Figure 5

Statistical comparison of genetic interaction sets. (a) Genetic buffering and genome-wide TF-DNA binding data (Workman dataset) is taken for evidence of putative regulatory relationships among TFs. R gives the fold enrichment defined as a ratio of frequencies relative to a random model, and P the probability of observing a ratio of equal or greater value by chance. Boldface is used to indicate statistically significant enrichments (P < 0.05) within the genetic interaction sets relative to the frequency (hits/N) of interactions within the Workman data. Enrichment is analyzed in four categories: (1) Genetic buffering of one TF by another, (2) Direct binding of one TF to the gene encoding another, (3) TF pairs that bind to the same target gene(s), (4) interactions supported by any of the categories (1)-(3). The number of hits in the Workman dataset among the 316 pairs tested for each category are 26, 10, 46 and 77, respectively. (b) Comparison of true- and false-positive rates associated with each methodology. (c) The predictive value of the fitness- and sensitivity-based methods at different P value thresholds. The predictive value is defined as the fraction of correctly identified interactions.