A Machine Learning Approach for Prediction of Gibberellic Acid Metabolic Enzymes in Monocotyledonous Plants

Abstract

Gibberellins (GA) are one of the most important phytohormones that control different aspects of plant growth and influence various developments such as seed germination, stem elongation and floral induction. More than 130 GAs have been identified; however, only a small number of them are biologically active. In this study, five enzymes in GA metabolic pathway in monocots have been thoroughly researched namely, ent-copalyl-diphosphate synthase (CPS), ent-kaurene synthase (KS), ent-kaurene oxidase (KO), GA 20-oxidase (GA20ox), and GA 2-oxidase (GA2ox). We have designed and implemented a high performance prediction tool for these enzymes using machine learning algorithms. ‘GAPred’ is a web-based system to provide a comprehensive collection of enzymes in GA metabolic pathway and a systematic framework for the analysis of these enzymes for monocots. WEKA-based classifiers (Naïve-Bayes) and Support Vector Machine (SVM) based-modules were developed using dipeptide composition and high accuracies were obtained. In addition, BLAST and Hidden Markov Model (HMMER-based model) were also developed for searching sequence databases for homolog’s of enzymes of GA metabolic pathway, and for making protein sequence alignments.