Skip to main content


Figure 8 | Journal of Cheminformatics

Figure 8

From: Are phylogenetic trees suitable for chemogenomics analyses of bioactivity data sets: the importance of shared active compounds and choosing a suitable data embedding method, as exemplified on Kinases

Figure 8

Average trend of SAR similarity based on shared active compounds for all kinases (181) showing the expected negative relationship between SAR similarity and bioactivity distance (excluding kinases from outlier group 1). Error bars indicate 1 standard deviation. The number of shared active compounds was normalized by the total number of active compounds in the variable kinase. Interestingly, the optimum in SAR similarity for kinases is not at the lowest distances: kinases show a lower degree of SAR similarity at distances smaller than 0.03 than the optimum which is at a distance of approximately 0.03. Furthermore, variance in shared active compounds decreases with increasing bioactivity distance (after distance is approximately 0.03), suggesting that it is easier to predict SAR similarity for distant kinases than for closely related kinases.

Back to article page