Figure 2
From: Predicting RNA-Protein Interactions Using Only Sequence Information
A. Predicted interactions using classifiers trained on the RPI2241 dataset. Among 254 known interactions, RPISeq-RF and RPISeq-SVM classifiers correctly predicted all except 5 and 2 edges, respectively. A protein hub, highlighted in yellow, shows interactions of a helicase (SEN1) with several snoRNAs. One of several RNA hubs, highlighted in purple, illustrates interactions of an snRNA (u4560) with various Sm-like proteins in the LSM complex.
B. Predicted interactions using classifiers trained on RPI369 dataset. Among 254 known interactions, RPISeq-RF classifier correctly predicted 168 (66%) and RPISeq-SVM correctly predicted 226 (89%). A protein hub highlighted in yellow, shows interactions of a helicase (SEN1) with 8 snoRNAs. One of several RNA hubs, highlighted in purple, illustrates interactions of an snRNA (u4560) with various Sm-like proteins in the LSM complex.
C. An enlarged view of the protein (SEN1) and RNA (snRNA) hubs described in B. above. Edges are labelled with the interaction probabilities predicted by RPISeq-RF (left) and RPISeq-SVM (right) classifiers, providing estimates of the relative pairwise interaction propensities.