I've performed gene set enrichment analysis with the limma-package's camera -function. I'm using limma's barcodeplot -function to visualize my results (one index set, no gene weights). I would like to know how the enrichment score on the plot (y-axis) is calculated - I can't find any mention of it in the help files or limma's user guide. An example of a barcode plot can be found here (figure 9): https://f1000research.com/articles/5-1438/v2.
It's done using the function tricubeMovingAverage, with the span defined with span.worm. Imagine a sliding window that is applied to the set of all genes (ordered by log-fold change for your contrast of interest). In each window, you calculate the "density" of genes in the specified gene set, where the contribution of each gene in the window is weighted according to its distance from the center. The computed value for each position of the window is the enrichment score. I would say that it's mostly for visual purposes and should not be used for any quantitative decision-making.
See also How barcode-plot enrichment is calculated?
Thank you Gordon. I don't know how I missed this post...