Time Series Prediction of Gene Expression in the SOS DNA Repair Network of Escherichia coli Bacterium Using Neuro-Fuzzy Networks

摘要

Identification of gene regulatory networks and prediction of gene behavior through their expression data have become major problems in computational molecular biology in the last few years. In this study, a novel approach for time series gene expression prediction based on Mamdani neuro-fuzzy networks is proposed. In this work, a set of gene expression data is studied and neuro-fuzzy networks are designed to obtain models for gene expression prediction over time. In each neuro-fuzzy network, one of the genes is considered as the output and the other genes as the inputs. The time series expression values of the output are predicted by analyzing the time series expression values of the inputs. We study a popular gene expression data set of SOS DNA repair network of Escherichia coli bacterium. By comparing the results of proposed algorithm with previous works, it is obvious that our networks are predicting more precisely.