From Proteins to Ligands Decoding Deep Learning Methods for Binding Affinity Prediction

Summary : (1st Read - January 2024)

1. Primary goal is to develop Deep Learning capable of predicting the binding affinities between proteins and Ligands .
   1. — Not clear why it is important. Related to drug discovery but how?
2. 1-D and 2-D encoding - Contact Map - Amino acids - Adjacency Matrix — Protein Language Model
3. AlphaFold2 , ESM-1B, Pcons4 - Generating protein contact maps
4. Impact of Ligand Encoding on binding affinity prediction
   1. — Result - Randomization and perturbations result is drastic drop in performance
      1. —- Hence Ligands is important