Differential search algorithm in deep neural network for the predictive analysis of xylitol production in Escherichia coli

Abstract

Xylitol is one of the bio-based chemical products that received well recognition and highly demanded from both food and pharmaceutical industries which led to various experiment to be carried out on various organism to produce xylitol. Recently, E. coli has become the spotlight to be one of the organisms that can be metabolically engineered to produce xylitol by using gene knockout strategy. However, gene knockout strategy required laborious, expensive and time-consuming when conducted in vivo. Motivated by this, in silico experiment has been done to simulate and manipulate the model of E. coli to construct a new E. coli model that will focus on producing xylitol by using Flux Balance Analysis (FBA). In this paper, a new hybrid method called DNNDSA is proposed which consists of both Deep Neural Network (DNN) method and Differential Search Algorithm (DSA) to do the predictive analysis on a newly constructed model of E. coli to predict which gene knockout condition is the best to be applied in metabolic pathway of E. coli to improve the xylitol production.