From: David Hardy (dhardy_at_ks.uiuc.edu)
Date: Fri Sep 28 2018 - 10:35:46 CDT
I can say for sure that NAMD populates a data structure (array of Lphost) for each lone pair in which it stores four atom indices, the first for the lone pair itself, and the next two or three for its support atoms, and then the three floating point parameters labeled distance, angle, and dihedral, used for repositioning the lone pair.
The PSF section NUMLP is read in two loops, the first reading the parameters and whether it is a lone pair with 2 or 3 support atoms, and the second loop to read in the atom indices. It does use that integer in the second column as a loop consistency check.
Now, in terms of routines that redistribute forces from lone pairs and reposition a lone pair based on its support atoms, NAMD does have special case versions just for the SWM4 water molecules. So for a pure SWM4 water simulation, NAMD might not be directly using this array of Lphost (or maybe just using the first entry found for all subsequent water molecules), but it does at least store this data from the PSF.
-- David J. Hardy, Ph.D. Beckman Institute University of Illinois at Urbana-Champaign 405 N. Mathews Ave., Urbana, IL 61801 dhardy_at_ks.uiuc.edu, http://www.ks.uiuc.edu/~dhardy/ > On Sep 27, 2018, at 9:22 PM, Brian Radak <brian.radak_at_gmail.com> wrote: > > Dave is right, but I was more arguing that those entries possibly don't actually matter because internally the hosts are inferred/imposed by the specialized routines and not taken from the PSF (this is at least true for 4 point models in order to support AMBER topologies, which don't contain lonepair host entries). > > It didn't occur to me, but of course makes perfect sense, that the Drude particle is not in the host list because it has no bearing on the bisector placement. With that in mind, you could check my supposition by arbitrarily changing one of the SWM4 host indices and seeing if NAMD complains. It's possible that it will notice that a set of indices don't match the expected pattern and then raise an error. In retrospect I would prefer to be wrong and have this be the case. > > BKR > > On Thu, Sep 27, 2018, 7:07 PM David Hardy <dhardy_at_ks.uiuc.edu <mailto:dhardy_at_ks.uiuc.edu>> wrote: > Hi Julian, > > Although Brian is correct in his statement regarding NAMD's handling of water molecules, his answer actually does not address the question. > > The answer is that the second column is not a list of atom indices. Rather it is a list that indexes the array of integers that follows the parameters section of NUMLP. > > For the entries in the NUMLP section (I'll restate the lines here): > > 3 1 F -0.240345 0.00000 0.00000 > 3 5 F -0.240345 0.00000 0.00000 > 3 9 F -0.240345 0.00000 0.00000 > 3 13 F -0.240345 0.00000 0.00000 > > the 3 in the first column means that the lone pair has 3 support atoms, and the 1, 5, 9, 13, in the section column refers to the index list that follows this force parameters list. For the index list that follows, position 1 is the index of the first lone pair with indices for its 3 supporting atoms, then position 5 is the index of the second lone pair with indices for its 3 supporting atoms, etc. The 3 real numbers for each entry in NUMLP provide the parameters needed for repositioning a lone pair with respect to its support atoms. > > Best regards, > Dave > > -- > David J. Hardy, Ph.D. > Beckman Institute > University of Illinois at Urbana-Champaign > 405 N. Mathews Ave., Urbana, IL 61801 > dhardy_at_ks.uiuc.edu <mailto:dhardy_at_ks.uiuc.edu>, http://www.ks.uiuc.edu/~dhardy/ <http://www.ks.uiuc.edu/~dhardy/> >> On Sep 27, 2018, at 8:41 AM, Brian Radak <brian.radak_at_gmail.com <mailto:brian.radak_at_gmail.com>> wrote: >> >> Hi Julian, >> >> Without looking back at the code, I don't have an answer which speaks directly to your inquiry. However, I will say that all water models in NAMD (including lone pair models) have specialized routines that often ignore some or all of the LPHOST section of the PSF. Thus, while I suspect there is a logical explanation for that format, it's also possible that it will not exactly match the NAMD internals,which instead use heuristics based on atom masses, etc. >> >> HTH, >> BKR >> >> On Thu, Sep 27, 2018, 3:49 AM Julian Müller <mueller_at_pctc.uni-kiel.de <mailto:mueller_at_pctc.uni-kiel.de>> wrote: >> Hello, >> >> i am trying to set up different solvent boxes using the drude >> polarizable SWM4-NDP water model. I downloaded the example files for 216 >> SWM4-NDP water from https://www.ks.uiuc.edu/Research/Drude <https://www.ks.uiuc.edu/Research/Drude>. I am a >> little puzzled by the !NUMLP NUMLPH section, especially the exact >> meaning of the second entry in the following lines: >> >> 3 1 F -0.240345 0.00000 0.00000 >> 3 5 F -0.240345 0.00000 0.00000 >> 3 9 F -0.240345 0.00000 0.00000 >> 3 13 F -0.240345 0.00000 0.00000 >> >> Searching available documentation it seems this number identifies the >> host atom the lone pair is attached to. Therefore i would intuitively >> expect the numbers to be 1, 6, 11 (i.e. the atom IDs of the Oxygen >> atoms) and ascending from there up to the total number of atoms in the >> system, since the number of atoms per SWM4-NDP water is 5 including the >> drude particle. Is there a reason that this number is incremented by 4 >> per lone pair and not by 5? Am i interpreting the line incorrectly or is >> there maybe an error with this .psf file due to the conversion from a >> four atom TIP4P model? Any help here would be greatly appreciated. >> >> best regards >> >> Julian Müller >> >> -- >> Dr. Julian Müller >> Institute for Physical Chemistry >> Christian-Albrechts-Universität zu Kiel >> Max-Eyth Strasse 2 >> 24118 Kiel >> >> phone: 0049 431 880-5820 >> e-mail: mueller_at_pctc.uni-kiel.de <mailto:mueller_at_pctc.uni-kiel.de> >> >
This archive was generated by hypermail 2.1.6 : Wed Dec 11 2019 - 23:20:14 CST