I am preparing receptor files for docking.
I have a crystal structure which is a homodimer. It has some residues missing in 1of the monomer but resolved in the other. It also has some residues absent in both monomers( 10 in the N-terminus & 3 residues at C-terminus). I modelled the protein's missing portions with Modeller and the resulting structure has the following problems :
1. There is a 3-10 helix in 1 of the loops in the original PDB 's A chain but this is not being modelled by Modeller in the B chain in which it is missing .....
I checked the documentation of Modeller but could not find anything about modelling 3-10 helices .....
2. The model is poor in quality compared to original PDB...There are 5 ramachandran outliers while the original PDB has none.
I used Molprobity to fix the model but still I am getting the same 5 outliers in Ramachandran Plot.
How to improve the quality of the model ?
Is there any tool to model 3-10 helices ?