The Research On Time Course Experiment Of Gene Regulation Network Analysis With Porcine Skeletal Muscle Applying MDL Principle

The study was based on feeding n-3 polyunsaturated fatty acids (n-3 PUFA) diet at different periods of porcine skeletal muscle expression profile data, used appropriate algorithms to process the data in case of the time course factor and nutrient factor. To further study the n-3 PUFA regulation of gene expression in skeletal muscle formation mechanism, construct the gene regulatory networks. At the same time based on the research background of the influence of time course and nutrients do not significantly change, established a specific pre-analytical framework.The collection of experiment samples was taken from three pigs longissimus dorsi muscle mass for each group as three repeated microarray experiments randomly, each repeat was a chip, added up to 12 chips. The major results as follows:(1) Using EDGE, ANOVA and SlopeMiner algorithm to differentiate four simulated gene chip datas. There were significant differences between the two methods. Such as using "Q<0.01" simulation data model, the FDR value and FNR value screened by EDGE were 0.0066 and 0.1048, the FDR value and FNR value screened by ANOVA were 0.0150 and 0.1012, the FDR value and FNR value screened by SlopeMiner were 0.9672 and 0.1795. FDR screened by EDGE which is lower than ANOVA and SlopeMiner algorithms, and FNR values were lower too. It shows that EDGE can work on microarray data under the influence of time course, the result is real reliable. Therefore, the experiment used the differentially expressed genes made by EDGE to analysis gene network.(2) According to the differential genes that have been screened,52 genes about fatty production and metabolism were selected by gene function annotation and 75 genes about the formation and differentiation of muscle-related were used to do the gene network analysis. Based on the regulation of this background of time course and nutrient, this experiment was designed to construct the dynamic bayesian networks and MDL algorithm network successfully. Be compared of the different network results, then we may correspond the regulatory networks in inquiry of the molecular genetic pathways(3) Through a more compact presentation of the network, query the biological genetic information on KEGG database, trying to find the genes relationship that regulat the pig skeletal muscle by n-3 PUFA.The conclusion of the experiment as follows: (1) EDGE has better performance than ANOVA and SlopeMiner algorithm on the gene chip data under the influence of time course for this experimental, it reflects biological process more realistic.(2) MDL algorithm is better than the dynamic Bayesian network model. It reflects the relationship between genes more accurately and more vivid manifestation the key nodes in the regulatory network, more suitable for regulatiing the gene networks under the time course factor.(3) LEF1, FBXW11 and PPP2R1A through Wnt signaling pathway and its transcription factor affecting the formation and metabolism of pig fat possibly. FOXC1, EGF and MEF through MAPK signaling pathway and its transcription factor affecting the formation and differentiation of pig muscle possibly.