suppose that:

firstly, I have a MD trajectory of a protein-protein complex, containing 500 snapshots. Then the binding free energy is calculated by combining MM/PBSA and NMA or QUSI commands.

secondly, I want to check the contribution to the binding free energy by one of the residues, which is a GLN. At first step, I use the MKPRES command to generate a rtf file including a PATCH from GLN to ALA. Then the original PSF and the coordinates in trajectory are read. Finally, the PATCH is used to modify the original PSF from GLN to ALA.

Thirdly, the binding free energy is calculated again by MM/PBSA and NMA or QUSI commands.

Do I get the idea?

Any suggestion is greatly appreciated.

Thanks in advance.

----
Jeffry

Re: computational alanine scanning
jeffrey #17402 03/11/08 03:12 PM
Joined: Nov 2003
Posts: 36
Forum Member
Offline
Forum Member
Joined: Nov 2003
Posts: 36
No, MMPB-SA is a completely different approach from FEP. I haven't done FEP myself, so can't offer advice on the specifics. You should be able to find a bunch of references about it the literature.

Another keyword to search for is Thermodynamic Integration.

I assume that MMPB-SA is going to be much less accurate and even meaningless here. Maybe it could still work, but the technique is designed for binding studies, and if you apply it to this case, you aren't going to account for the energy loss associated with annihilating the extra atoms of the GLN sidechain when you mutate it to ALA. You really need to try and conserve mass when you apply it.


Joshua Ward Graduate Student Purdue University Department of Medicinal Chemistry and Pharmacology
Re: computational alanine scanning
jmwict13 #17403 03/11/08 04:29 PM
Joined: Mar 2007
Posts: 148
J
jeffrey Offline OP
Forum Member
OP Offline
Forum Member
J
Joined: Mar 2007
Posts: 148
Thanks,Josh.

Suppose there are two chains of a complex, A and B.
What I want to do is to calculate the relative change of the binding free energy which can be compared with experimental results of mutations from A(GLN)*B to A(ALA)*B.

So, if I can change the psf from A(GLN)*B to A(ALA)*B and use the A(GLN)*B trajectory, in which we suppose that the mutation from A(GLN)*B to A(ALA)*B only change the local conformation, then much time can be saved.

What confuses me now is whether I can use the patch file generated by MKPRES to implement the mutaion?


Thanks very much for any advice.


----------
Jeffry

Re: computational alanine scanning
jeffrey #17404 03/12/08 12:42 AM
Joined: Sep 2003
Posts: 8,506
rmv Online Content
Forum Member
Online Content
Forum Member
Joined: Sep 2003
Posts: 8,506
Two separate simulations are the most rigorous way to do this for any P-B based approach to estimating binding free energy. Also, each trajectory must exactly match (number and order of atoms) the PSF used to create it, so you really need trajectories for both cases, Gln and Ala mutant.

Using an FEP method such as PERT may be a reasonable option as well, but you really need to get familiar the intricacies of topology files to be very successful.

For a single case, manually changing the coordinate file and making a new PSF with GLN replaced by ALA may be all you need to do.

If I had to do a lot of this for multiple amino acid types, I'd probably create a custom RTF file (designed for APPEND usage) with a series of patches (PRES definitions) that delete anything past Cbeta for each residue type, and adds Cbeta for Gly.

Re: computational alanine scanning
rmv #17405 03/12/08 01:46 PM
Joined: Mar 2007
Posts: 148
J
jeffrey Offline OP
Forum Member
OP Offline
Forum Member
J
Joined: Mar 2007
Posts: 148
Thanks very much for the reply, Rick.
Indeed, the method I am going to use is an approximate one. To run the seperate MD simulations of both GLN and ALA mutant is an exact way to calculate the relative free energy change. But when the mutation doesn't change the conformation significantly, the error should be small enough between the two methods.


--------
Jeffry

Re: computational alanine scanning
jeffrey #17406 03/12/08 04:26 PM
Joined: Sep 2003
Posts: 8,506
rmv Online Content
Forum Member
Online Content
Forum Member
Joined: Sep 2003
Posts: 8,506
The problem is that the Gln->Ala mutation might be expected to induce conformational changes in nearby residues, which could affect the contact surface, and thus the binding free energy.


Moderated by  BRBrooks, lennart, rmv 

Link Copied to Clipboard
Powered by UBB.threads™ PHP Forum Software 7.7.4
(Release build 20200307)
Responsive Width:

PHP: 5.6.33-0+deb8u1 Page Time: 0.015s Queries: 30 (0.007s) Memory: 0.9524 MB (Peak: 1.0682 MB) Data Comp: Off Server Time: 2020-10-29 07:34:34 UTC
Valid HTML 5 and Valid CSS