Next-generation text-mining mediated generation of chemical response-specific gene sets for interpretation of gene expression data

Table 2 Fibrate drug prediction ranks

	ToxProfiler		Weighted global test		GeneCodis single annotation
	CTD	TM	CTD	TM	CTD	TM
clofibrate	7 (6.1e-10)	2 (<1e.-10)	22 (1.0e-03)	71 (1.0e-06)	5 (1.1e-04)	2 (7.6e-29)
gemfibrozil	53 (2.0e-01)	8 (<1e.-10)	35 (2.0e-03)	106 (5.4e-06)	-	6 (3.5e-19)
bezafibrate	54 (2.5e-01)	7 (<1e.-10)	42 (2.1e-03)	105 (5.3e-06)	-	5 (9.3e-22)
fenofibrate	3 (<1e.-10)	10 (<1e.-10)	27 (1.6e-03)	115 (7.1e-06)	3 (2.2e-07)	9 (1.5e-18)
ciprofibrate	33 (4.2e-02)	5 (<1e.-10)	21 (8.7e-04)	97 (2.6e-06)	-	3 (3.1e-25)
WY14643	2 (<1e.-10)	4 (<1e.-10)	12 (5.5e-05)	118 (7.2e-06)	1 (3.2e-79)	4 (6.5e-22)

Fibrate drug prediction ranks and false discovery rate-corrected p-values (within parentheses) for the PPARA gene expression data set for the different gene set analysis tools. Significant results (false discovery rate-corrected p<0.05) are in bold. GeneCodis only reports results with a false discovery rate-corrected p-value smaller than 0.05. Missing values denote such non-significant results. ToxProfiler does not report false discovery rate-corrected p-value less than 1.e-10. These are denoted as <1.e-10 in the table.

ISSN: 1755-8794