Provide a better means for drilling into underlying data

[ppavlidis]

Other than 'gene contributions', users have little way to see exactly why a particular condition comparison comes up, at the moment. It will help to show more information about the underlying experiments (see #18) but can we do more?

A couple of ideas:

• For each experiment show the specific conditions / factor values that were relevant, rather than just the one that was used in the analysis (which is often an inferred higher-level term). This would help a lot because the terms are often cryptic (e.g. "breast disease vs. mammary cancer")
• Showing the actual fold-changes/p-values for the genes. We could show this as a distribution per gene or in a heatmap.
• Allowing drilling down to the expression data, kind of like we do for coexpression. This is unwieldy when there are a lot of experiments involved, but it works easily for 10.

[k8iechen]

• understanding how condition comparisons listed correlate to ontology step settings: why some show up & how far they are