Hydrophobicity and charge signals on protein surface

An image-oriented modality of describing protein surfaces in terms of hydrophobicity and charge distributions seen as signals has been used to classify surface atom neighborhoods. Global similitude and interaction measures of pairs of surface atom neighborhoods are defined, as well as vector descriptions of hydrophobicity and charge density distributions around each surface atom. Surface atom neighborhoods have been classified in terms of both resemblance and vector description. The implementation of the clustering algorithm is presented.