Illumina WG-6 BeadChip strips should be normalized separately

BMC Bioinformatics

Table 2 Functional analysis for DE genes at Tpl2^-/- 180 mins from different normalization strategies

		Array-level		Strip-level
Source	Term	Count	FDR	Count	FDR
INTERPRO	Four-helical cytokine, core	4	0.94	9	7.5E-7
SP_PIR_KEYWORDS	cytokine	5	0.21	9	1.6E-4
KEGG_PATHWAY	Cytokine-cytokine receptor interaction	5	0.84	10	0.001
KEGG_PATHWAY	Toll-like receptor signaling pathway	-	-	7	0.002
KEGG_PATHWAY	Jak-STAT signaling pathway	3	1.0	8	0.002
GOTERM_MF_ALL	cytokine activity	5	1.0	9	0.007
GOTERM_CC_ALL	extracellular region part	15	0.072	22	0.008
GOTERM_CC_ALL	extracellular space	14	1.0	21	0.008
GOTERM_CC_ALL	extracellular region	16	0.072	23	0.013
KEGG_PATHWAY	Regulation of autophagy	-	-	4	0.038
INTERPRO	Interferon alpha	-	-	4	0.039
GOTERM_MF_ALL	receptor binding	7	1.0	12	0.043
SMART	IFabd	-	-	4	0.043
SP_PIR_KEYWORDS	glycoprotein	17	0.16	24	0.048
SP_PIR_KEYWORDS	Secreted	10	0.2	14	0.048

The first column is the data source from which terms in the second column come. The third column gives the number of DE genes found by array-level normalization which are associated with the term in the same row in the second column. FDR (False Discovery Rate) from the enrichment test for these genes in all the DE genes found by array-level normalization is given in the fourth column (see METHODS for details). The fifth and sixth column are similar with the third and fourth column respectively except that DE genes examined are from strip-level normalization rather than from array-level normalization.

ISSN: 1471-2105