Comparison of hybridization-based and sequencing-based gene expression technologies on biological replicates

BMC Genomics

Table 4 Comparison of differentially expressed gene identification between microarrays and MPSS

	Affymetrix		Amersham		Mergen		ABI
	MPSS (17-bp)	MPSS (20-bp)	MPSS (17-bp)	MPSS (20-bp)	MPSS (17-bp)	MPSS (20-bp)	MPSS (17-bp)	MPSS (20-bp)
Number of genes included in the comparison	1912	1827	2248	2127	1518	1452	1628	1535
Number of genes considered as differentially expressed in both microarray and MPSS	29	30	49	46	17	16	81	79
Number of genes considered as having NO differential expression in both microarray and MPSS	1119	1084	1350	1277	937	905	900	858
Number of genes that are considered as differentially expressed in MPSS but not in the microarray	714	664	761	717	550	517	514	473
Number of genes that are considered as differentially expressed the microarray but not in MPSS	50	49	88	87	14	14	133	125

One-way ANOVA was conducted for all microarray platforms that had been analyzed with technical replicates (Affymetrix, Amersham, Mergen, ABI) to identify differentially expressed genes between samples, MRP1 and MRP2. For MPSS libraries, a Z-test was used to identify differentially expressed genes, since technical replicates were not available. A threshold p-value of 0.001 was used for both statistical tests to define differential expression. Based on UniGene cluster mapping, identification of differentially expressed genes from MPSS was compared to identification of differentially expressed genes from the microarrays. For each microarray platform, a column corresponding to comparison with the 17-bp tag library and a column corresponding to the 20-bp tag library are shown. The detailed comparisons of agreement/disagreement were restricted to the common UniGene clusters where there was a one-to-one correspondence with MPSS tags.

ISSN: 1471-2164