A gene sets approach for identifying prognostic gene signatures for outcome prediction

BMC Genomics

Table 7 Top 20 gene sets with high prediction accuracy (analysis with six datasets)

Gene set	category	GTG	GTP	PTG	PTP	accurary	sensitivity	specificity
11823860_ST2	br	3175	1325	559	746	0.6755	0.7056	0.5716
transferase activity	mf	3264	1236	658	647	0.6737	0.7253	0.4958
ligase activity	mf	3204	1296	633	672	0.6677	0.712	0.5149
11823860_ST3	br	3200	1300	632	673	0.6672	0.7111	0.5157
transcription factor activity	mf	3268	1232	701	604	0.667	0.7262	0.4628
16141321_SDC2	br	3169	1331	607	698	0.6661	0.7042	0.5349
oxidoreductase activity	mf	3209	1291	648	657	0.666	0.7131	0.5034
14737219_CSR	br	3165	1335	606	699	0.6656	0.7033	0.5356
12917485_ST9	br	3162	1338	611	694	0.6643	0.7027	0.5318
catalytic activity	mf	3209	1291	661	644	0.6637	0.7131	0.4935
RNA polymerase II transcription factor activity	mf	3235	1265	689	616	0.6634	0.7189	0.472
transport	bp	3186	1314	645	660	0.6625	0.708	0.5057
transcription	bp	3241	1259	701	604	0.6624	0.7202	0.4628
transporter activity	mf	3171	1329	631	674	0.6624	0.7047	0.5165
14737219_USR	br	3094	1406	555	750	0.6622	0.6876	0.5747
12917485_ST7	br	3140	1360	602	703	0.662	0.6978	0.5387
ATP binding	mf	3185	1315	647	658	0.662	0.7078	0.5042
kinase activity	mf	3205	1295	669	636	0.6617	0.7122	0.4874
metabolism	bp	3199	1301	666	639	0.6612	0.7109	0.4897
regulation of progression through cell cycle	bp	3108	1392	575	730	0.6612	0.6907	0.5594

*category: br – breast and other cancer gene set; mf – molecular functions; bp – biological processes
**GTG – Good prognosis group predicted as Good; GTP – Good prognosis group predicted as Poor; PTG – Poor prognosis group identified as Good; PTP – Poor prognosis group identified as Poor
^ accuracy = (GTG+ PTP)/(GTP+GTP+PTG+PTP); sensitivity = GTG/(GTG+GTP); specificity = PTP/(PTG+PTP)

ISSN: 1471-2164