angle projected onto a plane

[James Krieger]

Hi,

The documentation for the TORSION CV says it can be used for an angle projected onto a plane. I was wondering whether I can define this plane using something other than an orthogonal vector between two atoms e.g. using three atoms.

Thanks
James

[Giovanni Bussi]

We did something similar in the past creating a virtual atom with GHOST based on three atoms, then taking the distance between one of the three atoms and the ghost atom. In our case we used it to compute ANGLEs between planes, but I guess you can do the same for a TORSION

Giovanni

--
You received this message because you are subscribed to the Google Groups "PLUMED users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to plumed-users...@googlegroups.com.
To post to this group, send email to plumed...@googlegroups.com.
Visit this group at http://groups.google.com/group/plumed-users.
To view this discussion on the web visit https://groups.google.com/d/msgid/plumed-users/126db653-148b-4dd6-ad68-9aa2d33be78d%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[James Krieger]

I suppose I'd like to place the GHOST such that it forms an axis with one of the atoms that is perpendicular to the plane I wish to define. I'm not sure I understand how to place the GHOST. Are the COORDINATES distances from the three atoms?

Thanks

[Giovanni Bussi]

I think they are coordinates from the first atom, but the best thing is to use DUMPATOMS to double check where the ghost is placed.

To view this discussion on the web visit https://groups.google.com/d/msgid/plumed-users/676ca3d7-af03-4b87-94b9-4b0875fcb105%40googlegroups.com.

[jamesm...@gmail.com]

Ok thanks 

You received this message because you are subscribed to a topic in the Google Groups "PLUMED users" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/plumed-users/GfMqob-aeLE/unsubscribe.
To unsubscribe from this group and all its topics, send an email to plumed-users...@googlegroups.com.

To post to this group, send email to plumed...@googlegroups.com.
Visit this group at http://groups.google.com/group/plumed-users.

To view this discussion on the web visit https://groups.google.com/d/msgid/plumed-users/CAPLm8ZL%3DBasJBf0x0NULW3qTjM69R9xLG2kCK1A%3De69DyDOuDg%40mail.gmail.com.

[James Krieger]

Yes the GHOST is placed relative to the first atom. The second coordinate is perpendicular to the place containing the three atoms.

--
You received this message because you are subscribed to a topic in the Google Groups "PLUMED users" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/plumed-users/GfMqob-aeLE/unsubscribe.
To unsubscribe from this group and all its topics, send an email to plumed-users...@googlegroups.com.

To post to this group, send email to plumed...@googlegroups.com.
Visit this group at http://groups.google.com/group/plumed-users.

To view this discussion on the web visit https://groups.google.com/d/msgid/plumed-users/CAPLm8ZL%3DBasJBf0x0NULW3qTjM69R9xLG2kCK1A%3De69DyDOuDg%40mail.gmail.com.

[James Krieger]

The GHOST atoms moves around all over the place. Am I somehow defining it wrong?

COM ATOMS=com1,com3 LABEL=com13
GHOST ATOMS=com13,com1,com3 COORDINATES=0,0,3 LABEL=g1

Thanks
James

On Friday, June 12, 2015 at 11:56:39 AM UTC+1, James Krieger wrote:

Yes the GHOST is placed relative to the first atom. The second coordinate is perpendicular to the place containing the three atoms.

On Thu, Jun 11, 2015 at 6:46 PM, Giovanni Bussi <giovann...@gmail.com> wrote:

I think they are coordinates from the first atom, but the best thing is to use DUMPATOMS to double check where the ghost is placed.

On Thu, Jun 11, 2015 at 7:41 PM, James Krieger <jamesm...@gmail.com> wrote:

I suppose I'd like to place the GHOST such that it forms an axis with one of the atoms that is perpendicular to the plane I wish to define. I'm not sure I understand how to place the GHOST. Are the COORDINATES distances from the three atoms?

Thanks

On Thursday, June 11, 2015 at 6:13:03 PM UTC+1, Giovanni Bussi wrote:

We did something similar in the past creating a virtual atom with GHOST based on three atoms, then taking the distance between one of the three atoms and the ghost atom. In our case we used it to compute ANGLEs between planes, but I guess you can do the same for a TORSION

Giovanni

On Thu, Jun 11, 2015 at 6:49 PM, James Krieger <jamesm...@gmail.com> wrote:

Hi,

The documentation for the TORSION CV says it can be used for an angle projected onto a plane. I was wondering whether I can define this plane using something other than an orthogonal vector between two atoms e.g. using three atoms.

