CAPRI is a community wide experiment to assess the capacity of protein docking methods to predict protein-protein interactions. Groups participate in rounds by submitting blind structure predictions for protein-protein target complexes based on known structure of the component proteins. CAPRI aims to stimulate research in protein-protein docking and to act as a proving ground to evaluate prediction methods and assess their reliability. The predictions are data submitted to the CAPRI Management, which will be assessed and published as they see fit. Participants are free to disagree with any conclusions, but submission of predictions explicitly waives all potential claims against CAPRI Management OR THEIR STAFF OR CONSULTANTS, paid or unpaid.
CAPRI relies on the generosity of experimentalists willing to communicate unpublished atomic coordinates on a confidential basis. Targets are generally soon-to-be-released protein structures where the component structures must be known independently. Participating in CAPRI gives access to this confidential information. Scientists in academia and in companies are an integral part of the protein structure scientific community and all groups are welcome provided confidentiality can be guaranteed.
Each group submitting a target to CAPRI is recognised as the owner of all information relating to a target and as such may release information concerning a target when they choose. It is requested that owners of target details would advise the CAPRI management team when information is released so predictions should be submitted prior to such release. Information provided by groups submitting targets will remain confidential until it is released by these groups in publications or in the PDB. This date will then appear on the CAPRI Web-site. The scope of confidential information consists of atomic coordinates and all information concerning the target complex including the component sequences, the component taxonomy information and the information that the target even exists as a complex between known components.
For all targets and the participant predictions, prior to publication of the structures, participants may not report to third parties any formation about targets, this includes the size of each component, whether the components were bound, unbound, or homology modelled, and whether or not there were symmetry considerations. The identity of the molecules and the identity of any structures or sequences used in modelling, matching, or scoring must not be divulged before the coordinates are officially released. This includes pictures of the target and any figures or tables showing any structural, chemical, or biological information about the target.
Participation in a CAPRI round is by registration and at the start date entrants will be supplied with password access to the component PDB entries as a modified PDB ensemble. Registration will require entrants to sign a statement that they acknowledge the receipt of confidential information to be used only in the context of the CAPRI experiment and recognize their personal and professional obligation to keep it confidential until it is published or the original authors give CAPRI management written permission to release the information. Target coordinates would not be released to a participant until we actually had their signature. Target prediction coordinate sets will be made freely available on the www in an unrestricted manner once the target owner's have given their permission.
The assessment work is carried out to the best of the assessment team's ability by procedures agreed with the CAPRI management team and outlined for each Round on the appropriate web page. The assessors understand that there may be other ways to evaluate predictions however the methods used are fixed for each round. Assessment is open to scrutiny of the scientific community. Once the results are released, participants can make any use of them they wish, but cannot state or imply that CAPRI endorses one method over another; it only reports a ranking of performance of particular methods in the hands of specific groups on those targets considered.
Use of Evaluated Results in Publicity:
CAPRI is not a competition. Experience has shown that different prediction methods have performed better in different types of protein-protein complexes and research workers and commercial software vendors should feel rewarded with the success they achieve. The assessment results are the property of the Assessment team and of the CAPRI management group and use of this material in press statements or advertising requires permission from the assessors.
Click to view details of past CAPRI rounds: