Fisher's combined p-value for detecting differentially expressed genes using Affymetrix expression arrays

BMC Genomics

Table 1 IQR of Ranks for True Positives.

Data	Method	Q1	Median	Q3
Golden Spike	ANOVA	345.00	1624.00	3986.50
	Combined P	337.50	1173.00	3527.50
	Cyber-T	356.50	1490.00	4640.00
	Median t	339.50	1320.00	3719.00
	Moderated t	366.50	1543.00	5403.00
	Original t	445.50	1652.00	5545.50
	Ideal	333.50	666.00	998.50
Gene Logic Tonsil	ANOVA	182.25	363.50	1455.50
	Combined P	182.25	363.50	1031.00
	Cyber-T	182.25	363.50	3749.00
	Median t	182.25	363.50	2638.50
	Moderated t	182.25	365.50	27707.00
	Original t	182.25	390.50	39284.00
	Ideal	182.25	363.50	544.75
Affymetrix Latin Square	ANOVA	956.25	1951.50	3023.75
	Combined P	956.25	1946.50	2984.50
	Cyber-T	956.25	1957.50	2995.75
	Median t	956.25	1949.50	2998.75
	Moderated t	965.25	1961.50	3002.75
	Original t	965.25	1961.50	3002.75
	Ideal	956.25	1911.50	2866.75

This table shows the interquartile range (IQR) of ranks by method for true positives for each of the datasets. For the Golden Spike data there are a total of 1331 true positives. For the Gene Logic Tonsil data there are 66 comparisons on 11 differentially spiked transcripts for a total of 726 true positives. For the Affymetrix Latin Square data there are 91 comparisons on 42 differentially spiked transcripts for a total of 3822 true positives.

ISSN: 1471-2164