Prioritization of Modelled Protein Structures using Pareto points and TOPSIS

Abstract

It is conceivable to produce large number of models for a given protein sequence. Ranking the predicted models accurately and choosing the best predicted model from the hopeful pool stay as challenging assignments. The physiochemical properties of the protein affect the nature of their structure, these properties are used to distinguish native or native-like structure. In this study, Pareto point technique and TOPSIS method are used. These methods are analyzed with four qualitative parameters i.e. Root-mean-square deviation for the entire target structure (RMS_CA), Template Modelling Score (TM-Score), Global Distance Test Total Score (GDT_TS) and Z-Score[D] to determine the TOPSIS Score for ranking these modelled protein structures. In similar work, protein structure prediction center can predict the ranking of modelled protein structures according to only one qualitative parameter i.e. GDT_TS Score but in the present work four qualitative parameters are used for ranking these modelled protein structures. This research work mainly focuses on the selection of the best modelled protein structures from the pool of decoy in the absence of its true native structure using Pareto point technique and TOPSIS method, Pareto point technique is used for finding the optimal modelled protein structures and the output of Pareto point technique is further used by TOPSIS method for find the best modelled protein structure from the decoy of optimal modelled protein structures. TOPSIS method uses four qualitative parameters to generate a single score value that is used to rank the modelled protein structures and select the best modelled protein structures from the decoy.