Solvation FE for proteins based on Charmm36m parameters

20 views
Skip to first unread message

Roozbeh R

unread,
Jul 9, 2023, 1:34:26 PM7/9/23
to ermod-users
Hi ,
I am trying to estimate the Solv. FE for a protein segment (comprising almost 200 aminoacid residues) using ERMOD . Upon running:  tools/NAMD/gen_structure to obtain the required parameter i encounter the error msg. : 

"File "/content/ermod/share/ermod/tools/NAMD/gen_structure", line 438, in <module>
    sigtable[i][j] = sgm NameError: name 'sgm' is not defined" 

As i checked into this error, i think this is due to the inability of gen_structure script to read LJ potentials parameters from Charmm36m topology/parameter files which have different structure than the charmm22 topology files which are easily readable by gen_structure script.

Is there any solution/work around for this problem to calculate the solv. FE. for protein simulations based on charmm36m forcefield? 

thank you in advance for the support.

best regards,
roozi

Roozbeh R

unread,
Jul 9, 2023, 1:48:05 PM7/9/23
to ermod-users
just forgot to mention that i used ermod 0.3.8 (last stable version on sourceforge page of ermod) and my protein/solution MD simulation are run on NAMD3.
kind regards,
Roozi

Shun Sakuraba

unread,
Jul 9, 2023, 9:33:43 PM7/9/23
to ermod...@googlegroups.com
Hi,

It seems like a bug in the script. Could you tell us the version of
CHARMM36m file (e.g., "toppar_c36_jul18.tgz")?
> --
> You received this message because you are subscribed to the Google Groups "ermod-users" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to ermod-users...@googlegroups.com.
> To view this discussion on the web visit https://groups.google.com/d/msgid/ermod-users/363ea14d-d3f8-42e6-96e0-46b4566e81dbn%40googlegroups.com.



--
Shun Sakuraba

Roozbeh R

unread,
Jul 10, 2023, 9:27:11 AM7/10/23
to ermod-users
Hi,
I use toppar_c36_jul22.tgz. Also please note that in my .conf , there are several lines for topology files, like: 
parameters              toppar/par_all36m_prot.prm
parameters              toppar/toppar_water_ions.str
parameters              toppar/par_all36_na.prm
parameters              toppar/par_all36_carb.prm
parameters              toppar/par_all36_cgenff.prm
parameters              toppar/par_interface.prm
parameters              toppar/toppar_all36_moreions.str
parameters              toppar/toppar_ions_won.str
parameters              toppar/cam.str

This is because unlike charmm22 , in Charmm36m, force filed  parameters are written in different files according to the molecule type , so Ermod script shall be able to read the  required FF parameters for Solv. Fe calculation from different files with different structure , which i assume is the root cause of the problem am facing.


PS.: I just tried a workaround which did not work but i think it is a good idea to share it with you:
I edited charmm36m parameter file, par_all36m_prot.prm , by adding nonbonded interaction parameters (from charmm36m Water & Ions parameter files) into the Lennard-Jones section of it (at the end of  par_all36m_prot.prm) so that nonbonded parameters for all  atom types present in my MD run are stored in one file(the modified par_all36m_prot.prm). then I also edited my config. file so the topology file section refer to only one file, like below:
parameters              toppar/par_all36m_prot.prm

then I ran gen-structure in the hope that this procedure may resolve the problem, but again i got the following error msg:
"File "/content/ermod/share/ermod/tools/NAMD/gen_structure", line 438, in <module>
    sigtable[i][j] = sgm
NameError: name 'sgm' is not defined" 

I thoght sharing this ,may be of sme help to find a better way to solve the problem.

best regards,
roozi



Shun Sakuraba

unread,
Jul 11, 2023, 11:44:47 AM7/11/23
to ermod...@googlegroups.com
From the code, it seems that the error comes from typo.

sig = sgm_from_rmin2 * rmin
sigtable[i][j] = sgm

This apparently should be "sig" instead of "sgm".

But bigger problem is that this part of the code is likely untested for long and I am not sure whether it really works.
Since the very first commit the bug was there. If I remember correctly I did testing using NAMD. Probably I renamed except this one after testing and before the commit?

Could you test whether it works with the change above? Please let us know if you find any suspicious behavior.
Thank you,
> > To view this discussion on the web visit https://groups.google.com/d/msgid/ermod-users/363ea14d-d3f8-42e6-96e0-46b4566e81dbn%40googlegroups.com <https://groups.google.com/d/msgid/ermod-users/363ea14d-d3f8-42e6-96e0-46b4566e81dbn%40googlegroups.com>.
>
>
>
> --
> Shun Sakuraba
>
> --
> You received this message because you are subscribed to the Google Groups "ermod-users" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to ermod-users...@googlegroups.com <mailto:ermod-users...@googlegroups.com>.
> To view this discussion on the web visit https://groups.google.com/d/msgid/ermod-users/cf6fbb76-0528-4dc2-ab2e-d70d95dc3c45n%40googlegroups.com <https://groups.google.com/d/msgid/ermod-users/cf6fbb76-0528-4dc2-ab2e-d70d95dc3c45n%40googlegroups.com?utm_medium=email&utm_source=footer>.

--
Shun Sakuraba

Roozbeh R

unread,
Jul 15, 2023, 4:30:42 PM7/15/23
to ermod-users
Yes it worked when i replaced "sig" instead of "sgm" . Just one point i like to mention: when i check generated LJtable files, it seems that sigma values are different from values in charmm ff parameter files for LJ potential. for example for TIP3P model of water , sigma for OT (oxysgen atom) is 1.7682 in charmm parameter file but as i check in the LJ file generated by the code, the value listed in ljtable  is : +3.150574e+00 for this type of atom. i assume it is converted ( e.g. via divided by some characteristic length to facilitate the calculations ) but i prefer to ask for your advice regarding this issue and the reason for the difference. 

best regards,
Roozi

Shun Sakuraba

unread,
Jul 16, 2023, 11:15:56 PM7/16/23
to ermod...@googlegroups.com
CHARMM parameter files for CHARMM and NAMD uses rmin/2 as "radius" while ljtable use sigma-epsilon notation. (Note SltInfo/MolPrm files use rmin/2 as input if there are no "NBFIX" specified).

Best,

On 2023/07/16 5:30, Roozbeh R wrote:
> Yes it worked when i replaced "sig" instead of "sgm" . Just one point i like to mention: when i check generated LJtable files, it seems that sigma values are different from values in charmm ff parameter files for LJ potential. for example for TIP3P model of water , sigma for OT (oxysgen atom) is 1.7682 in charmm parameter file but as i check in the LJ file generated by the code, the value listed in ljtable  is : +3.150574e+00 for this type of atom. i assume it is converted ( e.g. via divided by some characteristic length to facilitate the calculations ) but i prefer to ask for your advice regarding this issue and the reason for the difference.
>
> best regards,
> Roozi

--
Shun Sakuraba
Reply all
Reply to author
Forward
0 new messages