*** For details on how to be removed from this list visit the *** *** CCP4 home page http://www.ccp4.ac.uk ***
A classical example for procheck verboten residue that is in fact structurally conserved in serin hydrolases is Ser119, stressed but allowable in the newer distributions. image at http://ruppweb.dyndns.org/new_images/ser119.gif A. A. Azizi et al. Cancer Letters, online (2006). A classical example why it does not matter - just pick the right journal according to Kleywegt's Rule (the worse the structure the higher impact journal should be considered) - is in http://ruppweb.dyndns.org/new_images/whocares.gif Cheers, BR Prosecutor, Societas Eruditorum -----Original Message----- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Kevin Cowtan Sent: Wednesday, March 15, 2006 4:43 AM To: Frederico Moraes Ferreira Cc: CCP4 bulletin board Subject: Re: [ccp4bb]: assesing data and model *** For details on how to be removed from this list visit the *** *** CCP4 home page http://www.ccp4.ac.uk *** Grab a copy of '8abp' or '1ajr' from the PDB - both well refined high resolution structures. Residue 89 in '8abp' and 296 in '1ajr' are in the gamma turn conformation. If your validation program thinks this is a bad residue, it needs new Ramachandran data. MolProbity and Coot both use the Richardson's data for their Ramachandran plots, and so give this residue as fine. I tried WhatCheck on this website: http://swift.cmbi.kun.nl/wiw-cgi//ChecksCGI.py It flags both residues as having poor phi/psi, so I think that WhatCheck at least could do with a new Ramachandran plot. I also tried the EDS on 1ajr - this uses either moleman2 or WhatCheck. Both of these object to residue 296. The Ramachandran plot within AstexViewer however, launched from the EDS, shows this residue as good, so that's based on newer data. Kevin Frederico Moraes Ferreira wrote: > *** For details on how to be removed from this list visit the *** > *** CCP4 home page http://www.ccp4.ac.uk *** > > > hi Kevin, > what about whatcheck? I'm currently using the pretty old linux version > from 2003, which has always run into problems with water molecules > "Error: Water molecules without hydrogen bonds.", even when they are > clearly ok. has whatcheck distribution also frozen? > thanks in advance, > fred > > Kevin Cowtan wrote: > >> *** For details on how to be removed from this list visit the *** >> *** CCP4 home page http://www.ccp4.ac.uk *** >> >> >> Kevin Cowtan wrote: >> >>> Run procheck, but ignore the Ramachandran plot, which is based on >>> out-of-date information and wrong. If referees give you trouble, >>> ccp4bb has a number of authoritative scientists who will back you up. >> >> >> >> Since it is generating queries, let me expand on this. >> >> While procheck has done sterling service over the years, its >> Ramachandran plot is out of date. This has been demonstrated by the >> work of the Richardsons, see this web page for a tiny fragment of >> their outstanding contribution to the field: >> http://kinemage.biochem.duke.edu/validation/model.html >> >> For details of the problem, see this paper: >> http://kinemage.biochem.duke.edu/ftpsite/pub/datasets/pdfs/03PROT50_4 >> 37.StructValid.pdf >> >> In particular the figure at the top of page 2 and the text at the >> bottom of page 2 and top of page 3. The procheck plot marks the gamma >> turn region around phi=+75, psi=-60 as forbidden, when in fact it >> should be allowed. >> >> Unfortunately many people and referees use procheck uncritically and >> deliberately introduce errors into their structure before deposition >> to make the procheck output look better. >> >> Since procheck is no longer actively maintained by the author, there >> has been a suggestion that CCP4, as perhaps the primary distributor >> of procheck, should unilaterally update procheck to use the >> Richarsons' data. I for one would support this. >> > > >
