[gmx-users] Re: Carbohydrate simulation: problem with the terminal atoms
Justin A. Lemkul
jalemkul at vt.edu
Mon Apr 2 15:58:11 CEST 2012
khuchtumur wrote:
> Hi Justin,
>
> Thank you for writing this clarification. I'm still having a bit of a
> problem with the anomeric -OH group giving LINCS warnings during
> equilibration though.
>
> I am trying to parametrize a beta-lactose molecule in g53a6. Following are
> the steps I followed.
> 1) Draw lactose in Avogadro
> 2) Use PRODRG to get ITP template
> 3) Change the charges and charge groups in accordance with Oostenbrink et al
> 2004 and Lins et al 2005.
> 4) Change the bonds and angles in accordance with Oostenbrink et al, which
> matches Lins et al
> 5) Check the improper dihedrals for chiral centers
> 6) Set up the proper dihedrals using Lins et al.
> 7) Set the exclusion you mentioned here (between O5 and H1)
> 8) Change the C-C-C-O5 specification from Lins et al to CCCO1 specification
> as in your post (explained below)
>
> This gives the following charges and dihedrals:
>
> [ atoms ]
> ; nr type resnr resid atom cgnr charge mass
> 1 OA 1 LAT O2 1 -0.642 15.9994
> 2 H 1 LAT H2 1 0.410 1.0080
> 3 CH1 1 LAT C2 1 0.232 13.0190
> 4 CH1 1 LAT C3 2 0.232 13.0190
> 5 OA 1 LAT O3 2 -0.642 15.9994
> 6 H 1 LAT H3 2 0.410 1.0080
> 7 CH1 1 LAT C4 3 0.232 13.0190
> 8 OA 1 LAT O4 3 -0.642 15.9994
> 9 H 1 LAT H4 3 0.410 1.0080
> 10 CH1 1 LAT C5 4 0.376 13.0190
> 11 CH2 1 LAT C6 5 0.232 14.0270
> 12 OA 1 LAT O6 5 -0.642 15.9994
> 13 H 1 LAT H6 5 0.410 1.0080
> 14 OA 1 LAT O5 4 -0.480 15.9994
> 15 CH1 1 LAT C1 4 0.232 13.0190
> 16 OA 1 LAT O1 4 -0.360 15.9994
> 17 CH1 1 LAT C4' 4 0.232 13.0190 ;'
> 18 CH1 1 LAT C3' 6 0.232 13.0190 ;'
> 19 OA 1 LAT O3' 6 -0.642 15.9994 ;'
> 20 H 1 LAT H3' 6 0.410 1.0080
> 21 CH1 1 LAT C2' 7 0.232 13.0190 ;'
> 22 OA 1 LAT O2' 7 -0.642 15.9994 ;'
> 23 H 1 LAT H2' 7 0.410 1.0080
> 24 CH1 1 LAT C1' 8 0.232 13.0190 ;'
> 25 OA 1 LAT O1' 8 -0.538 15.9994 ;'
> 26 H 1 LAT H1' 8 0.410 1.0080
> 27 OA 1 LAT O5' 8 -0.480 15.9994 ;'
> 28 CH1 1 LAT C5' 8 0.376 13.0190 ;'
> 29 CH2 1 LAT C6' 9 0.232 14.0270 ;'
> 30 OA 1 LAT O6' 9 -0.642 15.9994 ;'
> 31 H 1 LAT H6' 9 0.410 1.0080
>
> [ dihedrals ]
>
> 2 1 3 15 1 gd_30
> ; 2 1 3 4 1 gd_30
> 1 3 15 16 1 gd_18
> ; 1 3 15 14 1 gd_18
> 1 3 4 5 1 gd_18
> 1 3 4 7 1 gd_17
> ; 3 15 16 17 1 gd_29
> 3 15 14 10 1 gd_29
> 3 4 5 6 1 gd_30
> 3 4 7 8 1 gd_17
> 3 4 7 10 1 gd_34
> 15 3 4 7 1 gd_34
> 15 3 4 5 1 gd_17
> 16 15 3 4 1 gd_17
> 16 15 3 4 1 gd_34
> 14 15 3 4 1 gd_17
> ; 14 15 3 4 1 gd_34
> ; 6 5 4 7 1 gd_30
> 5 4 7 8 1 gd_18
> 5 4 7 10 1 gd_17
> 4 7 8 9 1 gd_30
> 4 7 10 11 1 gd_34
> 4 7 10 14 1 gd_17
> ; 4 7 10 14 1 gd_34
> ; 9 8 7 10 1 gd_30
> 8 7 10 11 1 gd_17
> ; 8 7 10 14 1 gd_18
> ; 7 10 11 12 1 gd_1
> 7 10 14 15 1 gd_29
> 13 12 11 10 1 gd_30
> 12 11 10 14 1 gd_3
> 12 11 10 14 1 gd_35
> ; 11 10 14 15 1 gd_29
> 14 15 16 17 1 gd_2
> 14 15 16 17 1 gd_32
> 15 16 17 18 1 gd_30
> ; 15 16 17 28 1 gd_29
> 16 17 18 19 1 gd_18
> 16 17 18 21 1 gd_17
> 16 17 28 29 1 gd_17
> ; 16 17 28 27 1 gd_18
> ; 17 18 19 20 1 gd_30
> 17 18 21 22 1 gd_17
> 17 18 21 24 1 gd_34
> ; 17 28 29 30 1 gd_17
> 17 28 27 24 1 gd_29
> 19 18 17 28 1 gd_17
> 18 17 28 29 1 gd_34
> 18 17 28 27 1 gd_17
> ; 18 17 28 27 1 gd_34
> 21 18 17 28 1 gd_34
> 20 19 18 21 1 gd_30
> 19 18 21 22 1 gd_18
> 19 18 21 24 1 gd_17
> ; 18 21 22 23 1 gd_30
> 18 21 24 25 1 gd_17
> 18 21 24 25 1 gd_34
> 18 21 24 27 1 gd_17
> ; 18 21 24 27 1 gd_34
> 23 22 21 24 1 gd_30
> 22 21 24 25 1 gd_18
> ; 22 21 24 27 1 gd_18
> ; 21 24 25 26 1 gd_30
> 21 24 27 28 1 gd_29
> 26 25 24 27 1 gd_2
> 26 25 24 27 1 gd_32
> ; 26 25 24 27 1 180.0 20.0 1 ; restrain?
