normal Possibility of partial charges in Martini

  • malay
  • malay's Avatar Topic Author
  • Visitor
5 years 10 months ago #7653 by malay
I have a small molecule which has an overall +2 charge. I study the interaction of this molecule with a membrane (DPPE+DPPG). In atomistic model the atoms in the small molecule had partial charges which add up to +2 overall charge. But in Coarse Grained (CG) model I have not specified any partial charges, but instead assigned a charge +1 to two bead of the molecule (+2 charge overall). However, during MD I observe that the interaction between the phosphate bead (-1) of lipid molecule and the positive bead of the molecule is so strong that the molecule does not leave the surface of membrane, even in biased simulations. What I feel is that the Coulombic interaction between absolute charges are very strong. One thing that was not clear to me from the literature is if Martin supports partial charges or it only uses integral charges. Is there any way to specify the partial charges or any way to solve it? Thanking you

Please Log in or Create an account to join the conversation.

More
5 years 10 months ago - 5 years 10 months ago #7667 by bart
The martini forcefield comes with a set of predefined beads and advisory charges to use with them. The strength of charge in an aquatic environment should be roughly right, though the strength of charges in an aliphatic environment are not strong enough in martini due to a uniform dampening of the electrostatic charges based on distance. The dielectrostatic constant has to be set roughly to the screening value of water due to the lack of partial charges and therefore shielding of charge by water and many other molecules). Conclusively:

1) I doubt your charge charge interactions are too strong due to the Coulombic component.

2) Partial charges are possible but should not be used. Part of the electrostatic behaviour is added to the bead type (handled by the LJ). For more information take a look at the tutorial for parameterizing a new molecule.

3) Maybe you could share the small molecule with us so we can take a look at your bead types and mapping. Overmapped regions are known to cause strong interactions.
Last edit: 5 years 10 months ago by bart.

Please Log in or Create an account to join the conversation.

  • malay
  • malay's Avatar Topic Author
  • Visitor
5 years 10 months ago #7670 by malay
Replied by malay on topic Possibility of partial charges in Martini
Thank you very much for your reply.

(1) I didn't understand what does it mean by "overmapped region" ? And how can I avoid it?
(2) Below is the CG model of my molecule:

[moleculetype]
; molname      nrexcl
  DR1f          1

[atoms]
; id    type    resnr   residu  atom    cgnr    charge
   1    C2       1      DR1f    C2A      1      0
   2    C2       1      DR1f    C2B      2      0
   3    P3       1      DR1f    AM1      3      0
   4    Q0       1      DR1f    NH1      4      1.0
   5    C1       1      DR1f    C1A      5      0
   6    C1       1      DR1f    C1B      6      0
   7    C1       1      DR1f    C1C      7      0
   8    Q0       1      DR1f    NH2      8      1.0
   9    P3       1      DR1f    AM2      9      0
  10    C2       1      DR1f    C2C     10      0
  11    C2       1      DR1f    C2D     11      0

[bonds]
;  i  j         funct   length  force.c.
   1  2         1       0.26    1250
   2  3         1       0.41    1250
   3  4         1       0.32    1250
   4  5         1       0.36    1250
   5  6         1       0.51    1250
   6  7         1       0.51    1250
   7  8         1       0.36    1250
   8  9         1       0.32    1250
   9 10         1       0.38    1250
  10 11         1       0.26    1250

[angles]
;  i  j  k      funct   angle   force.c.
   1  2  3      2       172.0   25.0
   2  3  4      2       162.0   25.0
   3  4  5      2       156.0   25.0
   4  5  6      2       173.0   25.0
   5  6  7      2       180.0   25.0
   6  7  8      2       173.0   25.0
   7  8  9      2       155.0   25.0
   8  9 10      2       148.0   25.0
   9 10 11      2       167.0   25.0

Please Log in or Create an account to join the conversation.

More
5 years 10 months ago #7672 by Pim
An overmapped region would mean that you have too many beads close together, for example when you use less than 4 heavy atoms per bead, or sometimes it happens for cyclic compounds.

However, Bart is right. Usually, including for membrane simulations, we use a dielectric constant of eps = 15 for the whole simulation box. This defines the screening of electrostatic charges over distance. This works well for charges in Martini water. However, in an apolar membrane environment, the dielectric constant should actually be lower (~2 for alkanes), but it is very difficult to work with inhomogeneous epsilons. We take our losses there, but that means that the electrostatic interaction is actually too weak by (naively) a factor of 15/2. The real truth probably lies in the middle, the surface of the membrane is obviously charged and not like an alkane. Anyway, that's why Bart suggested that maybe another interaction is too strong, like the LJ interaction between beads that are very close together, a known problem.

What molecule is your topology supposed to represent?

Please Log in or Create an account to join the conversation.

More
5 years 5 months ago #7857 by Pim
Hello,
I'm a little confused by your question, polyethylene oxide does not have charges at the martini level.
You can read all about it at doi.org/10.1021/acs.jpcb.8b04760 .

Please Log in or Create an account to join the conversation.

Time to create page: 0.106 seconds