| Summary: | 博士 === 國立陽明大學 === 微生物及免疫學研究所 === 97 === To analyze significant expression changes of predefined gene sets, rather than individual genes, has become a main approach to describe functional characteristics associated with transcriptome profiles. Currently, most gene set approaches are based on the measurement of the correlation between gene expression levels and the samples’ phenotypes. However, in highly heterogeneous samples, such as tumors, the detection power of conventional methods is greatly reduced due to the lack of consistency in gene expression levels. To overcome such limitation, we contrive here a new algorithm to detect changes of individual genes within individual test samples, and define such changes as regulatory events. We found that such new algorithm can tolerate more transcriptomic heterogeneity than the conventional methods. Based on regulatory events, we develop here a novel method called the regulatory event-based Gene Set Analysis (eGSA). From our test data sets, eGSA can detect transcriptome functional changes in a more precise and robust manner than convention methods, especially when applied to heterogeneous data sets.
The heterogeneous nature of hepatocellular carcinoma (HCC) has been well recognized. We thereby apply eGSA to describe functional patterns of each pathological stage during HCV-induced hepatocarcinogenesis using a public HCC microarray data set. From such analysis, we are able to identify molecular mechanism underlying HCC carcinogenesis. We notice that Toll-like receptor, JAK-STAT, MAPK and T cell receptor signaling pathway are involved in the immune escape mechanism during early stage HCC development. In addition, cell proliferations during early HCC can be induced by deregulation of G2/M check point, and cyclin A2 and its accompany regulators are significantly up-regulated.
In conclusion, this study develops a novel function-orientated scheme for transcriptome analysis. When the primary function involved is identified by eGSA, we can then zoom in to all potential regulatory routes. Individual genes within a specific regulatory route can then be further studied. This function-to-gene strategy can be better applied to heterogeneous transcriptome data and explore more potential regulatory mechanisms than the conventional gene-to-function methodology.
|
|---|
