I am trying to model MHC class I molecules and have come across a problem which appears to be related to the alignment. MHC molecules consist of three individual peptide chains and I have therefore specified my input sequence in the alignment file with the chain break character / (see alignment file below).
When modelled, this gives rise to a model with a serious flaw: It appears that the chain break character is interpreted by MODELLER as a kind of "dummy residue". The result is that the first chain is modelled nicely, but the second and third chains are shifted by one and two residues, respectively. This is a BIG problem, especially as I am interested in the exact interactions of the third chain (a small oligopeptide) with the rest of the molecule.
I have come up with a temporary "solution" (hopefully) where I simply delete the last residue of the first and second chain in order to make room for the "dummy residue". This produces nice models (which are lacking two residues), but I would like to be able to model the entire molecule.

This is my alignment file which is modelled by the model.top routine:
 

>P1;1hhi
structureX:1hhi: 1:A : 9:C : MHC I: HLA-A0201:
*
>P1;1A11
sequence:1A11: 1:A  : @:@  : MHC I: HLA-A1101:
GSHSMRYFYTSVSRPGRGEPRFIAVGYVDDTQFVRFDSDAASQRMEPRAP
WIEQEGPEYWDQETRNVKAQSQTDRVDLGTLRGYYNQSEDGSHTIQIMYG
CDVGPDGRFLRGYRQDAYDGKDYIALNEDLRSWTAADMAAQITKRKWEAA
HAAEQQRAYLEGRCVEWLRRYLENGKETLQRTDPPKTHMTHHPISDHEAT
LRCWALGFYPAEITLTWQRDGEDQTQDTELVETRPAGDGTFQKWAAVVVP
SGEEQRYTCHVQHEGLPKPLTLRWE/
MIQRTPKIQVYSRHPAENGKSNFLNCYVSGFHPSDIEVDLLKNGERIEKV
EHSDLSFSKDWSFYLLYYTEFTPTEKDEYACRVNHVTLSQPKIVKWDRDM
/GILGFVFTK

*

Many thanks.

Thomas
 

--------------------------------------------------------------------------- 
Thomas Blicher, Department of Chemistry,
Universitetsparken 5, Copenhagen University, 2100 Copenhagen Denmark.
phone: 45-35320280, fax: 45-35320299
E-mail: thomas@kage.ki.ku.dk