Thanks
James

--
You received this message because you are subscribed to the Google Groups "PLUMED users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to plumed-users...@googlegroups.com.
To post to this group, send email to plumed...@googlegroups.com.
Visit this group at http://groups.google.com/group/plumed-users.
To view this discussion on the web visit https://groups.google.com/d/msgid/plumed-users/126db653-148b-4dd6-ad68-9aa2d33be78d%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "PLUMED users" group.

To unsubscribe from this group and stop receiving emails from it, send an email to plumed-users+unsubscribe@googlegroups.com.

To post to this group, send email to plumed...@googlegroups.com.
Visit this group at http://groups.google.com/group/plumed-users.

To view this discussion on the web visit https://groups.google.com/d/msgid/plumed-users/676ca3d7-af03-4b87-94b9-4b0875fcb105%40googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to a topic in the Google Groups "PLUMED users" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/plumed-users/GfMqob-aeLE/unsubscribe.

To unsubscribe from this group and all its topics, send an email to plumed-users+unsubscribe@googlegroups.com.

To post to this group, send email to plumed...@googlegroups.com.
Visit this group at http://groups.google.com/group/plumed-users.

[Giovanni Bussi]

Check with DUMPATOMS, dumping also com13,com1,com3.

Notice that the atoms part of a COM should be whole wrt periodic boundaries. This will be done by default in 2.2, but in 2.1 you need to use WHOLEMOLECULES.

Giovanni

To view this discussion on the web visit https://groups.google.com/d/msgid/plumed-users/676ca3d7-af03-4b87-94b9-4b0875fcb105%40googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to a topic in the Google Groups "PLUMED users" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/plumed-users/GfMqob-aeLE/unsubscribe.

To unsubscribe from this group and all its topics, send an email to plumed-users...@googlegroups.com.

To post to this group, send email to plumed...@googlegroups.com.
Visit this group at http://groups.google.com/group/plumed-users.

To view this discussion on the web visit https://groups.google.com/d/msgid/plumed-users/CAPLm8ZL%3DBasJBf0x0NULW3qTjM69R9xLG2kCK1A%3De69DyDOuDg%40mail.gmail.com.

For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "PLUMED users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to plumed-users...@googlegroups.com.
To post to this group, send email to plumed...@googlegroups.com.
Visit this group at http://groups.google.com/group/plumed-users.

To view this discussion on the web visit https://groups.google.com/d/msgid/plumed-users/ed6ae474-727f-422a-86d2-0515576c90c6%40googlegroups.com.

[jamesm...@gmail.com]

Thanks Giovanni,

I am dumping all the COM atoms as well. It looks like their positions are very stable. Also this movement is happening from the start of the run before the groups of atoms contributing to the COMs have a chance to reach the boundaries. Nevertheless I'll try using WHOLEMOLECULES and see if it helps.

Best wishes 

James 

To view this discussion on the web visit https://groups.google.com/d/msgid/plumed-users/CAPLm8Z%2BBAtqzkig27exYDfNEV1y%2BuPhGjb%2BR6-tWR_4SMxnSVA%40mail.gmail.com.

[James Krieger]

That doesn't seem to have helped and it also makes some of the COMs appear in unexpected places (at least one of them outside the boundary).

To unsubscribe from this group and stop receiving emails from it, send an email to plumed-users+unsubscribe@googlegroups.com.

To post to this group, send email to plumed...@googlegroups.com.
Visit this group at http://groups.google.com/group/plumed-users.

To view this discussion on the web visit https://groups.google.com/d/msgid/plumed-users/676ca3d7-af03-4b87-94b9-4b0875fcb105%40googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to a topic in the Google Groups "PLUMED users" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/plumed-users/GfMqob-aeLE/unsubscribe.

To unsubscribe from this group and all its topics, send an email to plumed-users+unsubscribe@googlegroups.com.

To post to this group, send email to plumed...@googlegroups.com.
Visit this group at http://groups.google.com/group/plumed-users.

To view this discussion on the web visit https://groups.google.com/d/msgid/plumed-users/CAPLm8ZL%3DBasJBf0x0NULW3qTjM69R9xLG2kCK1A%3De69DyDOuDg%40mail.gmail.com.

For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "PLUMED users" group.

