Hi Michael, I'm not sure what you mean with the 'partial structures' you have generated but I assume you have a reasonable model of your POI. I'd say that if the sequence identity between the homologous proteins/structures and your POI is high enough, it would give you a good idea of how the POI dimer is formed. However the interactions between the molecules in the dimer usually depend on very specific interactions (e.g. H-bonds, salt bridges) that are made by specific residues. Possibly, in the dimer interface regions there are different residues among the homologues, and this makes it difficult to predict how exactly the POI dimer will be formed.

If one of the homologues has a dimer structure available in the PDB, you could try to generate the similar dimer in PyMOL by superposition of two POI with that dimer. Also, there are websites that allow you to do protein-protein docking.

Hope this helps, good luck. Tjaard

Hi Michael,

There are several steps you have to do before you get a convincing result.

1. You have to run sequence alignment for your protein with the representatives from the protein family it belongs to. Use proteins with different sequence homology (90%, 80%, 70% up to 50%).

2. You have to identify all conserved regions and assign them to the proper functional (active site, binding site ...) and structural (dimerization) features of your protein.

3. You may have to remove some of proteins from your alignment if they do not follow the general pattern.

4. Find in literature solid evidences describing the parts of your protein involved in dimerization, and check if that is consistent with your highly conserved regions.

5. You can highlight those residues in PyMol and check where they are located. You can map them on the surface and check If they form patches.

6. Check the structure(s) you used to build your model if they are reported in PDB as a dimers.

7. You can superpose available homolog dimers from PDB with you complete model of your protein and check if the conserved regions (patches) match the dimerization surface.

8. Choose only those which match your conserved regions (see below).

In PyMol. Download your model (protein). Fetch the PDB code of the dimer structure. Align the dimer from PDB with your model. By default the program will align chain A of the PDB model. Download your protein once again with different name. Align your second molecule with chain B of the dimer from PDB. You should get a dimer of your protein. Check carefully for clashes. If clashes are severe (main chain atoms involved), then something is wrong with your model, or with the dimer.

If you are familiar with COOT from CCP4, you can do it easier and use stereo mode to make your analysis faster and better.

Download your model. Fetch the structure from PDB. Use SSM superpose with your model as the reference and PDB structure as a moving object (move a copy, not the actual model). You can choose a chain you want to superpose. Then superpose your protein (a moving copy) with the other chain of PDB model. After you finish, display only your original chain and the moved copy of it. These two molecules should build your dimer.

9. Check all available dimer from PDB.

Possible problems... Dimers in the PDB do not represent the true biological assembly. Run PISA and check if there are different possible dimers for the deposited structures. Generate them (there is an option in PyMol) and visualize them. Check if the "new" dimers are the true ones.

10. You may need to run some optimizations (MD) if you found a true dimer but could not avoid clashes.

11. After you get your dimer check again if conserved residues are involved in the dimerization. That will be your final check and you have do decide if your model makes sense based on your analysis.

It looks complicated, but I am sure you spent enough time with graphics and can do it.

Good luck!

Best regards,

George.

Thank you both so much!