Application of Network Methods for Describing Empirical Gene Expression Relationships in Quantitative Real Time PCR (qRT-PCR) Datasets

The advance of quantitative real time PCR (qRT-PCR) technique makes it possible to describe gene expression patterns and allows comparisons of the changes in gene expression levels across a wide range of experimental conditions with remarkable sensitivity and specificity. This research focuses on the investigation of methodologies amenable for quantitative data analysis of potentially large qRT-PCR datasets, particularly on the application of social network methods for establishing empirical gene expression relationships in distinctly designed experiments.  In this study, qRT-PCR gene expression data was generated from an experiment focusing on those genes in maize associated with resistance to Aspergillus flavus and alfatoxin accumulation.   RNA was isolated from developing kernels of three susceptible and three resistant maize inbred lines, each with three replications and two treatments.  Gene expression data were obtained by qRT-PCR technique and analyzed by statistical and network specific methods.  Of the 58 genes tested, 23 were differentially expressed among the tested maize inbred lines at a significance level of p<0.05.  To visualize gene expression relationships, a network was constructed with the 58 genes as vertices and the Euclidean distance of pc scores as edge values. The degree distribution of the vertices appeared to fit the Power Law distribution. Susceptibility related gene AI664980 was found adjacent to multiple RNA transport pathway genes, whereas resistance related genes TC231674 and BE050050 appeared as isolates. Application of network methods will be capable of revealing complex gene relational structures in qRT-PCR gene expression datasets from a wide range of experimental conditions.