Figure 2

Score-ppv plots of individual datasets. On the x-axis is the score for that dataset, on the y-axis the ppv. The ppv was calculated in all score intervals with bin-width 0.025. Red lines indicate ppv on the protein complex reference set, being the number of true positives in the complex reference set divided by the number of true positives and false positives in both reference sets. Blue lines indicate the ppv on the metabolic reference set, being the number of true positives in the metabolic reference set divided by the number of true positives and false positives in both reference sets. (a) Correlated mRNA expression (CoExp). (b) Shared binding of transcription factors (ChIP-chip). (c) Co-regulation (ChIP-chip*CoExp). (d) Conserved co-expression between four species (CoExp4Sp). (e) Conserved co-expression between two species (CoExp2Sp). (f) Paralogous conserved co-expression (CoExpPar). (g) Gene neighborhood conservation (GenNeigh). (h) Correlated phylogenetic profiles (CoOcc). (i) Shared genetic interactions (GenInt). (j) Yeast-two-hybrid (Y2H). (k) TAP-tag purifications (Gavin et al. [3]). (l) TAP-tag purifications (Krogan et al. [4]). For (k, l) the protein pairs that are never co-purified and thus have a SA score of 0 are in bin 0.2.