Pathway Bar plot should include the possibility of having multiple bars for each gene.

[alexlenail]

In the case in which someone has more than binary experimental conditions, a "case, normalized to control" signal for each gene is insufficient. Imagine an experiment in which someone wants to understand combinatorial experimental variables, e.g. AB vs Ab vs aB vs ab. Perhaps 3 bars per gene with AB, Ab, and aB all normalized against ab would be interesting. Perhaps something else. Generally speaking, it should be possible to plot multiple bars per gene, given a suitable data structure.