Hi,

 

May be you would like to have a look to MODELLER's ALPHA restraints. They take the helical gaps into consideration very nicely. Prevents kinking and keeps the faces nicely.

 

Hope it helps.

 

br,

 

Vivek Sharma

 

On 2/2/06, Guillaume LETELLIER <guillaume.letellier@cea.fr> wrote:

Hi all

I am new to modeller and i'm currently trying to build a model of a GPCR
using the bovine rhodopsin structure as template.
I have notice that in the middle of one of a region corresponding to a
transmembrane segment, there is a very unambiguous gap of two residue in
the sequence alignment.

Two model of this receptor have already been published and in both
cases, authors have used sequence alignment without gaps, which means
that half the residues of this TM are clearly not correctly aligned with
the template.

I am confuse of how I should handle this gap to build my model.
Using a 2 residues helix-gapped alignment in modeller would result in a
model with
   - ugly coil in the middle of the helix
   - a switch from one helix face to the opposite of amino acids
   - a longer loop next to the TM (or maybe a longer helix ?)
On the contrary, ignoring the gap and using a wrong alignment result in
a  much better looking helix.

Should I deliberately ignore the sequence similarity to provide modeller
an alignment with no gap in the TM ?


Best regards,
Guillaume LETELLIER


_______________________________________________
modeller_usage mailing list
modeller_usage@salilab.org
http://salilab.org/mailman/listinfo/modeller_usage