[gmx-users] Parametrisation of the cyclic nucleotides in Gromos force fields
James Starlight
jmsstarlight at gmail.com
Fri Dec 7 16:41:47 CET 2012
Today I've tried to simulate complexes of my protein with the cyclic
GMP parametrized by ATB's. (below the recent parametrisation for
charges of that molecule done by am1 algorithm)
[ moleculetype ]
; Name nrexcl
_N4P 3
[ atoms ]
; nr type resnr resid atom cgnr charge mass total_charge
1 NT 1 _N4P N2 1 -0.848 14.0067
2 H 1 _N4P H22 1 0.424 1.0080
3 H 1 _N4P H21 1 0.424 1.0080 ; 0.000
4 NR 1 _N4P N1 2 -0.684 14.0067
5 H 1 _N4P H1 2 0.437 1.0080
6 C 1 _N4P C2 2 0.779 12.0110
7 NR 1 _N4P N3 2 -0.678 14.0067
8 C 1 _N4P C4 2 0.440 12.0110
9 NR 1 _N4P N9 2 -0.294 14.0067 ; 0.000
10 C 1 _N4P C8 3 0.272 12.0110
11 HC 1 _N4P H01 3 0.141 1.0080
12 O 1 _N4P O6 3 -0.555 15.9994
13 C 1 _N4P C6 3 0.608 12.0110
14 C 1 _N4P C5 3 0.101 12.0110
15 NR 1 _N4P N7 3 -0.567 14.0067 ; 0.000
16 OE 1 _N4P O4* 4 -0.451 15.9994
17 CH1 1 _N4P C1* 4 0.451 13.0190 ; 0.000
18 OA 1 _N4P O5* 5 -0.394 15.9994
19 P 1 _N4P PAQ 5 0.974 30.9738
20 OM 1 _N4P OAR 5 -0.616 15.9994
21 OA 1 _N4P O3* 5 -0.390 15.9994
22 OA 1 _N4P OAS 5 -0.584 15.9994
23 H 1 _N4P H03 5 0.491 1.0080
24 CH2 1 _N4P C5* 5 0.281 14.0270
25 CH1 1 _N4P C4* 5 0.238 13.0190 ; 0.000
26 CH1 1 _N4P C3* 6 0.000 13.0190 ; 0.000
27 CH1 1 _N4P C2* 7 0.181 13.0190
28 OA 1 _N4P O2* 7 -0.645 15.9994
29 H 1 _N4P H8M 7 0.464 1.0080 ; 0.000
In all cases my system have always been crushed after equilibration
run ( during that phase with the posres applied on both protein as
well as ligand the system was stabile) with the errors like
Step 19, time 0.038 (ps) LINCS WARNING
relative constraint deviation after LINCS:
rms 1.203662, max 11.750027 (between atoms 4289 and 4287)
bonds that rotated more than 30 degrees:
atom 1 atom 2 angle previous, current, constraint length
4288 4287 89.5 0.2277 1.1291 0.1000
4292 4287 90.5 0.1304 0.9898 0.1350
4289 4287 90.4 0.1056 1.2750 0.1000
4318 4303 30.8 0.1531 0.1816 0.1530
4304 4303 45.2 0.1097 0.1531 0.1090
4309 4310 90.0 0.2629 0.3419 0.1000
4299 4290 84.2 0.1404 0.1641 0.1400
4292 4290 89.4 0.1362 0.5025 0.1380
4291 4290 89.2 0.1563 0.2142 0.1000
4293 4292 89.1 0.1252 0.5682 0.1320
4294 4293 78.1 0.1387 0.3906 0.1350
4300 4294 92.2 0.1385 0.4639 0.1390
4295 4294 91.1 0.1365 0.5711 0.1380
4303 4295 43.4 0.1482 0.1954 0.1470
4296 4295 36.6 0.1408 0.1593 0.1400
4301 4296 89.9 0.1323 0.5162 0.1320
4297 4296 32.5 0.1089 0.1343 0.1090
4299 4298 87.0 0.1230 0.2758 0.1230
4300 4299 89.8 0.1431 0.5922 0.1430
4301 4300 79.0 0.1381 0.1928 0.1380
Fatal error:
1 particles communicated to PME node 26 are more than 2/3 times the
cut-off out of the domain decomposition cell of their charge group in
dimension y.
