normal Reverse CG with Amber

  • is
  • is's Avatar Topic Author
  • Offline
  • Senior Boarder
More
13 years 1 month ago #532 by is
Replied by is on topic Reverse CG with Amber

xavier wrote:

is wrote:

xavier wrote:

is wrote: Hi XAvier,

So did as you suggested, I added mapping section for Amber in ffamber03m.rtp file for standard amino acids, I renamed files and linked them if necessary. However, I still get exactly the same error as before. How to check if gmx has potential forms for Amber ff?

is

Run an atomic simulation of your system. No need to rerun with gmx3.3.1, but a short run to check whether the amber force field is defined or not!

You may also want to go on the gromacs web site whether gmx3.3.1 has the amber force field you use implemented.

XAvier.


So, there is no problem with atomistic simulation. Gromacs doesn't complain, and I can generate needed files.

Ok so you should check your mapping.

As Andrzej said earlier every atom should be assigned to a CG bead. It is not fundamental when going from atomistic to CG but from CG to atomist it is.

Check and check again ... must be something funky somewhere :))


I'm checking again my mapping section. I tried RV-CG only with protein (without ligand bound, water, or ions) using ffamber03m.rtp file. And I still get segmentation fault.

What is more interesting is that when I don't include cg topology file in g_fg2cg command then I get:

Name of CG water in topology should be W

Fg pre-structure computed !

Back Off! I just backed up fg.gro to ./#fg.gro.3#

So structure is computed, but with zeros as coordinates. (Yes, I'm still checking my mapping:))
It seems that my cg top file causes this segmentation fault, but it looks as simple as possible, and I can't see error here:

; Include forcefield parameters
#include "martini_v2.1.itp"

; Include chain topologies
#include "prot-cg.itp"
;#include "lig-cg.itp"

[ system ]
; Name
Protein in water

[ molecules ]
; Compound #mol
Protein 1
;ligand 1

Suggestions?
Thank you
is

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

  • is
  • is's Avatar Topic Author
  • Offline
  • Senior Boarder
More
13 years 1 month ago #533 by is
Replied by is on topic Reverse CG with Amber
One more question, but I guess is more to Andrzej.

For example, we have SER where atoms have some order:
[ SER ]
[ atoms ]
N amber99_34 -0.541430 1
H amber99_17 0.345415 2
CA amber99_11 0.118140 3
HA amber99_19 0.142177 4
CB amber99_11 0.146998 5
HB1 amber99_19 0.040081 6
HB2 amber99_19 0.040081 7
OG amber99_43 -0.640312 8
HG amber99_25 0.446255 9
C amber99_2 0.483424 10
O amber99_41 -0.580829 11

If we then mapped them into beads this order will be disturbed:

[ mapping ]
; bead atom
1 N
1 H
1 CA
1 HA
1 C
1 O
2 CB
2 HB1
2 HB2
2 OG
2 HG

Should I put them in correct order, as in residue definition (SER), or as shown above? Does it matter?

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

More
13 years 1 month ago #535 by xavier
Replied by xavier on topic Reverse CG with Amber

is wrote: One more question, but I guess is more to Andrzej.

For example, we have SER where atoms have some order:
[ SER ]
[ atoms ]
N amber99_34 -0.541430 1
H amber99_17 0.345415 2
CA amber99_11 0.118140 3
HA amber99_19 0.142177 4
CB amber99_11 0.146998 5
HB1 amber99_19 0.040081 6
HB2 amber99_19 0.040081 7
OG amber99_43 -0.640312 8
HG amber99_25 0.446255 9
C amber99_2 0.483424 10
O amber99_41 -0.580829 11

If we then mapped them into beads this order will be disturbed:

[ mapping ]
; bead atom
1 N
1 H
1 CA
1 HA
1 C
1 O
2 CB
2 HB1
2 HB2
2 OG
2 HG

Should I put them in correct order, as in residue definition (SER), or as shown above? Does it matter?

Ok, this mapping is not perfect!

