In the 90s morphing of two unrelated images was popular and mostly it was used for entertainment purposes. For example: the famous video of Michael Jackson’s pop hit “Black or White”.
This morphing method was also used to analyze changes in protein motions, like in domain rearrangement. A popular webserver, where you can get an animated gif of your protein’s motion (assuming you have two distinct conformations), is the Morph server (http://www2.molmovdb.org/) from Gerstein’s Lab. In many cases this gave us insight of how the protein could dynamically change from one form to another.
The change in structural forms of a protein is not a trivial problem. We would need to generate ensembles of protein structures for many purposes. 1) Understand conformational transition paths, 2) Generating more realistic receptors for docking 3) in turn understand the flexible and rigid parts of the protein, and few other applications.
Till now, one could use Normal mode analysis and Molecular Dynamics methods to generate ensemble. It is here that ConTemplate tries to bring in fresh perspective to generate an ensemble of structures.
ConTemplate mines the PDB for existing structures and gives the user a set of possible conformations. The main presumptions are that for any given PDB structure, it has more than one available structure, and there are additional conformations available for proteins that undergo major conformational changes.
For the dataset created for ConTemplate the maximum RMSD between two structures of the same protein is 5 Angstroms. 69.2% of the proteins have less than 1 Angstroms RMSD. Thus, the method uses an interesting three-step process:
- using the query it searches for structural equivalents using GESAMT aligner. Here using the structural alignment sequence alignments are generated.
- it runs BLAST to identify additional conformations for all structural equivalents obtained in step 1. A representative template is identified
- Finally, Modeller is used to build model structures using this template in various conformations.
The advantage of ConTemplate is that it yields a more relevant set of conformations for the query protein. I tried running a query to the server and I would say that I got some interesting results. Screenshot below:
Narunsky A, Nepomnyachiy S, Ashkenazy H, Kolodny R, & Ben-Tal N (2015). ConTemplate Suggests Possible Alternative Conformations for a Query Protein of Known Structure. Structure (London, England : 1993), 23 (11), 2162-70 PMID: 26455800