To assist drug design, we are developing computational tools to predict ligand binding sites. In the past, we developed a structure based program using normal mode accompanied exposure changes to predict ligand binding sites with 90% accuracy. As one would expect we are currently attempting to increase the accuracy to 100%. In addition, we are developing ligand optimization programs based on local motion in the binding site. In particular, we are optimizing drugs which bind to acetylcholine receptors, and acetylcholine esterases, and improve concentration in patients with Alzheimer's and dementia.