To unsubscribe from this group and stop receiving emails from it, send an email to plumed-users+unsubscribe@googlegroups.com.

To post to this group, send email to plumed...@googlegroups.com.
Visit this group at http://groups.google.com/group/plumed-users.

To view this discussion on the web visit https://groups.google.com/d/msgid/plumed-users/ed6ae474-727f-422a-86d2-0515576c90c6%40googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to a topic in the Google Groups "PLUMED users" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/plumed-users/GfMqob-aeLE/unsubscribe.

To unsubscribe from this group and all its topics, send an email to plumed-users+unsubscribe@googlegroups.com.

[James Krieger]

ok I have set WHOLEMOLECULES appropriately for getting all virtual atoms (COMs and the GHOST) stable. I still have a problem which probably related: one of my COMs makes periodic boundary crossings. I think I'll try setting the whole protein as one ENTITY for WHOLEMOLECULES. My current plumed.dat contains the following.

WHOLEMOLECULES STRIDE=1 ENTITY0=1-1880,4172-6097 ENTITY3=6556-8431,10729-12654 ENTITY1=1881-4171,6098-6555 ENTITY2=8432-10728,12655-13113
# set up one variable for opening angle. A torsion is used to project the angle back into the right plane. The ghost atom is placed in order to define said plane.
COM ATOMS=1-1880,4172-6097 LABEL=com1
COM ATOMS=1881-4171,6098-6555 LABEL=com2
COM ATOMS=6556-8436,10729-12654 LABEL=com3
COM ATOMS=8437-10728,12655-13113 LABEL=com4
COM ATOMS=1-1880,4172-6097,6556-8431,10729-12654 LABEL=com13

GHOST ATOMS=com13,com1,com3 COORDINATES=0,0,3 LABEL=g1
t: TORSION VECTOR1=com2,com13 AXIS=com13,g1 VECTOR2=com13,com4

DUMPATOMS STRIDE=10 FILE=file.gro ATOMS=com1,com2,com3,com4,com13,g1

# Activate metadynamics in angle/torsion t
# depositing a Gaussian every 500 time steps,
# with height equal to 1.2 kJoule/mol,
# and width 0.35 rad for both CVs.
#
metad: METAD ARG=t PACE=500 HEIGHT=1.2 SIGMA=0.35 FILE=HILLS

# monitor the angle variable and the metadynamics bias potential
PRINT STRIDE=10 ARG=t,metad.bias FILE=COLVAR

[Giovanni Bussi]

Notice that both COM and WHOLEMOLECULES will be very expensive with thousand atoms.

I suggest you to:

1. Prune the list (e.g. use COM of C-alpha) for COM and WHOLEMOLECULES

2. You don't need all the C-alpha for WHOLEMOLECULES, just: those used in COM and a few intermediate ones to join the sublists. Put all of them in a single ENTITY. I think there is an example about distance between ends of a polymer in the WHOLEMOLECULES manual

Giovanni

To view this discussion on the web visit https://groups.google.com/d/msgid/plumed-users/32c2e58a-0820-44fe-8f40-526f02baacfd%40googlegroups.com.

[jamesm...@gmail.com]

Thanks Giovanni. I'll have a look probably on Monday. I have some tests running over the weekend anyway.

To view this discussion on the web visit https://groups.google.com/d/msgid/plumed-users/CAPLm8Z%2B_8Y9uXLEARzbVi_h_%2BLcffayvU44go%2B0bF-U_42yaxA%40mail.gmail.com.

[James Krieger]

With all in a single ENTITY for WHOLEMOLECULES, I get a rather surprising result: some COMs are in a periodic image consistently (see attached Picture 2). I can reconstruct what I want by reanalysing with the driver after fixing periodic boundary jumps with trjconv -pbc nojump and not using WHOLEMOLECULES in the plumed.dat for that. In that case the COLVAR isn't littered with periodic boundary jumps but the variation is small. This is likely due to the GHOST dancing around COM13 in the actual metadynamics run and not consistently biasing the same angle.

Any ideas what I'm doing wrong?

Thanks
James

[Giovanni Bussi]

Hi,

read carefully in the manual the description of how WHOLEMOLECULES work. Unfortunately this is something complicated to set up, and the only way is to understand how it works in the details. Fixing with trjconv is unfortunately not valid :-)  you need the atoms to be in place during the MD.

Are the four domains part of the same chain? If not it's more complicated.

