On 03/27/2010 07:26 PM, Song, Hyundeok (songhk) wrote:
I generated the dimer by Modeller. Fasta sequence number of model(400) is 70 less than Fasta sequence number of template(470). Sequence identity between template and model is 24%.
If I understand you correctly, you have a loop that is about 70 residues long. There's no way Modeller will be able to accurately model that (the limit is about 15 residues). I can't tell for sure without seeing your alignment file though.
after selecting the template, I aligned them, and then I did loop modeling. (I attached the code at the bottom.) Finally I got final pdb structure, but when I checked my model, I found the below warning message. "Unusual bond length, unusual bond angle, Chirality deviations "
Looks from your script like you built 50 loops, so you can check the other 49 PDB files to see if another loop looks better. But if your loops are really 70 residues long, Modeller won't be able to help you. Your best bet there is simply to remove those 70 residues from your model sequence, and build a model of part of the protein.
Ben Webb, Modeller Caretaker
--
http://www.salilab.org/modeller/
Modeller mail list: http://salilab.org/mailman/listinfo/modeller_usage