QUANTILE TRANSFORM IN STRUCTURAL BIOINFORMATICS PROBLEMS

In this paper we study features of the multivariate empirical quantum function implementation for which sample is distributed at the mesh points of the regular grid. We present an algorithm for continuous and discrete quantile transform based on recursive definition of the multivariate quantile function. We perform numerical study of the presented algorithm and demonstrate it computational complexity according to representation of the sample. We present the results of using evolutionary optimization algorithm with quantile transform for solving the problems in structural bioinformatics: protein structure prediction from amino acid sequence and protein-peptide docking with known binding site and linear peptide structure.