PathwayDenester

Determining the likelihood that a pathway is enriched on its own and not due to intersection with more significant pathways


Parent pathways contain very few unique selected genes outside the intersection; highlighting them would be misleading.
Child pathways would be misleading to discuss if they have a smaller density of selected genes than the MSP.

Figure 1

Figure 2

For overlapping pathways, it is hard to intuitively assess if it is a Hitchhiker or if it brings sufficient new information. Only those with enough additional information should be discussed.
Due to the structure of the ontology, many pathways in a pathway enrichment analysis (PEA) are Hitchhikers, which complicates the analysis and requires pathways to be manually selected or grouped.


PathwayDenester performs statistical tests to check if a pathway is distinctive enough to be highlighted alongside the MSP.
The genes selected from a pathway are randomized, then, starting from the most significant pathway, every intersecting pathway is tested. The procedure is then repeated for every pathway kept by previous iterations.
The output is an assessment of all pathways (keep/exclude) based on the statistical evaluation.

Figure 3

Figure 4

Input PEA data using familiar point-and-click boxes. The whole analysis runs server-side, so no need to have Python or any kind of package installed.
By clicking on the “Preform Analysis” button, receive the results in a table (that can be downloaded as a tsv file) and see the input pathways represented as a graph (which can be downloaded as a JSON file). Nodes are connected to the most significant pathway they intersect and are colored grey if they are to be excluded.