I will write an example with two chains. Chain A goes from 1-100, chain B from 101-200. If atoms 100 and 101 are always close between each others it's simple. If not, you have to identify one atom of chain A close to one atom of chain B. E.g. if atom 70 of chain A is close to atom  145 of chain B you can do the following

list: GROUP ATOMS=1-100,100-70:-1,145-101:-1,101-200

WHOLEMOLECULES ENTITY0=list

Also notice that atoms 70 and 145 should be close (significantly less than half cell side) throughout the entire simulation.

Once you have this working (e.g. in analyzing from driver) you can try to prune out atoms to increase speed.

Giovanni

To view this discussion on the web visit https://groups.google.com/d/msgid/plumed-users/3091a845-84f2-46bc-ad40-99b4963603a9%40googlegroups.com.

[James Krieger]

Hi Giovanni,

It's a homodimer so each chain has two domains. Therefore the example you've set should work - thanks. I'll let you know how I get on.

Best wishes
James

[James Krieger]

I have found a pair of Calphas for residues at the dimer interface and set WHOLEMOLECULES as follows:

list: GROUP ATOMS=1-6097,6097-1413:-1,7454-6098:-1,6098-13113
WHOLEMOLECULES STRIDE=1 ENTITY0=list

Now COM2 doesn't work as it should - it mainly sits about halfway between COM1 and where COM2 should be. I have spikes for transient excursions to where COM2 should be and also to other places. Could this be because the range 6097-1413:-1 includes parts of COM1 and parts of COM2?

COM ATOMS=1-1880,4172-6097 LABEL=com1
COM ATOMS=1881-4171,6098-6556 LABEL=com2
COM ATOMS=6557-8436,10729-12654 LABEL=com3
COM ATOMS=8437-10728,12655-13113 LABEL=com4

chain 1 = com1+com2
chain 2 = com3+com4

Thanks
James

[James Krieger]

Actually I've assigned the chains wrong in the WHOLEMOLECULES command so that might be it. Will try again with that fixed.

[James Krieger]

Yes now it's working. Thanks Giovanni. Now I'll try pruning down to Calphas and then reducing residues.

[James Krieger]

I now have the following in my plumed.dat

MOLINFO STRUCTURE=npt.pdb
list: GROUP RESIDUES=2-378,378-1413:-1,86-381:-1,381-757 RES_ATOMS=CA
WHOLEMOLECULES STRIDE=1 ENTITY0=list

# set up one variable for opening angle. A torsion is used to project the angle back into the right plane. The ghost atom is placed in order to define said plane.

com1: COM RESIDUES=2-114,243-353 RES_ATOMS=CA
com2: COM RESIDUES=115-242,354-378 RES_ATOMS=CA
com3: COM RESIDUES=381-493,622-732 RES_ATOMS=CA
com4: COM RESIDUES=494-621,733-757 RES_ATOMS=CA
com13: COM RESIDUES=2-114,243-353,381-493,622-732 RES_ATOMS=CA

and it's throwing up the following error

********** STACK DUMP **********
plumed-mpi(_ZN4PLMD9Exception28abortIfExceptionsAreDisabledEv+0x46) [0x61d3c6]
plumed-mpi(_ZN4PLMD9ExceptionC2ERKSsS2_jS2_+0x30) [0x61d80a]
plumed-mpi(_ZN4PLMD10AtomNumber9setSerialEj+0x75) [0x66e5d5]
plumed-mpi(_ZN4PLMD5Tools7convertERKSsRNS_10AtomNumberE+0x2d) [0x66bd53]
plumed-mpi(_ZN4PLMD3PDB19readFromFilepointerEP8_IO_FILEbd+0x6f1) [0x656653]
plumed-mpi(_ZN4PLMD3PDB4readERKSsbd+0x52) [0x65729a]
plumed-mpi(_ZN4PLMD12SetupMolInfoC1ERKNS_13ActionOptionsE+0x85d) [0x576597]
plumed-mpi(_ZN4PLMD5setup19MolInfoRegisterMe846createERKNS_13ActionOptionsE+0x31) [0x612aa8]
plumed-mpi(_ZN4PLMD14ActionRegister6createERKNS_13ActionOptionsE+0x4b8) [0x5399ce]
plumed-mpi(_ZN4PLMD10PlumedMain14readInputWordsERKSt6vectorISsSaISsEE+0x492) [0x5643e4]
plumed-mpi(_ZN4PLMD10PlumedMain13readInputFileESs+0x2a6) [0x5648e4]
plumed-mpi(_ZN4PLMD10PlumedMain4initEv+0x358) [0x56621e]
plumed-mpi(_ZN4PLMD10PlumedMain3cmdERKSsPv+0x49c1) [0x56ad11]
plumed-mpi(plumedmain_cmd+0x1d9) [0x573c1c]
plumed-mpi(plumed_cmd+0x32) [0x68bae7]
plumed-mpi(_ZN4PLMD7cltools6DriverIdE4mainEP8_IO_FILES4_RNS_12CommunicatorE+0x256f) [0x4cfb1f]
plumed-mpi(_ZN4PLMD10CLToolMain3runEiPPcP8_IO_FILES4_RNS_12CommunicatorE+0x16e3) [0x551e53]
plumed-mpi(_ZN4PLMD10CLToolMain3cmdERKSsPv+0xe01) [0x55359d]
plumed-mpi(_ZN4PLMD10PlumedMain3cmdERKSsPv+0x9e78) [0x5701c8]
plumed-mpi(plumedmain_cmd+0x1d9) [0x573c1c]
plumed-mpi(plumed_cmd+0x32) [0x68bae7]
plumed-mpi(main+0xfd) [0x479670]
/lib64/libc.so.6(__libc_start_main+0xfd) [0x310d41ed5d]
plumed-mpi() [0x479489]
******** END STACK DUMP ********

