Francesco Pietra wrote:
On Wed, Oct 14, 2009 at 7:22 PM, Justin A. Lemkul <jalem...@vt.edu> wrote:
Francesco Pietra wrote:
Hi Justin:
Back from the Austrian mountains, to further isolate problems I
encountered before with my protein, I have tried to reproduce - at the
lowest stage - the CG tutorial for ubiquitin which is provided on the
MARTINI web page. I encountered the same problem as with my protein.
I started from files provided in
/output.exercises/protein/protein_data/ubiquitin, i.e.:
martini_v2.1.itp
cg.gro
cg.itp
cg.top
protein-in-vacuum.mdp
I commented out all W in cg.top.
At the command
$ grompp -np 4 -f protein-in-vacuum.mdp -c cg.gro -p cg.top -o cg.tpr
Back Off! I just backed up mdout.mdp to ./#mdout.mdp.1#
checking input for internal consistency...
WARNING 1 [file protein-in-vacuum.mdp, line unknown]:
For energy conservation with switch/shift potentials, rlist should be 0.1
to 0.3 nm larger than rcoulomb/rvdw.
calling /lib/cpp...
processing topology...
Generated 0 of the 465 non-bonded parameter combinations
Excluding 1 bonded neighbours for Protein 1
processing coordinates...
Warning: atom names in cg.top and cg.gro don't match (BCQd - BN0)
Warning: atom names in cg.top and cg.gro don't match (SCC5 - SC1)
Warning: atom names in cg.top and cg.gro don't match (BENda - BN0)
Warning: atom names in cg.top and cg.gro don't match (SEP4 - SC1)
Warning: atom names in cg.top and cg.gro don't match (BENda - BN0)
Warning: atom names in cg.top and cg.gro don't match (SEAC1 - SC1)
Warning: atom names in cg.top and cg.gro don't match (BENda - BN0)
Warning: atom names in cg.top and cg.gro don't match (SESC4 - SC1)
Warning: atom names in cg.top and cg.gro don't match (SESC4 - SC2)
Warning: atom names in cg.top and cg.gro don't match (SESC4 - SC3)
Warning: atom names in cg.top and cg.gro don't match (BENda - BN0)
Warning: atom names in cg.top and cg.gro don't match (SEAC2 - SC1)
Warning: atom names in cg.top and cg.gro don't match (BENda - BN0)
Warning: atom names in cg.top and cg.gro don't match (SEC3 - SC1)
Warning: atom names in cg.top and cg.gro don't match (SEQd - SC2)
Warning: atom names in cg.top and cg.gro don't match (BENda - BN0)
Warning: atom names in cg.top and cg.gro don't match (SEP1 - SC1)
Warning: atom names in cg.top and cg.gro don't match (BTNda - BN0)
Warning: atom names in cg.top and cg.gro don't match (STAC1 - SC1)
Warning: atom names in cg.top and cg.gro don't match (BTNda - BN0)
(more than 20 non-matching atom names)
WARNING 2 [file "cg.top", line 22]:
163 non-matching atom names
atom names from cg.top will be used
atom names from cg.gro will be ignored
double-checking input for internal consistency...
renumbering atomtypes...
converting bonded parameters...
# BONDS: 160
# G96ANGLES: 161
# PDIHS: 9
# IDIHS: 4
# CONSTR: 30
Walking down the molecule graph to make shake-blocks
initialising group options...
processing index file...
Analysing residue names:
Opening library file /usr/share/gromacs/top/aminoacids.dat
There are: 0 OTHER residues
There are: 76 PROTEIN residues
There are: 0 DNA residues
Analysing Protein...
-------------------------------------------------------
Program grompp, VERSION 3.3.3
Source code file: ../../../../src/kernel/readir.c, line: 798
Fatal error:
Group Non-Protein not found in indexfile.
Maybe you have non-default goups in your mdp file, while not using the
'-n' option of grompp.
In that case use the '-n' option.
=================
From the name "protein-in-vacuum.mdp" I assume that there is no "HOH"
You know what happens when you assume...
around in such file. Also, the tutorial was not for other non-protein.
I would be very grateful to the kind person who points out where I did
wrongly. Attached is also the mdout.mdp file.
Read the .mdp file you're using. There are two very clear appearances of
the Non-Protein term (energygrps and tc-grps). Don't simply trust an .mdp
file you found somewhere
was not found "somewhere". It is the file provided by the author of
that tutorial right for that tutorial (everything was taken from the
solution provided to the example).
Indeed, it may work in certain cases, but my general advice is simply this: if
you didn't make it yourself with manual in hand, you have to still comb through
it to make sure it addresses your system. If you have only protein, and
Non-Protein is defined somewhere, then the .mdp file does not suit your needs.
