Table 3.

Random forest breast cancer grade prediction accuracy with top important gene information. Pathways with cross-validation accuracy and prediction accuracy greater than 0.65 (all from PFOCR), plus the top ranked results from Reactome and KEGG. The result ranking was determined by min (cross-validation accuracy, prediction accuracy). The “Top Gene” corresponds to the gene with the highest feature importance score in each pathway. The “Ref.” column provides paper and pathway citations for each result

Rank	Pathway	Ref	Top Gene	Importance Score	CV Accuracy	Prediction Accuracy	Specificity	Sensitivity	F1
1	PMC2937358__F1	[45, 46]	CXCL6	0.61	0.72	0.72	0.61	0.83	0.69
2	PMC2653381__F1	[47, 48]	GSTM4	0.25	0.69	0.7	0.58	0.83	0.67
3	PMC5772637__F6	[49, 50]	CASP9	0.91	0.69	0.68	0.84	0.52	0.73
4	PMC5772637__F7	[49, 51]	CASP9	1.54	0.68	0.72	0.77	0.66	0.74
5	PMC8040471__F7	[52, 53]	CASP9	0.45	0.68	0.67	0.81	0.52	0.71
6	PMC6759650__F5	[54, 55]	CYCS	0.37	0.67	0.78	0.71	0.86	0.77
7	PMC7811378__F9	[56, 57]	CASP9	0.85	0.67	0.78	0.84	0.72	0.8
8	PMC5715135__F7	[58, 59]	CASP9	0.15	0.67	0.75	0.84	0.66	0.78
9	PMC2673236__F1	[60, 61]	ACSM3	1.32	0.67	0.72	0.74	0.69	0.73
10	PMC387764__F8	[62, 63]	CASP9	1.78	0.67	0.68	0.74	0.62	0.71
11	PMC6947643__F2	[64, 65]	HEY1	1.23	0.67	0.67	0.84	0.48	0.72
12	PMC7409684__F1	[66, 67]	WNT11	0.18	0.67	0.67	0.68	0.66	0.68
13	PMC8023395__F2	[68, 69]	EGFR	0.17	0.66	0.73	0.84	0.62	0.76
14	PMC4336604__F9	[70, 71]	MAPK9	0.33	0.66	0.72	0.74	0.69	0.73
15	PMC6499473__F1	[72, 73]	PARP1	0.23	0.66	0.72	0.61	0.83	0.69
16	PMC3219187__F5	[74, 75]	ZBTB7A	2.29	0.66	0.7	0.81	0.59	0.74
17	PMC6024909__F4	[76, 77]	CDK1	0.26	0.66	0.7	0.52	0.9	0.64
18	PMC2694962__F1	[78, 79]	ACSM3	1.52	0.66	0.68	0.48	0.9	0.61
19	PMC5256616__F6	[80, 81]	CASP9	0.36	0.66	0.68	0.48	0.9	0.61
20	PMC4407294__F2	[82, 83]	NOTCH2	0.7	0.66	0.67	0.52	0.83	0.62
21	PMC6305585__F2	[84, 85]	CPEB1	0.77	0.66	0.67	0.71	0.62	0.69
63	R-HSA-8864260	[86]	HSPD1	0.25	0.64	0.68	0.81	0.55	0.72
115	KEGG_hsa01522	[87]	BAX	0.11	0.62	0.85	0.9	0.79	0.86