From: Lennart Nilsson (Lennart.Nilsson_at_ki.se)
Date: Tue Oct 30 2018 - 09:31:59 CDT
Just send it to paramchem directly, before doing a minimization if you suspect there may be missing parameters; the CHARMM “energy” command will also report missing parameters right away (probably similar to an energy minimization in NAMD). It is hard to see how it could be done more efficiently, given that we do not trust magic and mind-reading.
Vänligen,
Lennart Nilsson
Tel: 08-524 81099
From: owner-namd-l_at_ks.uiuc.edu [mailto:owner-namd-l_at_ks.uiuc.edu] On Behalf Of Francesco Pietra
Sent: den 30 oktober 2018 23:17
To: gumbart_at_physics.gatech.edu
Cc: NAMD <namd-l_at_ks.uiuc.edu>; Joshua.Vermaas_at_nrel.gov
Subject: Re: namd-l: parameter request for PRES bonding
Both paramchem (faster) and FFTK (slower) will automatically identify the missing parameters.
However, after patching for a covalent bond between ligand and protein, it is only by running a minimization on .psf/.pdb, obtained from psfgen, that tells you about missing parameters for the new situation. Am I wrong, or is anything more efficient?
cheers
fp
On Mon, Oct 29, 2018 at 7:07 PM Gumbart, James C <gumbart_at_physics.gatech.edu<mailto:gumbart_at_physics.gatech.edu>> wrote:
So autogenerate just means that all possible angles and dihedrals given the topology of the molecule are added to the PSF. Each of these will require a corresponding parameter. If you are being asked for a parameter that doesn’t involve at least one of the changed atoms, then something might be wrong. But consider even just changing one atom type has the potential to require dozens of new angle and dihedral parameters if it hasn’t been used previously in that context in the force field.
Both paramchem (faster) and FFTK (slower) will automatically identify the missing parameters.
Best,
JC
On Oct 29, 2018, at 3:07 AM, Francesco Pietra <chiendarret_at_gmail.com<mailto:chiendarret_at_gmail.com>> wrote:
I was surprised that parameters were also requested for bonds rather removed from NE2-C9, such as between C10 and its adjacent C11. Besides that, requests seemed to be endless, for situations with non-obvious analogies. So that, with a natural-product ligand, covalently bonded to a protein receptor, running a decent MD may take months of efforts with QM calculations.
When that arises as a small part of a large project, involving many ligand-protein cases, one (myself) becomes frustrated. Unless one decides that each case is a publication, which, unfortunately, is a decision by many.
Thanks for your view
francesco
On Mon, Oct 29, 2018 at 12:09 AM Vermaas, Joshua <Joshua.Vermaas_at_nrel.gov<mailto:Joshua.Vermaas_at_nrel.gov>> wrote:
That seems perfectly normal to me. You changed the atomtypes on C9 and NE2 (and some others it looks like, but I haven't been following your exploits all that closely), so any term that touches those atoms (and there are likely a ton of them) will be missing parameters. What behavior did you expect?
-Josh
On 2018-10-27 01:34:07-06:00 owner-namd-l_at_ks.uiuc.edu<mailto:owner-namd-l_at_ks.uiuc.edu> wrote:
The following PRES generated correctly C9-NE2 bonding between organic ligand and HSP of protein
PRES ZAHI 1.116 ! patch for ligand-HSP bonding
! Patch must be 1-HSP and 2-ZPN
! use in patch statement
! follow with AUTOgenerate ANGles DIHEdrals command
GROUP !
ATOM 2C10 CG311 0.332 !
ATOM 2C9 CG314 0.199 !
ATOM 2H91 HGA1 0.09 !
ATOM 2C8 CG321 0.127 ! CD2--NE2--CE1
GROUP ! /
ATOM 2CD2 CPH1 0.219 ! C10--C9--C8
ATOM 2NE2 NR3 -0.20 ! |
ATOM 2CE1 CPH2 0.349 ! H91
DELETE ATOM 2H92
DELETE ATOM 1HE2
BOND 2C9 1NE2
IMPR NE2 CD2 CE1 C9
IMPR NE2 CE1 CD2 C9
Problems arise, however, on minimization. Besides the expected request for
BONDS
CG314 NR3 261.00 1.4900 !
CG311 HGA2 309.00 1.1110 !
ANGLES
HGA1 CG314 NR3 51.5 107.5 !
CG311 CG314 NR3 117.00 124.00 !
minimization also required angle
CG2O5 CG321 CG314
and many other angles outside the strict area of bonding. Even on satisfying these requests, dihedrals were also requested.
Question: is that normal despite the <AUTOgenerate ANGles DIHEdrals> or did I miss something?
thanks for advice
francesco pietra
När du skickar e-post till Karolinska Institutet (KI) innebär detta att KI kommer att behandla dina personuppgifter. Här finns information om hur KI behandlar personuppgifter<https://ki.se/medarbetare/integritetsskyddspolicy>.
Sending email to Karolinska Institutet (KI) will result in KI processing your personal data. You can read more about KI’s processing of personal data here<https://ki.se/en/staff/data-protection-policy>.
This archive was generated by hypermail 2.1.6 : Tue Dec 31 2019 - 23:20:17 CST