+++ Internal PLUMED error
+++ file AtomNumber.h, line 95
+++ message: assertion failed i>0, serial of an atom cannot be zero
[fmb333:19597] *** Process received signal ***
[fmb333:19597] Signal: Aborted (6)
[fmb333:19597] Signal code:  (-6)
[fmb333:19597] [ 0] /lib64/libpthread.so.0() [0x310d80f710]
[fmb333:19597] [ 1] /lib64/libc.so.6(gsignal+0x35) [0x310d432625]
[fmb333:19597] [ 2] /lib64/libc.so.6(abort+0x175) [0x310d433e05]
[fmb333:19597] [ 3] plumed-mpi(_ZN4PLMD9Exception6formatERKSsS2_jS2_+0) [0x61d466]
[fmb333:19597] [ 4] plumed-mpi(_ZN4PLMD9ExceptionC2ERKSsS2_jS2_+0x30) [0x61d80a]
[fmb333:19597] [ 5] plumed-mpi(_ZN4PLMD10AtomNumber9setSerialEj+0x75) [0x66e5d5]
[fmb333:19597] [ 6] plumed-mpi(_ZN4PLMD5Tools7convertERKSsRNS_10AtomNumberE+0x2d) [0x66bd53]
[fmb333:19597] [ 7] plumed-mpi(_ZN4PLMD3PDB19readFromFilepointerEP8_IO_FILEbd+0x6f1) [0x656653]
[fmb333:19597] [ 8] plumed-mpi(_ZN4PLMD3PDB4readERKSsbd+0x52) [0x65729a]
[fmb333:19597] [ 9] plumed-mpi(_ZN4PLMD12SetupMolInfoC1ERKNS_13ActionOptionsE+0x85d) [0x576597]
[fmb333:19597] [10] plumed-mpi(_ZN4PLMD5setup19MolInfoRegisterMe846createERKNS_13ActionOptionsE+0x31) [0x612aa8]
[fmb333:19597] [11] plumed-mpi(_ZN4PLMD14ActionRegister6createERKNS_13ActionOptionsE+0x4b8) [0x5399ce]
[fmb333:19597] [12] plumed-mpi(_ZN4PLMD10PlumedMain14readInputWordsERKSt6vectorISsSaISsEE+0x492) [0x5643e4]
[fmb333:19597] [13] plumed-mpi(_ZN4PLMD10PlumedMain13readInputFileESs+0x2a6) [0x5648e4]
[fmb333:19597] [14] plumed-mpi(_ZN4PLMD10PlumedMain4initEv+0x358) [0x56621e]
[fmb333:19597] [15] plumed-mpi(_ZN4PLMD10PlumedMain3cmdERKSsPv+0x49c1) [0x56ad11]
[fmb333:19597] [16] plumed-mpi(plumedmain_cmd+0x1d9) [0x573c1c]
[fmb333:19597] [17] plumed-mpi(plumed_cmd+0x32) [0x68bae7]
[fmb333:19597] [18] plumed-mpi(_ZN4PLMD7cltools6DriverIdE4mainEP8_IO_FILES4_RNS_12CommunicatorE+0x256f) [0x4cfb1f]
[fmb333:19597] [19] plumed-mpi(_ZN4PLMD10CLToolMain3runEiPPcP8_IO_FILES4_RNS_12CommunicatorE+0x16e3) [0x551e53]
[fmb333:19597] [20] plumed-mpi(_ZN4PLMD10CLToolMain3cmdERKSsPv+0xe01) [0x55359d]
[fmb333:19597] [21] plumed-mpi(_ZN4PLMD10PlumedMain3cmdERKSsPv+0x9e78) [0x5701c8]
[fmb333:19597] [22] plumed-mpi(plumedmain_cmd+0x1d9) [0x573c1c]
[fmb333:19597] [23] plumed-mpi(plumed_cmd+0x32) [0x68bae7]
[fmb333:19597] [24] plumed-mpi(main+0xfd) [0x479670]
[fmb333:19597] [25] /lib64/libc.so.6(__libc_start_main+0xfd) [0x310d41ed5d]
[fmb333:19597] [26] plumed-mpi() [0x479489]
[fmb333:19597] *** End of error message ***