This usually means that your system is not well equilibrated.
For more information and tips for troubleshooting, please check the GROMACS
After that my system have been crashed and than produced many
step*.pdb files in the work dirr.
It's intresting that with the prodrg topology of that mollecule ( with
worst charge distribution) I've never such problems. Might that error
be due to the wrong geometry parametrisation (e.g incorect dihedrals)
of the cGMP made by ATB ?
James
2012/12/6 James Starlight <jmsstarlight at gmail.com>:
> Justin,
>
> Thanks again for explanation.
>
>
> It's interesting that above parametrization made by ATB have cased the
> system to crash within first ps of modeling ;) (On the contrarythe
> system with the ligand made by prodrg have been very stable during
> 100ns). I ve tried to re-parametrized my molecule by another algorithm
> implemented in ATB ( am1 instead of pm3 which was used in the crashed
> simulation).
>
>
> James
>
> 2012/12/6, Justin Lemkul <jalemkul at vt.edu>:
>>
>>
>> On 12/6/12 2:39 AM, James Starlight wrote:
>>> Justin,
>>>
>>> Could you provide me with the example of the server where I could
>>> obtain Gromac's itp topologies for the charmm ff? I know many such
>>> servers which could be useful only for preparation systems for NAMD
>>> program.
>>>
>>
>> Google "CHARMM ligand topology in Gromacs" (without the quotes) - the first
>>
>> result is what you're looking for.
>>
>>>
>>> By the way recently I've made parametrization of my cGMP molecule by
>>> means of ATB server. In the below example you can see that the charge
>>> distribution is differs from the PRODRG example of that molecule which
>>> I've posted yesterday. Does that charge distribution more suitable for
>>> the 54a force field?
>>>
>>
>> Given that PRODRG generally produces very bad charges, just about anything
>> is
>> better ;)
>>
>> Nucleotide parameters already exist in 54A7, I don't see why you necessarily
>>
>> have to create them from scratch. In general, these charges look pretty
>> good,
>> but note that DGUA already exists and can describe most of your molecule
>> already. The cyclic part is the only trick, but the nucleobase parameters
>> should be the same in cGMP and DGUA, given the nature of Gromos96
>> parameterization.
>>
>> -Justin
>>
>>> [ atoms ]
>>> ; nr type resnr resid atom cgnr charge mass total_charge
>>> 1 NT 1 _N4H N2 1 -0.832 14.0067
>>> 2 H 1 _N4H H22 1 0.416 1.0080
>>> 3 H 1 _N4H H21 1 0.416 1.0080 ; 0.000
>>> 4 NR 1 _N4H N1 2 -0.715 14.0067
>>> 5 H 1 _N4H H1 2 0.427 1.0080
>>> 6 C 1 _N4H C2 2 0.775 12.0110
>>> 7 NR 1 _N4H N3 2 -0.691 14.0067
>>> 8 C 1 _N4H C4 2 0.431 12.0110
>>> 9 NR 1 _N4H N9 2 -0.227 14.0067 ; 0.000
>>> 10 C 1 _N4H C8 3 0.220 12.0110
>>> 11 HC 1 _N4H H01 3 0.162 1.0080
>>> 12 O 1 _N4H O6 3 -0.556 15.9994
>>> 13 C 1 _N4H C6 3 0.669 12.0110
>>> 14 C 1 _N4H C5 3 0.026 12.0110
>>> 15 NR 1 _N4H N7 3 -0.521 14.0067 ; 0.000
>>> 16 OE 1 _N4H O4* 4 -0.429 15.9994
>>> 17 CH1 1 _N4H C1* 4 0.429 13.0190 ; 0.000
>>> 18 CH1 1 _N4H C4* 5 0.000 13.0190 ; 0.000
>>> 19 OA 1 _N4H O5* 6 -0.422 15.9994
>>> 20 P 1 _N4H PAQ 6 0.971 30.9738
>>> 21 OM 1 _N4H OAR 6 -0.613 15.9994
>>> 22 OA 1 _N4H O3* 6 -0.382 15.9994
>>> 23 OA 1 _N4H OAS 6 -0.617 15.9994
>>> 24 H 1 _N4H H03 6 0.497 1.0080
>>> 25 CH2 1 _N4H C5* 6 0.319 14.0270
>>> 26 CH1 1 _N4H C3* 6 0.247 13.0190 ; -0.000
>>> 27 CH1 1 _N4H C2* 7 0.200 13.0190
>>> 28 OA 1 _N4H O2* 7 -0.614 15.9994
>>> 29 H 1 _N4H H8M 7 0.414 1.0080 ; 0.000
>>>
>>>
>>>
>>> James
>>>
>>> 2012/12/5 Justin Lemkul <jalemkul at vt.edu>:
>>>>
>>>>
>>>> On 12/5/12 1:39 PM, James Starlight wrote:
>>>>>
>>>>> Justin,
>>>>>
>>>>> Indeed the force field is the 54a7 ( modiffied version of the 54a6).
