how about this scenario:
the structure really is as published, or very like it, but upon
refinement, R-factors were high, other indicators were dodgy etc. so
the authors were afraid to publish as is and made up a dataset to
support their structure - this would be a bit less bad.
let's see how the authors, Nature (and the other journals) respond.
As some remarked, this is important, because if we let it pass, it
could damage the prestige of protein crystallography. I think, as the
crystallographic community, we should quickly act to put proper
controls in place like submission of raw images etc.
Mark
