Animation of a seek-and-identify process. The process of identifying a binding site on the surface of a protein is a combinatorial search using the CSS algorithm i.e. only combinations of water molecules that could be displaced by a drug are considered (shown in green). A typical bromodomain protein has about 2000 water molecules covering its surface within the first two solvation shells (shown in blue), which result in 1060 possible combinations of clusters of water molecules that could be displaced. The CSS algoritm reduces the search to 1014 combinations. Red cluster shows the location for the cluster of water molecules that is the easiest to displace, that turns out to be within a binding site of the protein. Orange cluster shows the location of a cluster that is the easiest to displace but it is not in the binding site (an allosteric site).