A knowledge discovery method for the characterization of protein unfolding processes

Abstract

This work presents a method of knowledge discovery in data obtained from Molecular Dynamics Protein Unfolding Simulations. The data under study was obtained from simulations of the unfolding process of the protein Transthyretin (TTR), responsible for amyloid diseases such as Familial Amyloid Polyneuropathy (FAP). Protein unfolding and misfolding are at the source of many amyloidogenic diseases. Thus, the molecular characterization of protein unfolding processes through experimental and simulation methods may be essential in the development of effective treatments. Here, we analyzed the distance variation of each of the 127 amino acids Cα (alpha carbon) atoms of TTR to the centre of mass of the protein, along 10 different unfolding simulations - five simulations of WT-TTR and five simulations of L55P-TTR, a highly amyloidogenic TTR variant. Using data mining techniques, and considering all the information of the 10 runs, we identified several clusters of amino acids. For each cluster we selected the representative element and identified events which were used as features. With Association Rules we found patterns that characterize the type of TTR variant under study. These results may help discriminate between amyloidogenic and non-amyloidogenic behaviour among different TTR variants and contribute to the understanding of the molecular mechanisms of FAP. © 2009 Springer-Verlag Berlin Heidelberg.