Analysis of structural measurements in correlation networks built from gene expression data across different tissue types in Mus musculus

Abstract

Big data analysis has been pervasively adopted as a method to analyze the tremendous amount of daily generated high throughput data in an efficient and accurate manner. Among the series of tools available in the field of big biomedical data, correlation networks are one of the most powerful tools for modelling gene expression, which is important in the study of disease and ageing. With the help of the correlation networks, insightful research has been done, such as distinguishing target genes for study within gene co-expression data. However, the utility of this model has not been thoroughly investigated as it pertains to applicability across and within tissue types. In this project, we address this gap in knowledge by investigating the range of outputs from analyzing correlation networks constructed from gene expression data. A total of 43 correlation networks were built using the gene expression data from 5 different tissues in Mus musculus. Then we compared a number of network measurements (degree distribution, assortativity coefficient, and clustering coefficient) across tissues to identify the span of possible ranges of each measure. We find that the average assortativity coefficient over all the networks is significantly different for networks between series, while the remainder of the parameters show no difference in average measure. Finally, we summarize the overall measurement ranges for number of nodes, number of edges, assortativity coefficient, clustering coefficient, and network density. This work is an investigation into the ability of the correlation network to represent gene expression data accurately, and the results that there are some common structural characteristics of data built across different tissues.