HIGHLIGHTS
SUMMARY
The decoys were extracted from ChEMBL and were chosen based on common molecular descriptors widely used for describing drug-likeness, such as the molecular weight (MW), the logarithm of octanol/water partition_coefficient (logP), the number of hydrogen_bond donors (HBD) and the number of hydrogen_bond acceptors (HBA). The repurposing model was a multi-class classification algorithm based on a multilayer perceptron neural_network (MLP NN) and was trained to discriminate not only between active and inactive molecules but also between TRPV1 agonists and antagonists. The machine_learning algorithm that the authors selected for task was the multilayer . . .
If you want to have access to all the content you need to log in!
Thanks :)
If you don't have an account, you can create one here.