Dear David and Thomas,
Thanks very much for the advice. As both of you have described, setting a small vdW radius (0.3) and epsilon (0.05) for the phosphate hydrogen appears to have resolved the problems with the instabilities I was observing; I must confess I have not yet tried stiffening the P-O-H angle, but suspect this too will yield the same result.
Will certainly look to validate this approach as I run more simulations down the line.
Thanks again,
Harvey
________________________________________________
Dr Harvey J. A. Dale
Research Fellow
Royal Commission for the Exhibition of 1851
John Henry Coates Research Fellow
Emmanuel College | Cambridge | CB2 3AP | hd487.cam.ac.uk
Postdoctoral Researcher
MRC Lab of Molecular Biology | Cambridge | CB2 0QH
_______________________________________________
-----Original Message-----
From: David A Case via AMBER <amber.ambermd.org>
Sent: Wednesday, March 22, 2023 3:07 PM
To: Steinbrecher, Thomas <thomas.steinbrecher.roche.com>; AMBER Mailing List <amber.ambermd.org>
Subject: Re: [AMBER] [Sender Not Verified] Re: Issues with simulating 5'-phosphorylated RNA strands
On Wed, Mar 22, 2023, Steinbrecher, Thomas via AMBER wrote:
>
>my experience here may be outdated, but the stability issues you
>mention are (or were some years ago) typical for MD simulations of
>protonated phosphates. As you describe, the imbalance between
>electrostatics and vdW repulsion is the issue here. A normal hydroxyl H
>does not need a LJ center, because it is sufficiently protected by both
>the O-LJ volume plus, and this is key, the relatively stiff angle
>bending in e.g. C-C-O groups, so the H is kept away from the heavy
>atoms three bonds away. In contrast, a phosphate has O-P-O angle terms
>which are softer plus relatively high charges on both H and O, so the
>collapse you describe between atoms 1 and 4 in a O-P-O-H substructure
>can occur. The ad hoc solution of giving the H non-zero vdW parameters
>will make the simulation stable (most likely) but in principle you'd be
>on the hook to show that this still describes the phosphate group well
>in the environment you simulate it in. The cowardly solution would be
>to claim that the phosphate has pK around 7 anyway and in the presence
>of all those nucleic acid associated metal cations, you may as well simulate it deprotonated.
Thomas' post seems spot-on to me. One other thing that I remember doing is to increase the force constant for the P-O-H angle in the monophosphate:
this will help keep the proton from finding a neighboring hydrogen. I'd double the force constant -- having a slightly-too-stiff angle is not really going to do anything bad to your sampling.
Let us know if either of both of these ideas (non-zero LJ [try the value for hydrogens bonded to nitrogen] or stiffer angle) help. We should update the parameters so that others don't see the problems you are having.
...thanks for the detailed reports....dac
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
https://eur03.safelinks.protection.outlook.com/?url=http%3A%2F%2Flists.ambermd.org%2Fmailman%2Flistinfo%2Famber&data=05%7C01%7Chd487%40universityofcambridgecloud.onmicrosoft.com%7Cc99e0f734ccc43d8eced08db2ae7323f%7C49a50445bdfa4b79ade3547b4f3986e9%7C1%7C0%7C638150944688465646%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=qOakLXLGCiUTGhHHDzZA3Kv0AlrYl9jwRWsmWn3GIS8%3D&reserved=0
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Thu Mar 23 2023 - 08:00:06 PDT
