PDB is an important data archive for protein structures and their structural information. Majority of structures available are obtained from X-ray crystallography. PDB file contains the coordinate information of the atoms including their spatial organisation. The water molecules are represented as HETATM (heteroatoms to the primary structure), which denotes the atoms that are not primary to the structure of the protein, i.e. non-main chain atoms.
During molecular dynamics studies of biological proteins, these PDB files are used. The protein structures need to be normalised before subjecting them to actual simulation process, so non-required information is discarded from PDB files. Majority of water molecules are removed but at certain places, water is retained. Why are selective water molecules retained?