Feed it common substructures and see what happens...
On Mon, 5 Mar 2007, Nat Echols wrote:
> I'm not sure what "in desperation" means. You're trying to solve a
> structure, so all options are open, right? They don't have to be elegant,
> it just has to work. :)
Yes, but it's not worth wasting time and resources on something that is
guaranteed to fail. In this case, it's a 50kD protein that doesn't have
any homologues in the PDB (the best match is ~25% over maybe 70 residues),
and PHYRE generates a model that probably approximates the topology
reasonably well but doesn't have a very good e-value and certainly doesn't
resemble the sequence of the original model. This is far worse than any
successful solution I've heard of, and given the number of times a 100%
identical model doesn't provide a solution, I'm not inclined to waste my
time on theoretical models or PDB-wide searches.
In this case, "desperation" means the crystals are nearly impossible to
reproduce, and only one data set was ever collected. (We could try phasing
using every structure from E. coli or at least 25% identical to an E. coli
protein, in case the crystals were of a contaminant. But if the protein
was what we wanted, this won't work.)
|