Application of new probabilistic graphical models in the genetic regulatory networks studies

摘要

This paper introduces two new probabilistic graphical models forreconstruction of genetic regulatory networks using DNA microarray data. One isan Independence Graph (IG) model with either a forward or a backward searchalgorithm and the other one is a Gaussian Network (GN) model with a novelgreedy search method. The performances of both models were evaluated on fourMAPK pathways in yeast and three simulated data sets. Generally, an IG modelprovides a sparse graph but a GN model produces a dense graph where moreinformation about gene-gene interactions is preserved. Additionally, we foundtwo key limitations in the prediction of genetic regulatory networks using DNAmicroarray data, the first is the sufficiency of sample size and the second isthe complexity of network structures may not be captured without additionaldata at the protein level. Those limitations are present in all predictionmethods which used only DNA microarray data.