In the mapping you should include all possible atoms of each residue. If an atom appears and is not defined in the mapping then gmx will crash.

This is typically the case for termini atoms and in particular H. Check the ffG53a6m.rtp and you'll see that the atoms H1, H2, H3 and H and defined and they are all possible names of H attached to the N backbone atom. The N1-3 are when a residue is the Nterminus and H is when it is in the middle of the protein ...
Same applies for the C terminus :))

You are almost there :))

XAvier.

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

  • andrzej
  • andrzej's Avatar
  • Visitor
13 years 1 month ago #536 by andrzej
Replied by andrzej on topic Reverse CG with Amber
For the 3.3.1 there were amber ports, that should be installed from the website ffamber.cnsm.csulb.edu/ . Then you have to make the mappings for all atoms that you will have in the atomistic gro file. I made it for all possible atoms, so e.g for ala the basic atoms in rtp are:
[atoms]
N N -0.31000 0
H H 0.31000 0
CA CH1 0.00000 1
CB CH3 0.00000 1
C C 0.450 2
O O -0.450 2
and mappings are defined for all possible additional atoms (H1, H2 ...O2), e.g. due to protonation state or end or start position of aminoacid:
[mapping]
1 N
1 H1
1 H2
1 H3
1 H
1 CA
1 CB
1 C
1 O
1 O1
1 O2

During the execution of the modified pdb2gmx with the command
pdb2gmx -ignh -missing -f aa.gro -p aa.top


you will see that the atoms that are in the mapping but are not in the gro file are listed ( so all the H1, H2 ... from the middle of your protein), this way you can dubble check if the mapping is ok.

At the end in the topology of the protein you have to have the mapping section.
I think the easiest is to make the topology for chain A and B separately and include them in the final topology file, together with the fg_w water itp file.

Andrzej

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

  • is
  • is's Avatar Topic Author
  • Offline
  • Senior Boarder
More
13 years 1 month ago #537 by is
Replied by is on topic Reverse CG with Amber

xavier wrote:

is wrote: One more question, but I guess is more to Andrzej.

For example, we have SER where atoms have some order:
[ SER ]
[ atoms ]
N amber99_34 -0.541430 1
H amber99_17 0.345415 2
CA amber99_11 0.118140 3
HA amber99_19 0.142177 4
CB amber99_11 0.146998 5
HB1 amber99_19 0.040081 6
HB2 amber99_19 0.040081 7
OG amber99_43 -0.640312 8
HG amber99_25 0.446255 9
C amber99_2 0.483424 10
O amber99_41 -0.580829 11

If we then mapped them into beads this order will be disturbed:

[ mapping ]
; bead atom
1 N
1 H
1 CA
1 HA
1 C
1 O
2 CB
2 HB1
2 HB2
2 OG
2 HG

Should I put them in correct order, as in residue definition (SER), or as shown above? Does it matter?

Ok, this mapping is not perfect!

In the mapping you should include all possible atoms of each residue. If an atom appears and is not defined in the mapping then gmx will crash.

This is typically the case for termini atoms and in particular H. Check the ffG53a6m.rtp and you'll see that the atoms H1, H2, H3 and H and defined and they are all possible names of H attached to the N backbone atom. The N1-3 are when a residue is the Nterminus and H is when it is in the middle of the protein ...
Same applies for the C terminus :))

You are almost there :))

XAvier.


OK, that was only example. Amber, beside SER, has also CSER and NSER definitions, where I included all atoms:
CSER:
[ mapping ]
; bead atom
1 N
1 H
1 CA
1 HA
1 C
1 OC1
1 OC2
2 CB
2 HB1
2 HB2
2 OG
2 HG

NSER:
[ mapping ]
; bead atom
1 N
1 H1
1 H2
1 H3
1 CA
1 HA
1 C
1 O
2 CB
2 HB1
2 HB2
2 OG
2 HG
You are saying that even there are CSER and NSER I should add: H1, H2, H3, OC1, OC2 to SER? Wouldn't that be wrong?

