“School of Biological”
Back to Papers HomeBack to Papers of School of Biological
| Paper IPM / Biological / 13474 |
|
||||||||||
| Abstract: | |||||||||||
|
Inference of gene regulatory networks (GRNs) from gene expression data is a major challenge in systems biology. There are several methods available to reconstruct a GRN from gene expression datasets. Bayesian network (BN) is one of the popular tools which have been successfully implemented in this field. Methods to reconstruct the structure based on BN approaches generally classified into three categories, constraint-based, score and search and hybrid methods. In this paper we introduced a hybrid algorithm (PCMIT) to reconstruct GRNs in which Conditional Mutual Information (CMI) tests is applied to obtain information about linear and nonlinear relations between genes. Then Hill Climbing (HC) algorithm is applied to give direction to the edges of skeleton of network. In HC algorithm, we use Mutual Information Test (MIT) score to measure the degree of fitness of a network to a data set.The PCMITalgorithm was evaluated by employing them on the DREAM3 Challengedata sets with n vertices and n samples (n = 10; 50; 100). Results indicated that applying the PCMIT algorithm, the precision of the inferred skeleton of the GRNs can considerably be improved.
Download TeX format |
|||||||||||
| back to top | |||||||||||
