Hello!Here is what I've tried for my protein's Cterm which did not align with any of my templates:
1- 2D structure prediction using only algorithms with at least 75% overall accuracy as determined monthly by EVA (http://salilab.org/~eva/), and using a consensus of all results as my final prediction 2 - This confirmed that the Cterm domain was probably all alpha since all proteins of known structure in this family of proteins have an all-alpha Cterm 3 - I then used the following script (it's an old .top file - there may be a new example using Python in the Modeller FAQ) for Modeller to set restraints for those last residues that were not aligned against any of the templates:
INCLUDE SET OUTPUT_CONTROL = 1 1 1 1 1 SET ALNFILE = 'alignment.ali' SET KNOWNS = 'xxx' 'xxx' 'xxx' SET SEQUENCE = 'xxxxxx' SET ATOM_FILES_DIRECTORY = '/xxx/xxxxxxx/xxxxxx/xxxx/' SET STARTING_MODEL = 1 SET ENDING_MODEL = 20 CALL ROUTINE = 'model' STOP SUBROUTINE ROUTINE = 'special_restraints' SET ADD_RESTRAINTS = on MAKE_RESTRAINTS RESTRAINT_TYPE = 'alpha', RESIDUE_IDS = '461' '465' RETURN END_SUBROUTINE -Charlotte Habegger-Polomat Lucas Bleicher wrote:
--- Charlotte Habegger-Polomat <> a écrit :Andy, There may be a better answer but depending on thelength of the gaps where structure information is lacking, you could use 2D structure prediction programs for those regions and then impose certain restraints on Modeller. For example, the last 10 residues in myprotein of interest did not have anything to align with in the sequences of the 4 templates.That's an interesting approach - I always cut the terminals of my modeled sequences when I have nothing that align with it. If it's not asking much, could you provide an example of how to do this in Modeller? Thanks in advance, Lucas___________________________________________________________________________ Appel audio GRATUIT partout dans le monde avec le nouveau Yahoo! Messenger Téléchargez cette version sur http://fr.messenger.yahoo.com