Could this be because I'm using virtual sites?

Best wishes
James

[James Krieger]

I managed to get around this by using NDX_GROUP instead of MOLINFO. Also I've spotted a bug: RES_ATOMS by WHOLEMOLECULES or GROUP using MOLINFO in the VMD plugin with Plumed 2.1.2.

[James Krieger]

I'm still having problems with the GHOST dancing around. The source code file ~/bin/plumed-2.1.2-src-dmpi/src/vatom/Ghost.cpp has the following in its +PLUMEDOC comments.

\warning If a force is added to a ghost atom, the contribution
to the virial could contain small artifacts. It is therefore discouraged to
use GHOST in a constant pressure simulation.

Do you think using constant volume would fix my problem?

Thanks
James

[Giovanni Bussi]

Hi,

the bug in virial that you mention is fixed in 2.1.3.

In any case, this is not expected to make the GHOST dancing around, so there might be some other issue

Giovanni

To view this discussion on the web visit https://groups.google.com/d/msgid/plumed-users/ce7bce9d-a4cd-46b9-8544-4ac1ec9b1b63%40googlegroups.com.

[James Krieger]

Thanks Giovanni.

To view this discussion on the web visit https://groups.google.com/d/msgid/plumed-users/CAPLm8ZLJu3BwAWz%3DABGddozTgCTYZOf_d0hV93bumjE1AgOR0A%40mail.gmail.com.

[jamesm...@gmail.com]

I just thought maybe the problem is I'm defining the ghost using 3 coms approximately on a line (com13 is the com for the joint group that includes the atoms for com1 and the atoms for com3). In that case small deviations from linearity would result in an altered coordinate frame for the ghost and so it moves. As such I might be able to fix it by assigning com13 as the direct com of com1 and com3. Do you agree?

To view this discussion on the web visit https://groups.google.com/d/msgid/plumed-users/CAPLm8ZLJu3BwAWz%3DABGddozTgCTYZOf_d0hV93bumjE1AgOR0A%40mail.gmail.com.

[James Krieger]

I think the problem is that one can't define the ghost using 3 coms approximately on a line. I have tried again with another com between the two lower groups and this works fine - it stays in a stable position and allows me to define an angle on a plane for biasing with metad.

My final plumed.dat is below. com13 is defined as the centre-point for the angle and com24 is used for defining the ghost and the plane. The metad parameters were set unreasonably high on purpose to test this CV.

Best wishes
James

RESTART

list: GROUP ATOMS=1-6556,6556-1413:-1,7454-6557:-1,6557-13113

WHOLEMOLECULES STRIDE=1 ENTITY0=list

# set up one variable for opening angle. A torsion is used to project the angle back into the right plane.

# The ghost atom is placed in order to define said plane.
gr1: GROUP NDX_FILE=index.ndx NDX_GROUP=ulC_&_C-alpha
gr2: GROUP NDX_FILE=index.ndx NDX_GROUP=llC_&_C-alpha
gr3: GROUP NDX_FILE=index.ndx NDX_GROUP=ulD_&_C-alpha
gr4: GROUP NDX_FILE=index.ndx NDX_GROUP=llD_&_C-alpha
gr13: GROUP NDX_FILE=index.ndx NDX_GROUP=ulC_ulD_&_C-alpha
gr24: GROUP NDX_FILE=index.ndx NDX_GROUP=llC_llD_&_C-alpha
COM ATOMS=gr1 LABEL=com1
COM ATOMS=gr2 LABEL=com2
COM ATOMS=gr3 LABEL=com3
COM ATOMS=gr4 LABEL=com4
COM ATOMS=gr13 LABEL=com13
COM ATOMS=gr24 LABEL=com24

