Considerations for feature selection using gene pairs and applications in large-scale dataset integration, novel oncogene discovery, and interpretable cancer screening

Table 1 Breast cancer datasets. Transcriptomic data from tumor and normal mammary tissue were downloaded from GEO. All datasets were combined to compare between cancer and control

Platform	Data Type	Control	Tumor	ER+	ER-	HER2+	HER2-	GSE ID
Affymetrix	Array	804	4193	1391	732	449	797	GSE4611, GSE7904, GSE10780, GSE10797, GSE11121, GSE15852, GSE18728^a, GSE18864^a, GSE20711^a, GSE21653^a, GSE22093^b, GSE23988^b, GSE26639^a, GSE42568^b, GSE45827, GSE48091, GSE48390^a, GSE53031^a, GSE54002, GSE65095^a, GSE78958, GSE93601^b, GSE124646, GSE129551^a, GSE131027
Agilent	Array	618	1295	701	259	178	743	GSE21974^a, GSE22820, GSE35186, GSE40206^a, GSE43973, GSE49175, GSE49481^a, GSE50939, GSE52604, GSE70905^a, GSE70947^a, GSE75678^a, GSE80999^a, GSE111601
Illumina	Array	0	1083	588	171	104	481	GSE20462, GSE36693, GSE37181^a, GSE45725^a, GSE46563^a, GSE60785^a, GSE103744^b, GSE111563
Illumina	RNA-seq	68	3779	3186	337	524	3074	GSE47462, GSE81538^a, GSE96058^a, GSE99680^a, GSE129508^a
	Total:	1490	10,350	5866	1499	1255	5095	52 datasets

^aDataset used for classification based on ER and HER2 status. ^b Dataset used for classification based on ER status

ISSN: 1755-8794