Reverse causal reasoning: applying qualitative causal knowledge to the interpretation of high-throughput data

BMC Bioinformatics

Table 6 HYPs meeting significance criteria in matched real and randomized data sets

Richness and concordance p-value threshold	TNF			PI3K inhibitor			High fat diet
Richness and concordance p-value threshold	Random	Real	Ratio	Random	Real	Ratio	Random	Real	Ratio
0.2	13	94	14%	20	81	25%	7	33	21%
0.15	9	84	11%	12	66	18%	4	25	16%
0.1	3	58	5%	4	45	9%	1	13	8%
0.05	1	41	2%	1	29	3%	0	5	0%
0.01	0	24	0%	0	14	0%	0	0	N/A
0.005	0	21	0%	0	6	0%	0	0	N/A

For each example data set (TNF, PI3K inhibitor, and high fat diet), 1000 matched random data sets were generated. The number of HYPs meeting significance criteria at each p-value threshold is indicated, for random data the median number of significant mechanisms is shown.
BOLD indicates the 0.1 p-value threshold applied to analysis of the three example data sets.

ISSN: 1471-2105