Difference between GSEA and GAGE analyses.

Question

[m.k.priebe]

We have been tasked with performing Gene Set Enrichment Analysis (GSEA) on GEO datasets as part of our masters project. These results should then be visualised as Heatmaps or even interaction networks.

The problems we are having is with finding appropriate R packages to do this. Ones that seem like they could be useful are the gage, gageData and pathview pacakages but these seem to perform Generally Applicable Gene-set Enrichment (GAGE).

What are the differences between GSEA and GAGE analyses? Would it be substitute one for the other? Are there any simple to use packages that will allow us to perform these analyses on GEO datasets and visualise the results?

Answer 1

[informatics bot]

The premise behind them is practically the same, they both use ranked lists to check where each gene falls on the distribution. If a proportionally large number of genes fall in the upper or lower part the the distribution, the gene set will be significant (you can adjust the directionality you are interested in). I prefer GAGE because the output can be easily added to path-view, also it's easy to make customized gene-sets from GEO. The GSEA R package is no-longer supported by Broad, which mean you have to use the GUI interface, which performs visualization for you.