Skip to main content
. 2023 Mar 20;19(5):e11325. doi: 10.15252/msb.202211325

Figure 6. Comparison of INDRA‐assembled mechanisms with a curated resource, BioGRID.

Figure 6

  1. Number of INDRA Statements representing PPIs (i.e., complex formation between two human proteins) grouped into bins by their belief score (as determined by a random forest belief model), differentiating whether the PPI represented by the Statement appears in BioGRID (orange) or not (blue).
  2. Fraction of INDRA Statements representing PPIs that appear in BioGRID grouped into bins by their belief score. A gray dashed line shows the expected fraction of correct Statements in each belief bin. The space between the gray and blue lines (i.e., between the expected fraction of correct Statements in each bin and the fraction of Statements that appear in BioGRID) represents an estimate of the set of correct Statements missing from BioGRID.
  3. Plot showing estimated curation yield if Statements were reviewed by decreasing belief score for inclusion into a curated resource. The blue line plots the number of correct Statements expected to be found as a function of the number of Statements reviewed, with green and pink dashed lines serving as guides showing 100% return (i.e., every reviewed Statement is correct) and 25% return (i.e., 1 out of 4 reviewed Statements is correct).