Reconstruction of gene regulatory networks using state space model

Abstract

State Space Model (SSM) is an approach to inferring gene regulatory networks. It requires less computational time than Dynamic Bayesian Network (DBN). There are two types of variables in the linear SSM, observed variables and hidden variables. SSM uses an iterative method, namely Expectation-Maximization, to infer regulatory relationships from microarray datasets. The hidden variables cannot be directly observed from experiments. How to determine the number of hidden variables has a significant impact on the accuracy of network inference. In this study, we investigated the effect of hidden variables in the state space model and their impact on inference accuracy. Ten different gene regulatory networks (GRNs) and synthetic gene expression datasets were generated using GeneNetWeaver. Both DBN and linear SSM were used to infer GRNs from the synthetic datasets and the inferred networks were compared with the true networks. The results show that inference accuracy varied with the change of the number of hidden variables. For some true networks, the inference accuracy of DBN is higher but in other cases SSM performs better. In the tested cases, the overall performance of SSM and DBN are compatible. However, SSM was much faster than DBN and can infer large networks that DBN cannot handle because of its significant computational cost. This study provides useful information in handling the hidden variables and improving the inference accuracy.