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

  • is
  • is's Avatar Topic Author
  • Offline
  • Senior Boarder
More
13 years 1 month ago #538 by is
Replied by is on topic Reverse CG with Amber

andrzej wrote: For the 3.3.1 there were amber ports, that should be installed from the website ffamber.cnsm.csulb.edu/ . Then you have to make the mappings for all atoms that you will have in the atomistic gro file. I made it for all possible atoms, so e.g for ala the basic atoms in rtp are:
[atoms]
N N -0.31000 0
H H 0.31000 0
CA CH1 0.00000 1
CB CH3 0.00000 1
C C 0.450 2
O O -0.450 2
and mappings are defined for all possible additional atoms (H1, H2 ...O2), e.g. due to protonation state or end or start position of aminoacid:
[mapping]
1 N
1 H1
1 H2
1 H3
1 H
1 CA
1 CB
1 C
1 O
1 O1
1 O2

During the execution of the modified pdb2gmx with the command
pdb2gmx -ignh -missing -f aa.gro -p aa.top


you will see that the atoms that are in the mapping but are not in the gro file are listed ( so all the H1, H2 ... from the middle of your protein), this way you can dubble check if the mapping is ok.

At the end in the topology of the protein you have to have the mapping section.
I think the easiest is to make the topology for chain A and B separately and include them in the final topology file, together with the fg_w water itp file.

Andrzej


OK, thanks.
So, to be clear (for example) SER, CSER, NSER will have the same mapping section in ffamber03m.rtp file?
I will let you know how it goes.

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

More
13 years 1 month ago #539 by xavier
Replied by xavier on topic Reverse CG with Amber

is wrote:

andrzej wrote: For the 3.3.1 there were amber ports, that should be installed from the website ffamber.cnsm.csulb.edu/ . Then you have to make the mappings for all atoms that you will have in the atomistic gro file. I made it for all possible atoms, so e.g for ala the basic atoms in rtp are:
[atoms]
N N -0.31000 0
H H 0.31000 0
CA CH1 0.00000 1
CB CH3 0.00000 1
C C 0.450 2
O O -0.450 2
and mappings are defined for all possible additional atoms (H1, H2 ...O2), e.g. due to protonation state or end or start position of aminoacid:
[mapping]
1 N
1 H1
1 H2
1 H3
1 H
1 CA
1 CB
1 C
1 O
1 O1
1 O2

During the execution of the modified pdb2gmx with the command
pdb2gmx -ignh -missing -f aa.gro -p aa.top


you will see that the atoms that are in the mapping but are not in the gro file are listed ( so all the H1, H2 ... from the middle of your protein), this way you can dubble check if the mapping is ok.

At the end in the topology of the protein you have to have the mapping section.
I think the easiest is to make the topology for chain A and B separately and include them in the final topology file, together with the fg_w water itp file.

Andrzej


OK, thanks.
So, to be clear (for example) SER, CSER, NSER will have the same mapping section in ffamber03m.rtp file?
I will let you know how it goes.

If AMBER differentiates the residues at the termini then it should fine not to add the atoms present in the termini in the mapping of the residues not in the termini. The situation is different for the GROMOS force fields that Andrzej has been working with.

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

  • andrzej
  • andrzej's Avatar
  • Visitor
13 years 1 month ago #540 by andrzej
Replied by andrzej on topic Reverse CG with Amber
I must admit that I forgot how terminal amino acids are defined in Amber ports. What is important is that in the mapping you have all the atoms that you want to have in the gro file ( i.e. the ones that are in the original pdb/gro file used to make atomistic topology). There should be no problems specific to Amber, i played some time ago with a few selected aminoacids and everything worked well.

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

  • is
  • is's Avatar Topic Author
  • Offline
  • Senior Boarder
More
13 years 1 month ago #541 by is
Replied by is on topic Reverse CG with Amber

andrzej wrote: I must admit that I forgot how terminal amino acids are defined in Amber ports. What is important is that in the mapping you have all the atoms that you want to have in the gro file ( i.e. the ones that are in the original pdb/gro file used to make atomistic topology). There should be no problems specific to Amber, i played some time ago with a few selected aminoacids and everything worked well.


