Re: Problems to parametrize molecules

From: Evandro Semighini (
Date: Wed Apr 15 2015 - 15:37:12 CDT


I generated the pdb and psf with charmm_gui and solvated the system with
water only, with the solvate package.

I tried this and there is the error "Unable to find bond parameters for
CG311 NG311 (Atoms 1 2)", which are bonded in the psf file.
I tried then to open the molecule and edit it with molefacture, correcting
the atom types, because it misses it, like you said.
I also corrected the double bonds (this one missing is simple) and saved
the molecule, solvated it and the same error shows up when I try to run

I tried also to parametrize the molecule with paramchem, but when I try to
generate the psf file, as described here:

I doesn't recognize the topology files of charmm36. (any of them)

There is any other method/program to parametrize the molecules?

Thank you again, Brian!


2015-04-15 15:04 GMT-03:00 Bennion, Brian <>:

> Hello,
> can you state how you used the molefacture plugin. I need to know what
> steps you are applying to solve this problem, EXACTLY.
> Molefacture will not automatically assign partial charges and in cases
> where the atom type is larger than 4 characters it will cut off the rest.
> It APPEARS that you are reading in the pdb file and molefacture is using
> that atom type.
> Please list the steps that you are following and the results you get from
> those steps and then what you THINK those steps should have done.
> Brian
> ------------------------------
> *From:* Evandro Semighini []
> *Sent:* Wednesday, April 15, 2015 9:00 AM
> *To:* Bennion, Brian
> *Subject:* Re: namd-l: Problems to parametrize molecules
> I got the psf from charmm_gui, but is seems that it's not working
> with my molecules.
> I tried now with another molecule and it doesn't recognize the bond
> parameters between two atoms that are bonded and listed at the BOND list of
> psf file.
> There is another way to parametrize this molecules ?
> Molefacture doesn't put the partial charges and misses the correct atom
> types (instead of CG311 it calls C3)
> Thank you
> 2015-04-14 17:57 GMT-03:00 Bennion, Brian <>:
>> still guessing but the psf file is not correct and that is what
>> counts. I am not sure how you got your psf file from the pdb or mol2 but
>> molefacture might be a plugin to use just to see the difference in the psf
>> file that it creates compared to the current psf file.
>> Brian
>> ------------------------------
>> *From:* Evandro Semighini []
>> *Sent:* Tuesday, April 14, 2015 1:51 PM
>> *To:*; Bennion, Brian
>> *Subject:* Re: namd-l: Problems to parametrize molecules
>> Thank you Brian,
>> I thought that this as the problem too, but I don't know how to fix it,
>> since the input files (pdb or mol2) have the correct bond orders.
>> I can change this with molefracture, before try the psfgen or the sites
>> I tried ?
>> 2015-04-14 17:45 GMT-03:00 Bennion, Brian <>:
>>> I am just guessing here...
>>> You need to pay attention to the what the error states. The errors
>>> suggest that the same atom is bonded to itself and it obviously cant find
>>> that in the parameter file.
>>> ------------------------------
>>> *From:* [] on behalf
>>> of Evandro Semighini []
>>> *Sent:* Tuesday, April 14, 2015 1:30 PM
>>> *To:*
>>> *Subject:* namd-l: Problems to parametrize molecules
>>> Hello NAMD community
>>> I am trying to run NAMD with the results of virtual screening assays I
>>> made, but I am not being able to parametrize the molecules properly.
>>> I read everything I found about it but, unfortunately, it didn't worked.
>>> I tried charmmgui and paramchem webservers to generate the psf files,
>>> but when I try to run NAMD, theres always some unknown atom type (this
>>> errors are from two different molecules)
>>> This atom types are at the par_all36_cgenff.prm and I added all the
>>> parameter files from the package to run this simulations.
>>> There were no high penalties for the atoms from the servers.
>>> Can you please help me to find where I am missing or show me where I
>>> should go to parametrize this molecules for the simulations ?
>>> Thank you,
>>> Evandro Semighini
>>> University of São Paulo - Brazil

This archive was generated by hypermail 2.1.6 : Tue Dec 27 2016 - 23:21:04 CST