GHOST ATOMS=com24,com1,com3 COORDINATES=0,0.5,0 LABEL=g2
t: TORSION VECTOR1=com2,com13 AXIS=com24,g2 VECTOR2=com13,com4

DUMPATOMS STRIDE=5000 FILE=file.gro ATOMS=com1,com2,com3,com4,com13,com24,g2

# Activate metadynamics in angle/torsion t
# depositing a Gaussian every 500 time steps,
# with height equal to 1.2 kJoule/mol,
# and width 0.35 rad for both CVs.
#

metad: METAD ARG=t PACE=500 HEIGHT=1.2 SIGMA=0.35 FILE=HILLS GRID_MIN=-pi GRID_MAX=pi GRID_SPACING=0.1

# monitor the angle variable and the metadynamics bias potential
PRINT STRIDE=10 ARG=t,metad.bias FILE=COLVAR

Thanks Giovanni.

To unsubscribe from this group and stop receiving emails from it, send an email to plumed-users+unsubscribe@googlegroups.com.

To post to this group, send email to plumed...@googlegroups.com.
Visit this group at http://groups.google.com/group/plumed-users.

To view this discussion on the web visit https://groups.google.com/d/msgid/plumed-users/ce7bce9d-a4cd-46b9-8544-4ac1ec9b1b63%40googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to a topic in the Google Groups "PLUMED users" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/plumed-users/GfMqob-aeLE/unsubscribe.

To unsubscribe from this group and all its topics, send an email to plumed-users+unsubscribe@googlegroups.com.

To post to this group, send email to plumed...@googlegroups.com.
Visit this group at http://groups.google.com/group/plumed-users.

[James Krieger]

Actually there's probably no bug there. I now noticed that I didn't swap all the atom numbers to residue numbers.

[Lakshman Verma]

Hi James, 

I am trying to use GHOST atom to define the normal vector to three atoms in a small molecule. I am trying to calculate the angle of this normal with respect to the z-axis. This angle jumps randomly, which apparently happening because the GHOST particle is jumping due to PBC, I checked trajectory at the corresponding time steps. I defined the molecule as WHOLE Entity and still, my ghost particle jumps because of the PBC. My plumed script is the following:

RESTART

WHOLEMOLECULES ENTITY0=1-3024 ENTITY1=3475-3516

a0: FIXEDATOM AT=0,0,0

az: FIXEDATOM AT=0,0,1

gh: GHOST ATOMS=3479,3482,3483 NOPBC COORDINATES=0,0.5,0  # atoms 1,2,3 define plane, one GHOST atom is product

ang: ANGLE ATOMS=a0,az,gh  # defines angle between normal to the plane

The molecule is free to move and rotate.

Can you please help me figure out what I am doing wrong?

Thanks

Lakshman

[jamesm...@gmail.com]

Hi Lakshan,

I had the same problem with the ghost jumping around. I don’t remember how I fixed it but I can try to work it out. I’m flagging this email to revisit when I have more time.

Best wishes 

James 

Visit this group at https://groups.google.com/group/plumed-users.
To view this discussion on the web visit https://groups.google.com/d/msgid/plumed-users/ec96a141-ec98-4319-8a98-9aca0046bf84%40googlegroups.com.

[jamesm...@gmail.com]

Hi Lakshan,

I think the problem is with how you define your system under WHOLEMOLECULES entities. If you have a dimer you need to define it in pieces with two pieces that form the interface being next to each other in the atom ranges. That’s what I did that fixed my problem with the GHOST moving around.

It could be useful for you to revisit Giovanni Bussi’s post on this thread from 6/25/15, which explains how this works better than I can.

Best wishes 

James 

Visit this group at https://groups.google.com/group/plumed-users.
To view this discussion on the web visit https://groups.google.com/d/msgid/plumed-users/ec96a141-ec98-4319-8a98-9aca0046bf84%40googlegroups.com.