I have also seen many instances of files that "work" in tutorials (topologies,
scripts, .mdp files) wind up not working when I look into them. This is not to
say that any particular individual has posted anything incorrectly, but always
check your input for yourself!
and are potentially having to hack at. Read the
input, determine the appropriate settings and why those you're using may not
work.
I'll look at. But I fear that by commenting out what alludes to
non-protein (if this is what I should do; if not, please explain what
I should do), it will not work. It was so with "my protein".
If your system has only protein, then a suitable setting for both the energygrps
and tc-grps would be either "Protein" or "System", since they are essentially
equivalent in this case. If you have only protein, then nothing should be
non-protein, right? The combination of "Protein Non-Protein" in these settings
is designed for a simulation of protein in solvent, and perhaps ions, to prevent
unsuspecting users from coupling ions separately to a thermostat (which usually
leads to a crash).
At any event, for a person like me coming from AMBER, it is surprising
that there is no "minimum" input that works with every standard
protein.
I think that this could be dangerous. Different force fields were parameterized
for different cutoffs, vdW and electrostatics methods, etc, and so it may be
desirable to adhere rigorously to the appropriate settings, given the diversity
of force fields available in GROMACS (GROMOS, OPLS, AMBER, CHARMM soon, CG
models). Many parameters can be left at the default, but others would certainly
require a bit of user input to be assured of their validity :)
-Justin
Thanks
francesco
-Justin
thanks
francesco pietra
On Mon, Oct 5, 2009 at 11:14 PM, Justin A. Lemkul <jalem...@vt.edu> wrote:
Francesco Pietra wrote:
Look at the .mdp file. You will find Non-Protein somewhere.
The .mdp I used is not from a tutorial; it is the general purpose - or
standard - mdp provided by Martini. It doe not contain the word
"protein"
The "protein-in-vacuum.mdp from the ubiquitin CG tutoria does contain
both "Protein" and "Non-Protein" words. deleting "Non-Protein", the
error from running grompp is:
Fatal error:
Invalid T coupling input: 1 groups, 2 ref_t values and 2 tau_t values
This comes up because you've defined one group now for T-coupling and
assigned coupling constants and temperatures for two. Haphazard changes
to
.mdp files will certainly generate these errors. Deleting until the
problem
goes away never works - it only creates more problems :)
Please see the manual for proper use of tc-grps, etc.
I'll settle the matter until I can generate ideas, or be less tired to
check again for the presence of that mysterious non-protein. The real
protein contains chloride ion ligands. I have cleared from them before
starting with CG and anyway no trace of cl CL Cl.
So that has nothing to do with it.
Hope to be able to come again here with a solution of this affair.
The best advice - don't blindly use some .mdp file you find because it
works
for a tutorial. Read it, read the manual, and make changes based on
what
you actually have in your system.
I fully agree. Using "protein-in-vacuum.mdp", although attractive
because of "vacuum", was a bold adventure. But the standard mdp I used
before is declared to be standard, and my protein is standard. As I
said, I'll work hard to come out but before I'll take a few days
wandering with my Brittany on the Alps. The Tetraonidae will refresh
my mind.
What is "standard" for one user is not necessarily "standard" for
another.
You must almost always make changes to suit your particular system, as a
general rule.
Thanks for all you did for me (for us)
You're welcome. Hopefully everything works out.
-Justin
francesco
-Justin
thanks
francesco
PS: the protein CG file contains TER lines separating the subunits,
but the CG .gro file has no TER.
-Justin
;
; mod21 | MARTINI 2.1
;
; Include force field poarameters containing all particle
definitions,
; the interaction matrix, plus the topology for water.
#include "martini_v2.1.itp"
; Then include the file(s) containing the topologies of other
; molecules present in your system.
; #include "martini_v2.0_lipids.itp"
; #include "martini_v2.0_salt.itp"
#include "mod21.itp"
; Define a name for your system
[ system ]
mod21
; Define the composition of your system
; The molecule names should correspond to those defined in the itp
file(s).
[ molecules ]
Protein 1
; DPPC 128
; W 2000
; NA+ 20
; CL- 20
It seems that commenting out ";" is placed correctly.
I am sincerely sorry to bother you again.
francesco
With command
$ grompp -np 4 -f mod21.cg.mdp -c mod21.cg.box.gro -p mod21.cg.top
-n
-o mod21.cg.tpr
NOTE:
System has non-zero total charge: -5.000000e+01
processing coordinates...
