Coral Restrains chain ID problem

Interactive modelling (MASSHA, SASpy) and global minimization programs (SASREF, BUNCH, CORAL, GLOBSYMM)
Post Reply
Message
Author
eM
Guest
Posts: 2
Joined: 2017.03.10 15:57

Coral Restrains chain ID problem

#1 Post by eM » 2017.03.10 16:47

Hey,

I have recently started to use Coral and I just can't set restrains right.

There are 3 pdb files - core_a.pdb and flexible_a.pdb make together one protein and core_b.pdb is another one. I am trying to set K140 from one protein (A) and K113 from another one (B) to be about 12 A apart. However it seems Coral is taking all my pdb files as one chain (see bellow). How can I set it correctly?

Code: Select all

 Contacts conditions file name .......................... : dist.cnd
Condition #  1: Distance  12.000
   Between chain #  1, Residues from GLY   140 to GLY   140
       and chain #  1, Residues from GLY   378 to GLY   378
 Contacts conditions penalty ............................ : 33.54
 Contacts penalty weight ................................ : 10.00
dist.cnd:

Code: Select all

dist 12.0
1 140 140 2 113 113
configuration:

Code: Select all

NTER 8
core_a.pdb
LINK 14
flexible_a.pdb
CTER 20
core_b.pdb
CTER 10

SaxsMax
Administrator
Posts: 343
Joined: 2007.08.03 19:16

Re: Coral Restrains chain ID problem

#2 Post by SaxsMax » 2017.03.14 11:31

Hi,

Internally Coral puts all the input proteins into one "superchain",
therefore the reported numbering is different from the dist.cnd.
Have you checked that the conbdition is fulfilled in the resulting model?

eM
Guest
Posts: 2
Joined: 2017.03.10 15:57

Re: Coral Restrains chain ID problem

#3 Post by eM » 2017.03.17 11:53

I have made 2 models with restrains and sadly no - the gap is about 40 A. However without restrains the models are much better.

Nyshae
Active member
Posts: 24
Joined: 2015.11.03 14:38

Re: Coral Restrains chain ID problem

#4 Post by Nyshae » 2017.06.07 10:41

Hi,

I thought I would just resuscitate this post because I have a similar issue... My setup is the following: one protein = domain A + short linker + domain B. The protein forms a dimer and we do know (from crystallographic data) that the domains A are dimerizing. I want to set CORAL with the pdbs of domains A and B and make it model the position of domain B in each case with the 8 AA linker.
So far, I am using a .con file with the information for the linker length (functional) and a P2 symmetry (ok). But the constraints file doesn't work...
The file is the following:
dist 5.0
1 41 62 2 262 238
dist 5.0
2 41 62 1 262 238

But I get this error:
- E -- Wrong format of conditions file
1 41 62 2 262 238
1 41 62 2 262 238

I want basically that the domains A are side by side in an anti-parallel way (aka amino acids 41/62 of domain A are next to amino acids 262/238 of domains A'). But I have no idea what is the chain numbers CORAL attributed to my bits... And I don't know if I can setup this constraints for the dimer interface only (aka not the protein monomer). Should I use different input pdbs? Any idea?
Many thanks!

SaxsMax
Administrator
Posts: 343
Joined: 2007.08.03 19:16

Re: Coral Restrains chain ID problem

#5 Post by SaxsMax » 2017.08.24 13:24

You could try the following conditions:

dist 5.0
1 41 41 2 262 262
dist 5.0
1 62 62 2 238 238

Nyshae
Active member
Posts: 24
Joined: 2015.11.03 14:38

Re: Coral Restrains chain ID problem

#6 Post by Nyshae » 2017.08.31 14:49

Hi,
Thanks for the tip! I managed since then to find out a working set of constraints (at least the program accepts it) that do resemble yours.
Now I have another problem... : I basically have a protein dimer which is in an anti-parallel way. So I input P2 symmetry and a set of constraints to have each monomer C-ter close to the other N-ter. But it seems that it's not working: in resulting PDB each monomer is very far away from the other (like 150 Å).
Any idea why? ^^'

SaxsMax
Administrator
Posts: 343
Joined: 2007.08.03 19:16

Re: Coral Restrains chain ID problem

#7 Post by SaxsMax » 2017.08.31 15:03

It might be that the dimer is not consistent with the SAXS data.

What is the resulting chi^2 and the contact penalty (in the header of the generated PDB file) ?

Nyshae
Active member
Posts: 24
Joined: 2015.11.03 14:38

Re: Coral Restrains chain ID problem

#8 Post by Nyshae » 2017.09.25 11:36

Sorry about the time lapse, I don't come so often ^^'

best Chi2 values: 3,6643
contact penalty: 0,00000

SaxsMax
Administrator
Posts: 343
Joined: 2007.08.03 19:16

Re: Coral Restrains chain ID problem

#9 Post by SaxsMax » 2017.09.25 12:54

Given that the penalty is zero, the contacts restraints should be fulfilled.
Could you post the input files to check?

Nyshae
Active member
Posts: 24
Joined: 2015.11.03 14:38

Re: Coral Restrains chain ID problem

#10 Post by Nyshae » 2017.09.25 14:31

Actually, my apologies, I did many trials and got lost in the file piles, the actual constraint penalty is around 27-29 (the 0 one is the one where I tried without constraints).

Post Reply