Skip to main content
Figure 5 | Journal of Cheminformatics

Figure 5

From: Proteochemometric modelling coupled to in silico target prediction: an integrated approach for the simultaneous prediction of polypharmacology and binding affinity/potency of small molecules

Figure 5

Compounds predicted to interact with DHFR by the target prediction algorithm, and predicted by the PCM model to have a pIC50 value higher than 7 pIC50 units. Compound IDs correspond to the TCMDC identifier given in the original dataset. The 23 compounds for which the IDs are accompanied by an upward-pointing arrow were identified by the two methods. The two compounds predicted to have a pIC50 value higher than 7 by the PCM model, but not predicted to interact with DHFR by the target prediction algorithm, are accompanied by a downward-pointing arrow. The 23 compounds predicted to be high-affinity DHFR inhibitors (upward-pointing arrows) share a common scaffold: a 5-methylpyrido[2,3-d]pyrimidine-2,4-diamine ring with an aryl substituent in the 6-position. Overall, it can be seen that these data indicate a high agreement between the target prediction algorithm and the PCM model to identify high-affinity DHFR inhibitors.

Back to article page