Then I'm doing something wrong (but no clue what). I will start again from the beginning.

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

More
13 years 1 month ago #542 by xavier
Replied by xavier on topic Reverse CG with Amber

is wrote:

andrzej wrote: I must admit that I forgot how terminal amino acids are defined in Amber ports. What is important is that in the mapping you have all the atoms that you want to have in the gro file ( i.e. the ones that are in the original pdb/gro file used to make atomistic topology). There should be no problems specific to Amber, i played some time ago with a few selected aminoacids and everything worked well.


Then I'm doing something wrong (but no clue what). I will start again from the beginning.


Good idea to start again. Let us know how it goes and we might have a look at your files if troubles persist.

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

  • is
  • is's Avatar Topic Author
  • Offline
  • Senior Boarder
More
13 years 1 month ago #548 by is
Replied by is on topic Reverse CG with Amber
Hi
I have something new:)

I started again from the beginning.
Then I used command:

pdb2gmx -f aa.gro -p aa.top

And I got:

Analyzing pdb file
.
.
.

Opening library file ffamber03m.atp
Atomtype 74
Reading residue database... (ffamber03m)
Opening library file ffamber03m.rtp
Residue 73
Program pdb2gmx, VERSION 3.3.1
Source code file: resall.c, line: 357

Fatal error:
in .rtp file in residue mappping at line:
1 N



First I thought that something is wrong with residue 73 (VAL) in my gro file, and mapping section for it. I checked both and I don't see anything unusual. What more VAL is listed in gro file earlier. What else can cause this error?

is

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

  • andrzej
  • andrzej's Avatar
  • Visitor
13 years 1 month ago #549 by andrzej
Replied by andrzej on topic Reverse CG with Amber
Fatal error:
in .rtp file in residue mappping at line:
1 N


mapping has 3 p in your error

Andrzej

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

  • is
  • is's Avatar Topic Author
  • Offline
  • Senior Boarder
More
13 years 1 month ago #550 by is
Replied by is on topic Reverse CG with Amber

andrzej wrote: Fatal error:
in .rtp file in residue mappping at line:
1 N


mapping has 3 p in your error

Andrzej


Thanks:)

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

  • is
  • is's Avatar Topic Author
  • Offline
  • Senior Boarder
More
13 years 1 month ago #551 by is
Replied by is on topic Reverse CG with Amber
Hi Andrzej,

When you played with Amber, did you modify your gro file (put names special for Amber).
How comment line should looked like?

is

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

  • is
  • is's Avatar Topic Author
  • Offline
  • Senior Boarder
More
13 years 1 month ago #552 by is
Replied by is on topic Reverse CG with Amber
It works:)
Thank you for your help Xavier and Andrzej.

I will test it on whole system (as I tried it only on a small protein), and see how it goes with water. I hope there will be no problems.
Did you, Andrzej write this mapping for ions? I would be still interested to try it.
Also if you are interested in this files, let me know.

is

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

  • is
  • is's Avatar Topic Author
  • Offline
  • Senior Boarder
More
13 years 1 month ago #553 by is
Replied by is on topic Reverse CG with Amber

is wrote: It works:)
Thank you for your help Xavier and Andrzej.

I will test it on whole system (as I tried it only on a small protein), and see how it goes with water. I hope there will be no problems.
Did you, Andrzej write this mapping for ions? I would be still interested to try it.
Also if you are interested in this files, let me know.

is


Works with water, works with NA+ ions (as my system has only this one:)).

