Recently, various machine learning methods have emerged for analyzing and interpreting the ever-expanding genetic data. In addition, new analytical tools for machine learning using statistical models, are being developed. Lim et al. [1] proposed the integrative deep learning to find the differentially expressed (DE) biomarkers using deep neural network with a single hidden layer. This method consists of the input layer, a hidden layer, the consolidation layer, and the output layer. They found that integrative deep learning method is stable and robust for analysis of the variation in the simulation datasets. In this study, we expanded the integrative deep learning method by including an additional hidden layer. The present expanded method consists of the input layer, two hidden layers, the consolidation layer and the output layer. The purpose of this study is to investigate the effect of the additional hidden layers on the performance of the previous method (integrative deep learning). We conducted a simulation study and compared the results with those from deep neural network with one hidden layer.