> http://compbio.biosci.uq.edu.au/atb/download.py?molid=2002
> ; 24 27 28 29 1 gd_29
> 27 28 29 30 1 gd_5
> 27 28 29 30 1 gd_37
> 28 29 30 31 1 gd_30
>
> ; 28 27 24 25 1 180.0 10.0 1
>
> ; Picked from Autieri et al J. Chem. Phys. 2010
> ; 10 7 4 5 1 180.0 2.4 2
> ; 15 3 4 5 1 180.0 2.4 2
> ; 4 3 15 16 1 180.0 0.5 2
> ; 11 10 14 15 1 180.0 0.5 2
> ; 7 4 3 1 1 180.0 0.5 2
> ; 28 17 18 19 1 180.0 2.4 2
> ; 24 21 18 19 1 180.0 2.4 2
> ; 18 21 24 25 1 180.0 0.5 2
> ; 29 28 27 24 1 180.0 0.5 2
> ; 17 18 21 22 1 180.0 0.5 2
> [ exclusions ]
> 26 27
> #The commented dihedrals are things I was trying#
>
> The molecule starts giving off LINCS warnings during NVT or NPT
> equilibration with a protein that I'm interested in. I have run just the
> molecule alone to check if it's a problem with the complex, but the molecule
> in water also gives the same error. The error is on the anomeric hydroxy
> group on the reducing end of the molecule, namely glucose O1-H1. My
> questions are:
>
> 1) In this post, you have set both gd_17 and gd_34 for C3-C2-C1-O1, while
> Lins et al 2005 have set up gd_17 and gd_34 for C3-C4-C5-O5 or C3-C2-C1-O5.
> My assumption is that Lins et al are treating the O5 as if it's a carbon
> within the sugar ring. My chemistry is not good, but the O5 here is quite
> different from that of the other oxygens and could require extra attention.
In the context of dihedrals, that's doubtful. Maybe on some quantum level there
is an important distinction, but not likely here. Take that up with the authors
of the force field itself ;)
> I'm not sure why the O1 is treated like this in your post. Is it because
> it's in a polymer? Also I noticed that you did not have the extra gd_34
> definition for the dihedrals involving the O5, was it left out on purpose?
The building blocks I made were derived directly from what was available in the
53A6 force field. As I recall, everything else matched aside from what I added.
I'll leave that as an exercise to you to determine. I did simple diagnostic
simulations to verify stability, nothing more.
> 2) I realize that there are different ways to set the dihedrals for complete
> coverage of the molecule (i. e. C2-C3-O3-H3/C4-C3-O3-H3). Is there any down
> side to defining all of the possible dihedrals redundantly?
If you define multiple dihedrals for a given bond, they are summed together to
give the final term. Unless you have specific reason to do that, don't.
> 3) I realize that the topology you've shown here is a template, but have you
> simulated this template successfully?
>
Yes.
> I have run 1'-deoxy-1'-methoxylactose (replacing the glucose OH with a
> terminal methoxy {according to the specifications from Lins et al} ) and it
> doesn't give LINCS warnings. I haven't made any quality checks, but this
> actually runs. So this simple change makes me think that the problem is not
> anything else but this anomeric OH termination.
> Other notes: running gmx v4.5.5, spc water, em run until Fmax <100
>
> I'm sorry if I rambled on. This is the first molecule I'm parameterizing.
> Please let me know if I made any careless mistakes or if there are other
> lit. articles I should have taken into account.
>
Why not try using ATB to build the topology?
http://compbio.biosci.uq.edu.au/atb/
I think the results should be vastly more reliable than trying to
reverse-engineer a faulty PRODRG topology. They also have a reasonably large
database of building block molecules with which you can check your work.
-Justin
--
========================================
Justin A. Lemkul
Ph.D. Candidate
ICTAS Doctoral Scholar
MILES-IGERT Trainee
Department of Biochemistry
Virginia Tech
Blacksburg, VA
jalemkul[at]vt.edu | (540) 231-9080
http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin
========================================
More information about the gromacs.org_gmx-users
mailing list