Detecting constituent sequences by means of HP pattern-based grammars to synthesize proteins: Inferring sequence-structure-function relationship

The detection of protein characters that could reveal how protein chains are constituted, is an important step to understand the main functions of specific classes of protetins. We made use of the concept of "HP Pattern-Based" grammars to study the connection between protein chains and protein functions. In order to consider the structure of the proteins the HP models were used. Amino acid sequences were treated as a formal language, and it was built a set of HP Pattern-Based grammars to describe this language by means the Teiresias pattern discovery tool. First, this methodology was tested on the class of Antimicrobial peptides (AmPs). The deduced derivation rules of HP Pattern-Based Grammars were validated by the regular grammar designed by [11] which was used to create new, unnatural, AmPs sequences. Then, our approach was applied to characterize a function of the Pleckstrin Homology domain(PH Domain) which represents an important three dimensional domain which bind to phosphoinositides. Nowadays, interactions among PH domain amino acids and inositol phosphate are not well characterized. For the first time, by means of an HP Pattern-Based grammar we highlight that this binding function can be described in terms of hydrophocity patterns. Our approach points out some fundamental aspects regarding the relationship between sequence, structure and function of proteins.