>>>>>
>>>>> The main reason of using GROMOS ff in that case was the topology of
>>>>> ligands which could be easily created by means of prodrg or ATB. On
>>>>> other hand I've never worked with the protein-ligand complexes in
>>>>> charmm ff for instance.
>>>>>
>>>>
>>>> Well, you get out what you put in. A recent paper
>>>> (dx.doi.org/10.1002/jcc.23055) showed that Gromos force fields performed
>>>> very poorly for simulating nucleic acids. There are others, but that's
>>>> just
>>>> a recent one. If you're choosing a force field because it makes life
>>>> easy,
>>>> be prepared to defend your results if they are of poor quality or defend
>>>> a
>>>> lot of wasted time while you re-do the simulations :)
>>>>
>>>> There are servers that produce CHARMM topologies and other programs that
>>>> will convert AMBER topologies into Gromacs format as well. I would
>>>> suggest
>>>> you evaluate all the options available.
>>>>
>>>>
>>>>> By the way is there any suitable builing blocks (implemented in the
>>>>> rtp enties of the gromos ff) which could be used for charge
>>>>> assignment?
>>>>>
>>>>
>>>> That depends on the functional group. If it's also found in proteins,
>>>> yes.
>>>> If not, then maybe but probably not.
>>>>
>>>>
>>>> -Justin
>>>> --
>>>> ========================================
>>>>
>>>> Justin A. Lemkul, Ph.D.
>>>> Research Scientist
>>>> Department of Biochemistry
>>>> Virginia Tech
>>>> Blacksburg, VA
>>>> jalemkul[at]vt.edu | (540) 231-9080
>>>> http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin
>>>>
>>>> ========================================
>>>> --
>>>> gmx-users mailing list gmx-users at gromacs.org
>>>> http://lists.gromacs.org/mailman/listinfo/gmx-users
>>>> * Please search the archive at
>>>> http://www.gromacs.org/Support/Mailing_Lists/Search before posting!
>>>> * Please don't post (un)subscribe requests to the list. Use the www
>>>> interface or send it to gmx-users-request at gromacs.org.
>>>> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>>
>> --
>> ========================================
>>
>> Justin A. Lemkul, Ph.D.
>> Research Scientist
>> Department of Biochemistry
>> Virginia Tech
>> Blacksburg, VA
>> jalemkul[at]vt.edu | (540) 231-9080
>> http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin
>>
>> ========================================
>> --
>> gmx-users mailing list gmx-users at gromacs.org
>> http://lists.gromacs.org/mailman/listinfo/gmx-users
>> * Please search the archive at
>> http://www.gromacs.org/Support/Mailing_Lists/Search before posting!
>> * Please don't post (un)subscribe requests to the list. Use the
>> www interface or send it to gmx-users-request at gromacs.org.
>> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>>
More information about the gromacs.org_gmx-users
mailing list