Application of Recurrence Quantification Analysis (RQA) in biosequence pattern recognition

Abstract

Recurrence Quantitative Analysis is a relatively new pattern recognition tool well suited for short, non-linear and non stationary systems. It is designed to detect recurrence patterns that are expressed as a set of Recurrence Quantification variables. In our work we made use of this tool on allosteric protein system to identify residues involved in the transmission of the structural rearrangements as an upshot of allostery. Allostery is the phenomenon of changes in the structure and activity of proteins that appear as a consequence of ligand binding at sites other than the active site. Here, we scrutinized the sequence landscape of 'ras' protein by partitioning its residues into windows of equal size. An 11 element characteristic vector, comprising of 10 features extracted from the Recurrence Quantification Analysis along with a feature relating to allosteric involvement, was defined for each windowed sequence set. By applying multivariate statistical analysis tools including Principal Component Analysis and Multiple Regression Analysis upon the characteristic feature vectors extracted from all the windowed sequence set, we could develop a significant linear model to identify the residues that are critical to allostery of 'ras' protein. © 2011 Springer-Verlag.