- Employing the use of biochemical properties e.g. b-factor (describes the movement of atoms—low movement corresponds to likely binding sites deep within the protein), hydrophobicity (small molecules prefer hydrophobic protein pockets) and conservation information.
- Running X-ray crystallography in the lab in order to experimentally determine the structure of a protein and then run 3DPocket to predict its binding sites.
- Use nuclear magnetic resonance (NMR) spectroscopy in order to experimentally determine binding sites of proteins and compare these results to the predicted binding sites by 3DPocket.
- Test algorithm on a wider variety of proteins (both DNA-binding proteins and enzymes) from the Protein Data Bank (PDB) in order to ensure reliability.
- Optimize the number of calculations done by 3DPocket by employing heuristics to reduce the number of combinations that need to be tested when calculating minimum distances from the Connolly surface to the convex hull.