I have been reading the tutorials and the 2007 McCoy (Acta D) paper about how to use the full-featured Phaser-MR tool (with the Phenix GUI) to do MR. I am struggling with how to define the Ensembles, and the Search procedure. Specifically, after setting up the Ensembles and the Search procedure, as I'd deduced that they should be done, I ran the script and get the following message:

>>Search request requires more scattering than defined in composition. Composition increased to accommodate search components.<<

I guess that I have not understood how to set up the multi-component search to do what I want for it to do: I'd wanted the tool to search for the best placement of individual domains, given a set of diffraction data and a collection/combination of pdb files that contains coordinates for individual domains.

This is the context:

The diffraction data is from a crystal of a three-domain protein. The first model I used for MR was a crystal structure of the largest of these domains, which had been expressed and purified alone. With this one domain as a search model (using the single-component Phaser-MR) I obtained a solution without a problem, but now have density into which I will need to place the other two domains. 

The second-largest domain has also been crystallized (also on its own). Its structure (as a monomer or dimer) is available from different organisms, some dimers engage different inter-chain interfaces, and some have ligands bound at different interfaces. There does exist a crystal structure for the three-domain protein, and its biological assembly is 4. I have reason to think that this new crystal is of this same protein in a different configuration - that is, in this crystal structure the three domains are oriented differently to one another than they were in the first 3-domain structure. In solution this protein can be a dimer or tetramer.

The ASU for this dataset is 2 chains, and I assume that each chain has all three domains intact.

This diffraction data is of relatively poor resolution (~3.8A).

So, I configured the Phaser-MR run as follows, using the AP2 and Beta-blip case studies as examples:

- In the 'Ensembles' tab I created 17 different ensembles, each with a different combination of single-domain structures (or with just the single largest domain, as in the original MR single-component search). 

- In the 'ASU contents' tab I provided the sequence of the full-length (3-domain) protein, and set 'Number of copies' to 2.

- In the 'Search strategy' tab I added each of the 17 Ensembles as a separate 'Component' with 'Copies to search for' set to 2. This seems redundant to what I'd put in to the 'Ensembles' tab.

I thought that a 'component' meant a single pdb file representing a single chain or domain (in this case) while an 'Ensemble' meant a collection of components so that, e.g., a 3-domain chain could be described as an ensemble of 3 components.

Do you have advice on how to set up this search correctly?

Regards,

Emily.

--