Drug-Target Affinity Prediction — Full ML Project

What are we trying to do?

In drug discovery, a drug molecule needs to bind to a protein target in the body to have an effect. The strength of this binding is called affinity — measured here as pKd (a log-scale score).

Higher pKd = stronger binding = more effective drug

Our dataset has 68 drugs × 379 protein targets = 30,056 drug-target pairs

Each pair has a measured pKd value we want to predict

Our goal: Build a machine learning model that, given a drug's molecular structure and a protein's amino acid sequence, predicts how strongly they bind.