(This talk is a collaboration with Prof. Andrey Morgun, OSU College of Pharmacy.) False Discovery Rate (FDR - Benjamini-Hochberg) is the most popular multiple hypothesis correction procedure for network inferences. However, it is a conservative procedure (i.e., has high false negative rate) that results in losing many of the real connections in a network. This loss of important information happens because FDR makes an assumption of independence between connections in dense gene networks. We proposed a new method that does not make an assumption of independence between connections in a network. The method consists of the identification of the proportion of correlations/connections, in a network, that do not correspond to regulatory relations. It is based on a fundamental principle that connects correlation and causation, providing advantage over traditional statistics that do not use the fact that changes in any biological system are a result of cause-effect relationships. We applyProportion of Unexpected Correlations (PUC) method for reconstruction of gene regulatory networks that model transition of biological system from one state to another. Our simulation analysis suggests that using PUC allows for more accurate identification of a proportion of random connections (errors) in a network than using FDR or other similar methods.