Inferring Gene Regulatory Network from Bayesian Network Model Based on Re-Sampling

Abstract

Nowadays, gene chip technology has rapidly produced a wealth of information about gene expression activities. But the time-series expression data present a phenomenon that the number of genes is in thousands and the number of experimental data is only a few dozen. For such cases, it is difficult to learn network structure from such data. And the result is not ideal. So it needs to take measures to expand the capacity of the sample. In this paper, the Block bootstrap re-sampling method is utilized to enlarge the small expression data. At the same time, we apply “K2+T” algorithm to Yeast cell cycle gene expression data. Seeing from the experimental results and comparing with the semi-fixed structure EM learning algorithm, our proposed method is successful in constructing gene networks that capture much more known relationships as well as several unknown relationships which are likely to be novel.