On Tue, Nov 29, 2011, thomas.fox.boehringer-ingelheim.com wrote:
>
> I would like to report on a subtle problem how leap handles protein systems
> which contain several chains.
Thanks for the example. It looks like the CONECT cards create cross-links
between the protein chains--do you know what these are? Have you ever tried
to remove the CONECT cards and recreate those bonds by hand?
> Now, if I remove the comment sign before the solvateoct command, and do
> the exact same thing, the first TER card in my pdb file disappears: in
> the leap-generated prot_1.pdb the TER card between ILE28 and ILE29 is
> present, however, when I generate a pdb file with ambpdb -p prot_1.top <
> prot_1.crd > prot_bug.pdb this TER card is missing.
I think this is because the cross links make more than one chain into a single
"molecule". If the chains that are linked are together in the pdb sequence,
things might be OK except for the absence of TER cards in the ambpdb output.
If the linked chains are disjoint, I'm not sure what all happens.
It's pretty simple to change ambpdb to put TER cards where you expect them:
Go to line 436 of ambpdb.f90 (or thereabouts) and find the lines:
if (ifbox .gt. 0) then
! --- prmtop file will have information needed to generate where the
! TER cards in the output should be:
Change the first line above to "if (.false.) then", and recompile. Warning:
this is untested, but it looks like it should work.
However, we still need to address the issue that the ATOMS_PER_MOLECULE
section seems to have bad entries....
...dac
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Tue Nov 29 2011 - 09:00:04 PST
