Gene-based and probe-based network modeling in some bovine tissues using DNA microarray data

The aim of this study was to extract genes and probes hub based on Bayesian networks on probe and gene transcriptomic data over different tissues in Bovine species. Using raw probe and gene transcriptomic data in each tissue, genes and probes with the highest expression variances were extracted and fitted to Bayesian networks. The hub genes and hub probes were identified using the ratio of in-and-out degree.The size of the Markov Blanket was different in different networks. This might be indicative of existing of substructures topology of aforementioned network on different tissues. Using gene based network, in muscle (CBR1 and LOC788826); in mammary of (NID2, COL5A2, LOC616942 and FXYD3); in liver (LOC100132279; MGC127133; MBOAT2; CLDN2; ANKRD1; IGFBP1; DGAT2; CKMT1, ISG15, CKMT1), and in the ovary (LOC286871 and INHBA) were extracted as hub genes. It was shown that the hubs were different in different tissues. These results can be used for more accurate bioassays of each tissue. Using probe based network, two genes BOLA-DQB and JSP.1 would have function in liver and uterus tissues. The results of this study can reduce the gap between phonemics-and Genomics distance on investigated tissues in bovine species.