The Biological General Repository for Interaction Datasets (BioGRID) is a database of physical and genetic interactions developed by Mike Tyers' group. It contains high-throughput (HTP) interaction data for Saccharomyces cerevisiae, Caenorhabditis elegans, Drosophila melanogaster, Homo sapiens, and other organisms. For yeast, BioGRID also contains many genetic and protein interactions curated from the primary literature. SGD has incorporated BioGRID's interaction data and will also continue to manually curate yeast interactions. What follows is a description of the BioGRID curation procedure along with some issues and details to keep in mind while curating this information.

Once you have logged in to the Interaction Management System (contact an admin for details), you will be able to add new interactions or correct existing ones for a particular paper. Clicking on one of the PubMed IDs in the high or low priority lists, or else searching on a PubMed ID will bring you to the curation page for that paper. We are keeping track of articles and abstracts that have been read. If you enter a PubMed ID that has not yet been curated, you may determine based on reading the abstract alone that the paper does not contain interaction data. If so, then mark the paper as 'Abstract' by selecting the relevant option displayed in the tool. This option should also be selected for reviews (even if the review doesn't have an abstract) because we are not curating reviews at this time. If you read an abstract and cannot determine whether or not the paper contains interaction data, then access the full text article. If you cannot obtain the full text article online, then select the 'Unavailable Full Text' option. We are keeping track of these papers in order to curate them at a later date. Similarly, if the paper refers to an organism other than S. cerevisiae (baker's yeast), then mark it as 'Not This Project' and select the relevant organism, if applicable, or 'Miscellaneous Curation' if the organism isn't listed. If, on the other hand, the paper does not contain any interactions, then mark it as finished by selecting the 'Full Text' option. However, if a paper has interaction data, then it is best to access the full-text article and curate the interaction information following the steps described below.