To search, Click below search items.


All Published Papers Search Service


Comparative Study of Score Functions for Edge Orientation Problem in Network Estimation


Hitoshi Afuso, Takeo Okazaki, Morikazu Nakamura


Vol. 12  No. 9  pp. 32-38


In this paper, we compared the score functions for edge orientation problem in estimation of genetic network from DNA microarray data. We focused four score functions, Bayesian Information Criterion(BIC), Bayesian-Dirichlet Metric(BDM), K2 Metric and PageRank Orientation Metric(PROM). To compare and evaluate the performance of each score function in various situations, we generated genetic networks changes of gene expression level, artificially. To generate such artificial networks, we used subsampling technique from large-scale real transcriptional regulatory network. To determine the set of genes that observed their expression level changed, we considered not only the structures of generated networks, but also the certain type of gene regulation that it makes difficult to detect the change of expression levels in knock-out or knock-down DNA microarray experiments.


TRNs estimation, Edge Orientation Problem, Bayesian score functions, PageRank Orientation Metric