[ccp4bb] Validation Question

[Simon Kolstoe]

Dear ccp4bb,

I have heard from many people that molprobity is the gold standard as  
far as validation is concerned. As I am just about to deposit a  
structure I therefore figured it was best to use molprobity in guiding  
the final stages of my refinement. I do not have enough data to  
justify adding hydrogens in my refinement, however as molprobity  
suggests adding hydrogens for validation purposes I obediently did so  
and (surprise surprise) found that I had a number of clashes  
(clashcore 17.6). Just as an experiment I re-ran refmac including  
hydrogens and this time my clash score was much lower and apparently  
acceptable (6.27), whilst my R factors were pretty much identical to  
before. So my question is this - do I ignore the normal obs:params  
calculation and just refine my structure with hydrogens (resolution  
1.9, 99.9% complete) as according to molprobity this gives me a  
"better" structure, or do I just ignore the clashscore targets in  
molprobity based upon not having enough data to justify refining this  
component?

Thanks,

Simon