To answer the following question you will be analyzing the results you generated in step 1.
Q 3.1 What are the predicted bioactivity values (i.e. pChEMBL values) of the input molecules for thyroid hormone receptor alpha?
Q 3.2.1 How similar is Liothyronine to the most similar molecule (nearest neighbor) from the model’s training set? How about Rotenone? (Tip: have a look at the “Comprehensive output” tab)
Q 3.2.2 How is the similarity calculated, which properties is it based on? (Tip: have a look at the QPRF document)
Q 3.2.3 Where does the experimental value from the nearest neighbor originate from? Give an example of one of the experiments that was done.
Q 3.3.1 Do the input molecules fall within the applicability domain (AD) of the model?
Q 3.3.2 How is the AD calculated? (Tip: have a look at the QPRF document)
Modify the structure of Liothyronine (for example by removing some side chains or adding additional ones)(This can be done using PubChem Sketcher)
Q 4.1 What happens to the predicted bioactivity value? (Tip: also include Liothyronine in your run so that you can compare the results)
Q 4.2 Is this prediction within the applicability domain of the model?