Skip to main content
. 2015 Oct 5;16:745. doi: 10.1186/s12864-015-1944-z

Fig. 8.

Fig. 8

Assessment of BINM, direct PE and indirect PE scores on the Collins dataset. Interactions are ranked by corresponding scores. Performance of all methods is measured by plotting the number of top-ranking inferred direct interactions against the number of these interactions that are validated by reference sets. a HINT reference set, b Y2H reference set, c PCA reference set