Explanation of datasets
AI-1MAIN (main screen): A systematic dataset containing 5,664 protein-protein interactions among 2,661 Arabidopsis proteins. This dataset was produced by screening all pair-wise combinations in the Space-1 search space twice. This dataset should be used for network and other global analyses.
AI-1REPEAT (repeat screen): A dataset containing 1,066 protein-protein interactions among 673 Arabidopsis proteins. This dataset was produced by screening the indicated subspace 4 times in addition to the 2 repeats done to produce AI-1MAIN.
AI-1: The union of all interactions in AI-1MAIN and AI-1REPEAT, containing 6,205 interactions among 2,774 Arabidopsis proteins. This dataset should be used for manual inspection and the development of specific hypotheses for biological follow-up.
LCI: 5,270 interactions among Arabidopsis proteins reported by IntAct, TAIR and BioGrid as of April 2010. These interactions were assembled by extracting interaction information from the literature and contain both direct protein-protein interactions and indirect co-complex associations.
LCIBINARY: The subset of 4,252 interactions in LCI that are likely direct, binary interactions based on the experimental evidence supporting them.
LCICORE: The subset of interactions in LCIBINARY that are supported by >1 experimental method, or >1 paper.
Reference sets: Arabidopsis thaliana Positive Reference Set (AtPRS) and Random Reference Set (AtRRS) for used for assay calibration and determination of assay sensitivity. AtPRS contains 118 well-documented protein interaction pairs that were randomly sampled from LCICORE and manually re-curated by experts, and ORF reagents for both partners were available in Space-1. AtRRS contains 146 random protein pairs sampled from Space-1 and serves as a negative control.
Space-1: The 2-dimensional space of all pair-wise combinations tested among 8,583 ORFs encoded by 8,429 TAIR7 genomic loci.
For additional questions regarding the Arabidopsis interactome datasets, please contact Michael Calderwood.