I am also in favor of two versions of the pdb: one archive version with all models as originally deposited including retracted and corrected versions, which are useful for educational purposes, and a curated version with only models that meet a minimum of validation criteria, including credible ligand density. Whether crowd-perfected or original submissions that pass the validation criteria does not matter to me. Herman
Von: CCP4 bulletin board [mailto:CCP4BB@JISCMAIL.AC.UK] Im Auftrag von Pavel Afonine Gesendet: Freitag, 16. Mai 2014 07:19 An: CCP4BB@JISCMAIL.AC.UK Betreff: Re: [ccp4bb] PDB passes 100,000 structure milestone What about structures that are obviously wrong based on inspection of the density, but no one has bothered to challenge yet? The TWILIGHT database helps some, if that counts, but it doesn't catch everything. How about this utopia.. Imagine PDB has two versions: one is the original data and model deposited as is, and never ever changed no matter what. Another version is a curated one obtained in a quest-like way: anyone can take an original entry, improve it and deposit (into the curated version) with his/her name tag on it. And of course anyone can take and update that improved entry and re-deposit it again with his/her name tag, etc. If desired one could keep track of all the revisions, like in svn or so. Sounds like a sport with an element of public service that might yield crowd-perfected models -:) ! Pavel
