Various subsets of the DIP interaction data are available in a variety of formats: PSI-MI (versions 1.0 and 2.5), MITAB2.5 and our legacy XIN format. The sequences of the proteins participating in DIP interactions are provided in FASTA format. In addition, some more detailed information about the interactions of S.cerevisiae proteins are also provided in the form of tab-delimited files.

Please note that, in general, the interaction data we offer should be customized before they are used to address any particular problem. This includes, but is not limited to, making informed choices about:

• Acceptable interaction types. DIP, in line with other interaction databases, relies on experiments that might provide only partial information about protein-protein interactions. In particular, some of the experiments might only demonstrate that a pair of proteins is a part of a larger complex but without actually making a direct contact. Depending on the addressed problem, such interactions might have to be excluded from further analysis. Please, use the information about the interaction type demonstrated by each experiment and make an informed choice.
  
  
• Acceptable level of protein ambiguity. DIP, in line with other interaction databases, relies on experiments that might provide only partial information about the identities of the proteins reported in the protein-protein interaction experiments. In particular, some of the experiments might not distinguish between various protein isoforms that might have arisen as a rerult of alternative splicing events. In case such ambiguity is relevant to the problem on hand our original IMEx curation records should be consulted in order to identify experiments with unambigiously identified proteins.

The above issues are not specific to the datasets provided by DIP but reflect the very nature of the experimental data that the knowledge of protein-protein interactions is based upon. Therefore they should be taken into consideration when utilizing any other protein-protein interaction datasets made available by any other interaction data provider.

The standard data sets provide the basic information on all interactions within DIP including these inferred through the spoke expanson of the multi-protein associations identified in affinity-based experiements but no details on the interacting molecules.

DIP-IMEx data sets contain records curated according to the IMEx Curation Rules. They provide information on the experimentally determined details of the interactions. These might include interacting regions of the proteins, mutations affecting interaction and relevant posttranslational modifications.