Previous Thread
Next Thread
Print Thread
Joined: Aug 2009
Posts: 134
L
lqz Offline OP
Forum Member
OP Offline
Forum Member
L
Joined: Aug 2009
Posts: 134
Dear charmm users:

I plan to calculate PMF for a protein dimer translocation on lipid membrane using umbrella sampling in combination with WHAM program. I tried to confine the center of mass of the protein dimer along z-axis at different reference locations (set up by @height )using the following commands, while allowing the protein dimer to move freely in the x and y directions.

cons hmcm force 10.0 weig refz @height sele segid HD1 .or. segid HD2 end

MMFP
GEO sphere RCM -
force @force zref @height droff 0.0 -
sele segid HD1 .or. segid HD2 end
END

I am wondering if above commands can be good enough? Thank you very much.

Joined: Aug 2009
Posts: 134
L
lqz Offline OP
Forum Member
OP Offline
Forum Member
L
Joined: Aug 2009
Posts: 134
Based on my check on the trajectory, the center of mass was fixed, not only the height in the z-direction.

I am not sure if that is OK. If not, then how to set up mmfp so that the dimer still can move in the x and y direction?

Joined: Sep 2003
Posts: 8,605
Likes: 24
rmv Online Content
Forum Member
Online Content
Forum Member
Joined: Sep 2003
Posts: 8,605
Likes: 24
If both HMCM and MMFP restraints were applied simultaneously, that would explain the result; you've restrained the dimer COM to the point 0,0,10 with MMFP GEO SPHERE RCM, as the default values for e.g. XREF are zero if not specified.

It would be best to use just one method, either CONS HMCM alone, or MMFP, but with a PLANAR shape; either would confine the COM to a plane with the given z value.

If you intend to use DOMDEC, you need to use CONS HMCM, as the MMFP restraints are not supported. DOMDEC does support umbrella sampling via the commands described in umbrel.doc and adumb.doc


Rick Venable
computational chemist

Joined: Aug 2009
Posts: 134
L
lqz Offline OP
Forum Member
OP Offline
Forum Member
L
Joined: Aug 2009
Posts: 134
Thanks for the comments. I modified the script as below:

coor stat mass select segid HD1 .or. segid HD2 end

MMFP
GEO plane -
xref ?xave yref ?yave zref ?zave zdir @height -
force @force droff 0.0 sele segid HBD1 .or. segid HBD2 end
END

although it can work, it crashed soon with the following error:

EPHI: WARNING. bent improper torsion angle is

far from minimum for;
IPHI= 10 with deltaPHI= -99.8838 MIN= 0.0000 ATOMS: 57 55 58 59
EPHI: WARNING. bent improper torsion angle is

far from minimum for;
IPHI= 45 with deltaPHI=-139.9396 MIN= 0.0000 ATOMS: 288 286 290 289
UPDECI: Nonbond update at step 1
NBONDA>> Maximum group spatial extent (12A) exceeded.

Joined: Aug 2009
Posts: 134
L
lqz Offline OP
Forum Member
OP Offline
Forum Member
L
Joined: Aug 2009
Posts: 134
I had a typo in above post. The command in charmm is

coor stat mass select segid HD1 .or. segid HD2 end

MMFP
GEO plane -
xref ?xave yref ?yave zref ?zave zdir @height -
force @force droff 0.0 sele segid HD1 .or. segid HD2 end
END


But I got the same error message as shown above.

Joined: Sep 2003
Posts: 8,605
Likes: 24
rmv Online Content
Forum Member
Online Content
Forum Member
Joined: Sep 2003
Posts: 8,605
Likes: 24
Without the RCM keyword (see mmfp.doc), the restraint is attempting to collapse all of the peptide atoms into a plane.


Rick Venable
computational chemist

Joined: Aug 2009
Posts: 134
L
lqz Offline OP
Forum Member
OP Offline
Forum Member
L
Joined: Aug 2009
Posts: 134
Thank you very much for your guidance, Rick.
After using rcm keyword, the simulation no longer crashes. But I have one question about xref, yref and zref. What kind of values should I use in the command line in order to let the molecule movable in the x and y direction but with its height fixed?

When I used ?xave, ?yave, and @height, the molecule did not move at all.

Joined: Sep 2003
Posts: 8,605
Likes: 24
rmv Online Content
Forum Member
Online Content
Forum Member
Joined: Sep 2003
Posts: 8,605
Likes: 24
Based on my reading of mmfp.doc, I believe XDIR etc. should be unit vector values; I've always used ZDIR 1.0 for bilayer based planar restraints.

Peptides in a membrane may not have much lateral motion on the timescale of a short MD simulation, e.g. less than a few hundred nanoseconds. You should be comparing to a control with no restraints over the same time period.

I've often found it useful to set up small, simple simulation systems in order to test my understanding of restraint behavior, esp. with MMFP. For instance, it should be easy to determine if an ion in water is being restrained to a point or a plane.


Rick Venable
computational chemist

Joined: Aug 2009
Posts: 134
L
lqz Offline OP
Forum Member
OP Offline
Forum Member
L
Joined: Aug 2009
Posts: 134
Thank you very much for the information. I am doing the long simulation tests now.


Moderated by  BRBrooks, lennart, rmv 

Link Copied to Clipboard
Powered by UBB.threads™ PHP Forum Software 7.7.5
(Release build 20201027)
Responsive Width:

PHP: 7.3.31-1~deb10u1 Page Time: 0.013s Queries: 32 (0.009s) Memory: 0.7705 MB (Peak: 0.8421 MB) Data Comp: Off Server Time: 2022-09-28 00:19:30 UTC
Valid HTML 5 and Valid CSS