Warning: atom names in mod21.cg.top and mod21.cg.box.gro don't
match
(BCQd - BN0)
Warning: atom names in mod21.cg.top and mod21.cg.box.gro don't
match
(SCAC1 - SC1)
Warning: atom names in mod21.cg.top and mod21.cg.box.gro don't
match
(BCP5 - BN0)
Warning: atom names in mod21.cg.top and mod21.cg.box.gro don't
match
(SCP1 - SC1)
Warning: atom names in mod21.cg.top and mod21.cg.box.gro don't
match
(BSP5 - BN0)
Warning: atom names in mod21.cg.top and mod21.cg.box.gro don't
match
(SSAC1 - SC1)
Warning: atom names in mod21.cg.top and mod21.cg.box.gro don't
match
(BCP5 - BN0)
Warning: atom names in mod21.cg.top and mod21.cg.box.gro don't
match
(SCC3 - SC1)
Warning: atom names in mod21.cg.top and mod21.cg.box.gro don't
match
(SCQd - SC2)
Warning: atom names in mod21.cg.top and mod21.cg.box.gro don't
match
(BCP5 - BN0)
Warning: atom names in mod21.cg.top and mod21.cg.box.gro don't
match
(SCN0 - SC1)
Warning: atom names in mod21.cg.top and mod21.cg.box.gro don't
match
(SCQd - SC2)
Warning: atom names in mod21.cg.top and mod21.cg.box.gro don't
match
(BHN0 - BN0)
Warning: atom names in mod21.cg.top and mod21.cg.box.gro don't
match
(BHNd - BN0)
Warning: atom names in mod21.cg.top and mod21.cg.box.gro don't
match
(SHAC1 - SC1)
Warning: atom names in mod21.cg.top and mod21.cg.box.gro don't
match
(BHNd - BN0)
Warning: atom names in mod21.cg.top and mod21.cg.box.gro don't
match
(SHSC4 - SC2)
Warning: atom names in mod21.cg.top and mod21.cg.box.gro don't
match
(SHSP1 - SC1)
Warning: atom names in mod21.cg.top and mod21.cg.box.gro don't
match
(SHSC4 - SC3)
Warning: atom names in mod21.cg.top and mod21.cg.box.gro don't
match
(SHSC4 - SC4)
(more than 20 non-matching atom names)
WARNING 2 [file "mod21.cg.top", line 28]:
2875 non-matching atom names
atom names from mod21.cg.top will be used
atom names from mod21.cg.box.gro will be ignored
double-checking input for internal consistency...
renumbering atomtypes...
converting bonded parameters...
# BONDS: 2724
# G96ANGLES: 2860
# PDIHS: 328
# IDIHS: 187
# CONSTR: 745
Walking down the molecule graph to make shake-blocks
initialising group options...
processing index file...
-------------------------------------------------------
Program grompp, VERSION 3.3.3
Source code file: ../../../../src/gmxlib/futil.c, line: 345
File input/output error:
index.ndx
==========
Non-matching because of the many -S-S- bonds? Why asking about
index.ndx?
Non-matching names occur because MARTINI doesn't assign them
correctly
in
the initial system setup. Using the .top names is fine.
The I/O error comes from the fact that you have passed the -n flag
to
grompp
without having index.ndx in the working directory. Note that
having
an
index file doesn't necessarily solve your problem; the warning is
printed
in
case you have non-default groups for, e.g., T-coupling or
energygrps.
If
you have a "moleculetype not found," then you haven't #included the
.itp
file and an index file won't save you.
-Justin
Thanks for help
francesco
On Mon, Oct 5, 2009 at 12:33 PM, Justin A. Lemkul
<jalem...@vt.edu>
wrote:
Francesco Pietra wrote:
Hi:
As I found it difficult to manage solvation - the way I would
like
-
with a pore protein (a trimer) partly immersed into a bilayer, I
started with the protein alone in vacuum, in order to try to fix
the
problems step-by-step. Unfortunately, I can't fix simple
problems
with
grompp.
1) Generated mod21.itp from mod21.pdb
2) Generated mod21.cg.pdb with awk script)
3) editconf -f mod21.cg.pdb -o mod21.cg.gro
4) genbox -cp mod21.cg.gro -box 10 17 10 -o mod21.cg.box.gro
5) grompp -np 4 -f mod21.cg.box.mdp -c mod21.cg.box.gro -p
mod21.cg.box.top -warnmax 10
Error: no such moleculetype Protein.
.............................
.mdp is merely a renamed martini_v2.0.example.mdp, taking care
that
"cpp = /usr/bin/cpp".
.top was from editing martini_v2.0.example.top and reads (in
between
===):
======
;
; mod21 | MARTINI 2.1
;
; Include force field parameters containing all particle
definitions,
; the interaction matrix, plus the topology for water.
#include "martini_v2.1.itp"
; Then include the file(s) containing the topologies of other
; molecules present in your system.
