ORCS curators capture experimental evidence supporting CRISPR screens from the primary literature. Our aim is to curate all genome-wide screens in the literature for humans and some model organisms, in addition to some targeted screens we may come across. Screens discussed in reviews or re-analyzed by new analysis methods are not recorded. The BioGRID ORCS curation workflow can be viewed here.

ORCS uses gene identifiers to describe all screen participants. For each screen the gene identifiers reported by the authors are used, or if gene names are reported they are mapped to relevant gene IDs whenever possible. Ambiguities or names that cannot be mapped are reviewed and included as published if no further identification can be determined. Scores for each gene provided by the author using their analysis method of choice are included. Thresholds provided by the author for determining “hit” genes are applied. If authors do not explicitly state a threshold, then commonly accepted thresholds or those established in a reference article for that analysis method may be applied.

Metadata for each screen is also curated including the cell line the experiment is performed in and other relevant experimental details such as duration, any selection conditions the cells may be exposed to such as the presence of a chemical, toxin or virus, the format of the screen including pooled or arrayed and whether it is a positive, negative or phenotype selection screen. Additional CRISPR specific information captured includes the name of the gRNA library and the endonuclease enzyme used. Curators also capture the phenotype used to sort cells. Specific information on the vocabularies and ontologies used for curation of these fields can be found in the ORCS Curation Guide.