Analyzing protein motif combinations using a data mining technique

Abstract

It is apparent that some regions in a protein have been more conserved than others during evolution. These regions are called `motif``s that have representative features of functions and structures. From a view of function, to understand activities of a motif combination is also, sometimes even more, important as those of a single motif. Motifs act as a `functional module`` in combination, and give more complicated and specified function to a protein. Therefore, searching fully informative NRFC (Non-Redundant Frequently Co-occurring) motif combinations and analyzing the biological meanings of them are needed for understanding functions of proteins. I proposed an expert system MOCHA (MOtif Co-occurrence cHaracteristics Analysis) for biological interpretation about NRFC combinations. The 76.5% of the searched NRFC motif combinations were previously unknown. NRFC combinations are full of functional information in proteins or protein families. With MOCHA system, three kinds of information can be obtained from NRFC combinations: (1) as a ``functional module````, (2) as a information provider for un-annotated motif(s) and (3) as a reflection of the complexity of an organism. MOCHA system is reachable through Would Wide Web at http://mocha.kaist.ac.kr.