; #include "martini_v2.0_lipids.itp"
; #include "martini_v2.0_salt.itp"
#include mod21.itp
This #include statement has the wrong syntax. See any of the
above
for
the
correct way to do it.
-Justin
--
========================================
Justin A. Lemkul
Ph.D. Candidate
ICTAS Doctoral Scholar
Department of Biochemistry
Virginia Tech
Blacksburg, VA
jalemkul[at]vt.edu | (540) 231-9080
http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin
========================================
_______________________________________________
gmx-users mailing list gmx-users@gromacs.org
http://lists.gromacs.org/mailman/listinfo/gmx-users
Please search the archive at http://www.gromacs.org/search before
posting!
Please don't post (un)subscribe requests to the list. Use the www
interface
or send it to gmx-users-requ...@gromacs.org.
Can't post? Read http://www.gromacs.org/mailing_lists/users.php
--
========================================
Justin A. Lemkul
Ph.D. Candidate
ICTAS Doctoral Scholar
Department of Biochemistry
Virginia Tech
Blacksburg, VA
jalemkul[at]vt.edu | (540) 231-9080
http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin
========================================
_______________________________________________
gmx-users mailing list gmx-users@gromacs.org
http://lists.gromacs.org/mailman/listinfo/gmx-users
Please search the archive at http://www.gromacs.org/search before
posting!
Please don't post (un)subscribe requests to the list. Use the www
interface
or send it to gmx-users-requ...@gromacs.org.
Can't post? Read http://www.gromacs.org/mailing_lists/users.php
--
========================================
Justin A. Lemkul
Ph.D. Candidate
ICTAS Doctoral Scholar
Department of Biochemistry
Virginia Tech
Blacksburg, VA
jalemkul[at]vt.edu | (540) 231-9080
http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin
========================================
_______________________________________________
gmx-users mailing list gmx-users@gromacs.org
http://lists.gromacs.org/mailman/listinfo/gmx-users
Please search the archive at http://www.gromacs.org/search before
posting!
Please don't post (un)subscribe requests to the list. Use the www
interface
or send it to gmx-users-requ...@gromacs.org.
Can't post? Read http://www.gromacs.org/mailing_lists/users.php
--
========================================
Justin A. Lemkul
Ph.D. Candidate
ICTAS Doctoral Scholar
Department of Biochemistry
Virginia Tech
Blacksburg, VA
jalemkul[at]vt.edu | (540) 231-9080
http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin
========================================
_______________________________________________
gmx-users mailing list gmx-users@gromacs.org
http://lists.gromacs.org/mailman/listinfo/gmx-users
Please search the archive at http://www.gromacs.org/search before
posting!
Please don't post (un)subscribe requests to the list. Use the www
interface
or send it to gmx-users-requ...@gromacs.org.
Can't post? Read http://www.gromacs.org/mailing_lists/users.php
--
========================================
Justin A. Lemkul
Ph.D. Candidate
ICTAS Doctoral Scholar
Department of Biochemistry
Virginia Tech
Blacksburg, VA
jalemkul[at]vt.edu | (540) 231-9080
http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin
========================================
_______________________________________________
gmx-users mailing list gmx-users@gromacs.org
http://lists.gromacs.org/mailman/listinfo/gmx-users
Please search the archive at http://www.gromacs.org/search before
posting!
Please don't post (un)subscribe requests to the list. Use the www
interface
or send it to gmx-users-requ...@gromacs.org.
Can't post? Read http://www.gromacs.org/mailing_lists/users.php
--
========================================
Justin A. Lemkul
Ph.D. Candidate
ICTAS Doctoral Scholar
Department of Biochemistry
Virginia Tech
Blacksburg, VA
jalemkul[at]vt.edu | (540) 231-9080
http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin
========================================
_______________________________________________
gmx-users mailing list gmx-users@gromacs.org
http://lists.gromacs.org/mailman/listinfo/gmx-users
Please search the archive at http://www.gromacs.org/search before posting!
Please don't post (un)subscribe requests to the list. Use the www interface
or send it to gmx-users-requ...@gromacs.org.
Can't post? Read http://www.gromacs.org/mailing_lists/users.php
--
========================================
Justin A. Lemkul
Ph.D. Candidate
ICTAS Doctoral Scholar
Department of Biochemistry
Virginia Tech
Blacksburg, VA
jalemkul[at]vt.edu | (540) 231-9080
http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin
========================================
_______________________________________________
gmx-users mailing list gmx-users@gromacs.org
http://lists.gromacs.org/mailman/listinfo/gmx-users
Please search the archive at http://www.gromacs.org/search before posting!
Please don't post (un)subscribe requests to the list. Use the
www interface or send it to gmx-users-requ...@gromacs.org.
Can't post? Read http://www.gromacs.org/mailing_lists/users.php
