(93an) A Computational Framework for the Analysis of Time-Series Transcriptomic Data

DNA microarray analysis transformed the way in which problems in the life sciences are approached. The ability to monitor simultaneously the expression of thousands of genes is extremely valuable in the reconstruction of the gene regulation network. This implies, however, that sophisticated algorithms exist to correctly reveal the information content of the experimental data.

We programmed a computational framework comprising of algorithms developed in our group for the significance analysis of time-series transcriptomic data. To date, significance analysis methods had been developed for non-dynamic data, having limitations in extracting the time-dependent information from transcriptomic profiles. The suite of tools is written in C and consists of four parts. Specifically, it enables (a) the determination of the differentially expressed genes at each time point, (b) the quantification of the change in the expression of a gene, (c) the most highly correlated time points and (d), the GO categorization of the differentially expressed genes at each time point. The computational framework will be demonstrated in the context of a dataset acquired as presented by B. Dutta at the National AIChE Conference 2005 sessions 4as and 495c.