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

  • is
  • is's Avatar Topic Author
  • Offline
  • Senior Boarder
More
13 years 1 month ago #555 by is
Replied by is on topic Reverse CG with Amber
Hi,
I get an error at the stage of mdrun:

Warning: 1-4 interaction between 4899 and 4905 at distance 1.217 which is larger than the 1-4 table size 1.200 nm
These are ignored for the rest of the simulation
This usually means your system is exploding,
if not, you should increase table-extension in your mdp file
step 0
Program mdrun, VERSION 3.3.1
Source code file: nsgrid.c, line: 226

Range checking error:
Explanation: During neighborsearching, we assign each particle to a grid
based on its coordinates. If your system contains collisions or parameter
errors that give particles very high velocities you might end up with some
coordinates being +-Infinity or NaN (not-a-number). Obviously, we cannot
put these on a grid, so this is usually where we detect those errors.
Make sure your system is properly energy-minimized and that the potential
energy seems reasonable before trying again.

Variable ci has value -2147483648. It should have been within [ 0 .. 2197 ]
Please report this to the mailing list (This email address is being protected from spambots. You need JavaScript enabled to view it.)

So, I increased table-extension in fg.mdp file, in order to see if it would help. It did not.
I search forums (this one and Gromacs) and one of the advices was to check the RMSD.
I know that my protein changes its conformation during simulation, thus there is a jump in RMSD, but then protein finds it stable state. It does not collapse. Thus, I have a question.
To generate tpr file for simulating annealing, I use atomistic mdp, gro and top files (grompp step) and then in mdrun I use cg structure. I expect that both fg and cg structures are compared, but to what degree?

grompp -f fg.mdp -c fg.gro -p topol.top -o sa_formd.tpr

mdrun -coarse cg.gro -v -s sa_formd.tpr

Isn't it that Gromacs will always print error about exploding, as these two structures represent two different conformations?
Or maybe there is a simpler explanation (mistakes in commands...), and thus simpler solution?

Thanks for suggestions.
is

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

More
13 years 1 month ago #560 by xavier
Replied by xavier on topic Reverse CG with Amber

is wrote: Hi,
I get an error at the stage of mdrun:

Warning: 1-4 interaction between 4899 and 4905 at distance 1.217 which is larger than the 1-4 table size 1.200 nm
These are ignored for the rest of the simulation
This usually means your system is exploding,
if not, you should increase table-extension in your mdp file
step 0


Program mdrun, VERSION 3.3.1
Source code file: nsgrid.c, line: 226

Range checking error:
Explanation: During neighborsearching, we assign each particle to a grid
based on its coordinates. If your system contains collisions or parameter
errors that give particles very high velocities you might end up with some
coordinates being +-Infinity or NaN (not-a-number). Obviously, we cannot
put these on a grid, so this is usually where we detect those errors.
Make sure your system is properly energy-minimized and that the potential
energy seems reasonable before trying again.

Variable ci has value -2147483648. It should have been within [ 0 .. 2197 ]
Please report this to the mailing list (This email address is being protected from spambots. You need JavaScript enabled to view it.)

So, I increased table-extension in fg.mdp file, in order to see if it would help. It did not.
I search forums (this one and Gromacs) and one of the advices was to check the RMSD.
I know that my protein changes its conformation during simulation, thus there is a jump in RMSD, but then protein finds it stable state. It does not collapse. Thus, I have a question.
To generate tpr file for simulating annealing, I use atomistic mdp, gro and top files (grompp step) and then in mdrun I use cg structure. I expect that both fg and cg structures are compared, but to what degree?

At that stage of the transformation the AA structure does not look like anything reasonable. Atomistic positions are random. It is only after the simulated annealing that it will become realistic. The atomistic positions should however be close to the CG beads. Did you check?

grompp -f fg.mdp -c fg.gro -p topol.top -o sa_formd.tpr

mdrun -coarse cg.gro -v -s sa_formd.tpr

Looks good.

Isn't it that Gromacs will always print error about exploding, as these two structures represent two different conformations?

I am not sure what you mean here.

Or maybe there is a simpler explanation (mistakes in commands...), and thus simpler solution?

The command lines seem to be fine!

Did you look at what is actually exploding? Visualization could help. That might give you some idea where to look. It might be a funny thing in your topology!

