[modeller_usage] Modelling of multi-domain protein from oligomeric template

[Chugunov Anton <>]

Dear modellers,

I'm building a model of ionic channel that consists of four pseudo-homologous domains forming a pore between all of them. My template is Kv1.2 channel that is organized similarly, but it is homo-tetramer rather than a single-chain four-domain protein. The template is a four-chain pdb file (a "biological assembly" from pdb-site) (A-D), where all protomers are already oriented to form a pore.

My alignment is like:

 

>P1;2r9r_assembly1

structureX:2r9r_assembly:32    :A:  417:D::::

AAAAAA------------------/

------BBBBBB------------/

------------CCCCCC------/

------------------DDDDDD*

 

>P1;Channel
sequence:Channel:1    : :XXX: ::::

DOM-1_DOM-2_DOM-3_DOM-4_*

 

Here, AA, BB, CC, DD is the same sequence of the whole Kv1.2, aligned with four domains of the channel (DOM1-4) .

 

 

What I expect, is a single-chain model that consists of four domains, each built from respective template's chain according to the alignment. What I get (as far as I see from the *.ini file, since no models output due to high energy-function) is the first domain followed by a huge non-structured loop -- there's no structure assignment for the last three domains. Additionally, to use the large template, I use "max_ca_ca_distance = 150" option.

Can anyone help me with the alignment -- there's obviously something wrong there?

 

Anton.