Case Study - Angiotensin II Blocker

Scenario: We want to study the effect of angiotensin II antagonists. The question arises if there are different structure classes that can act as antagonists. 

We selected 96 compounds from MDDR by searching in the activity filed with "like %angiotensin II%". For these 96 compounds we predicted about 58 activities using PASS. We clustered the resulting EXCEL spreadsheet with DataMiner3D,  by trial an error moving different activities on axis and objects.  A pretty separation into three clusters were found by using the activities "angiotensin II antagonist, cognition disorder treatment, anticoagulant, and antidiabetic. For instance using the activity antiparkinsonian for shape did not bring anything in terms of clustering.

We wanted to see if we could cluster the same structure file by using different activities. This time we predicted all (about 500) activities with PASS and selected for the display about 20 activities, which had a Pa > 4 in some arbitrary selected compounds.

The figure below shows again a separation in two clusters. It also interesting to note that the compounds, which are predicted to be active as Angiotensin AT2 receptor antagonist are also active as Neuropeptide antagonist.