Which parameters for the simulated annealing did you use?

Have you tried to decompose you transformation in each of the molecules you have in the system and check what is going wrong ... may be the mapping is not perfect :))

Thanks for suggestions.
is

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

  • is
  • is's Avatar Topic Author
  • Offline
  • Senior Boarder
More
13 years 1 month ago #576 by is
Replied by is on topic Reverse CG with Amber

xavier wrote:

is wrote: Hi,
I get an error at the stage of mdrun:

Warning: 1-4 interaction between 4899 and 4905 at distance 1.217 which is larger than the 1-4 table size 1.200 nm
These are ignored for the rest of the simulation
This usually means your system is exploding,
if not, you should increase table-extension in your mdp file
step 0


Program mdrun, VERSION 3.3.1
Source code file: nsgrid.c, line: 226

Range checking error:
Explanation: During neighborsearching, we assign each particle to a grid
based on its coordinates. If your system contains collisions or parameter
errors that give particles very high velocities you might end up with some
coordinates being +-Infinity or NaN (not-a-number). Obviously, we cannot
put these on a grid, so this is usually where we detect those errors.
Make sure your system is properly energy-minimized and that the potential
energy seems reasonable before trying again.

Variable ci has value -2147483648. It should have been within [ 0 .. 2197 ]
Please report this to the mailing list (This email address is being protected from spambots. You need JavaScript enabled to view it.)

So, I increased table-extension in fg.mdp file, in order to see if it would help. It did not.
I search forums (this one and Gromacs) and one of the advices was to check the RMSD.
I know that my protein changes its conformation during simulation, thus there is a jump in RMSD, but then protein finds it stable state. It does not collapse. Thus, I have a question.
To generate tpr file for simulating annealing, I use atomistic mdp, gro and top files (grompp step) and then in mdrun I use cg structure. I expect that both fg and cg structures are compared, but to what degree?
At that stage of the transformation the AA structure does not look like anything reasonable. Atomistic positions are random. It is only after the simulated annealing that it will become realistic. The atomistic positions should however be close to the CG beads. Did you check?


Yes, I did. It seems to be OK. The CA from fg structure and BB from cg structure are very close, but they do not overlap. The rest of atoms and beads have also close positions. None of the beads or atoms are "flying" around.

grompp -f fg.mdp -c fg.gro -p topol.top -o sa_formd.tpr

mdrun -coarse cg.gro -v -s sa_formd.tpr
Looks good.

Isn't it that Gromacs will always print error about exploding, as these two structures represent two different conformations?

I am not sure what you mean here.


I was thinking about something that is not relevant, sorry for that.

Or maybe there is a simpler explanation (mistakes in commands...), and thus simpler solution?

The command lines seem to be fine!

Did you look at what is actually exploding? Visualization could help. That might give you some idea where to look. It might be a funny thing in your topology!

Simulation crashes before it actually starts, thus I have only starting .xtc frame.

Which parameters for the simulated annealing did you use?


For Protein and Non-Protein (water and ions) I changed number of steps to 1000 and I started with:

annealing_time = 0 1.5 0 1.5
annealing_temp = 1300 300 400 300

as suggested in tutorial. Rest I left without changes.

Have you tried to decompose you transformation in each of the molecules you have in the system and check what is going wrong


I will do that right away.

... may be the mapping is not perfect :))


Maybe:)

Thanks for suggestions.
is

[/quote]

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

  • is
  • is's Avatar Topic Author
  • Offline
  • Senior Boarder
More
13 years 1 month ago #577 by is
Replied by is on topic Reverse CG with Amber
The problem was with one of the top files.

I used command:

pdb2gmx -f aa.gro

to generate top files with mapping sections for protein and ligand. However, instead of that I got itp files: protein_A and protein_B with mapping sections for protein and IONS (should be for ligand), respectively.
Which caused all problems.
I fixed that and it works.

Thank you for your suggestions Xavier.

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

Time to create page: 0.131 seconds