Discovery and validation of gene classifiers for endocrine-disrupting chemicals in zebrafish (danio rerio)

BMC Genomics

Table 3 Summary statistics of qualified gene classifiers from network-specific searches

Chemical/tissue	All tissues (495 search / condition)^a	Brain (496 search / condition)	Ovary (495 search / condition)	Testis (497 search / condition)	Subtotal by chemical
Prochloraz	&	177	68	76	321
Flutamide	&	NA	124	0	124
Fipronil	&	71	17	20	108
17α-ethynyl estradiol	0	41	---	50	91
Vinclozolin	0	NA	28	1	29
Ketoconazole	&	18	7	---	25
Muscimol	&	22	---	NA	22
Fadrozole	&	4	0	NA	4
17β-trenbolone	&	---	0	NA
trilostane	0	NA	---	0
Subtotal by tissue		333	243	147

A total of 517 individual networks were analyzed by chemical-tissue conditions using genetic algorithm-K nearest neighbors. Within a condition, each TF network contributed one classifier composed of multiple gene features. “NA”, a condition not included in the study; “---“, insufficient sample size for classifier search; “0”, searched but no qualified classifier; “&”, no search conducted.
^aBetween 20 to 22 of the 517 networks did not contain sufficient number of genes to be searched due to variations in their availability in different datasets.

ISSN: 1471-2164