Hi Nat, playing with fake data you know the answer. Real data are good too, although you have to keep in mind how its completeness will affect what you look at. I can't see how B-factors are relevant if you take a model from PDB: in this case the B-factors are what comes from PDB - real ones.
Pavel. On Wed, Jun 1, 2011 at 5:01 PM, Nat Echols <[email protected]>wrote: > On Wed, Jun 1, 2011 at 4:45 PM, Pavel Afonine <[email protected]> wrote: > >> to get feeling about how maps may look like at different resolutions do >> the following learning exercise: download a structure from PDB and compute >> Fcalc maps at different resolutions: >> >> phenix.fmodel model.pdb high_res=1 >> phenix.fmodel model.pdb high_res=2 >> phenix.fmodel model.pdb high_res=3 >> phenix.fmodel model.pdb high_res=4 >> phenix.fmodel model.pdb high_res=5 >> ... >> phenix.fmodel model.pdb high_res=10 >> phenix.fmodel model.pdb high_res=20 >> >> then load them in Coot and you will get your answer. >> > > But fake data will always look much better than real data - to start with, > the B-factors will be all wrong. > > There are actually a handful of PDB entries that are between 8.0 and 12.0 > and have experimental data deposited: > > 1H1K, 1PNS, 1PNX, 1VCR,1ZBB, 2QIJ, 2QZV, 3LVH, 3PCQ > > and 1VOQ, 1VOR, 1VOS, 1VOU, 1VOV, 1VOW, 1VOX, 1VOY, 1VOZ, 1VP0, which are > all the same structure split over multiple files. > > I'd suggest downloading a few of these and looking at the maps. > > -Nat > >
