Significance calculations for pathway analysis

[kwcurrin]

For the pathway analysis, we plan to:

1. Identify the kegg pathways in which a metabolite is involved.
2. For each pathway, run TwoSampleMR for each component metabolite.
3. For each pathway, determine the percent of metabolites that were significantly associated with the outcome variable.
4. Generate a bargraph showing the top few pathways ranked by percent of metabolites significantly associated with outcome.

Problem: This ranking is qualitative and does not identify if the top ranked pathway has a higher percent of significantly associated metabolites than expected by chance.

Approaches:

1. Calculate a null distribution using 100 to 1,000 pathways. This is not feasible to calculate dynamically. Instead, we could precompute the null distribution and use it for every tested pathway.
   2....