Identifying cancer driver genes based on multi-view heterogeneous graph convolutional network and self-attention mechanism

Table 4 Co-citer analysis of the top 30 genes predicted by MRNGCN

	#MRNGCN	#MTGCN	#EMOGI	#GAT	Cancer	Driver	Biomarker	Drug target	in_NCG	%interact with drivers (%)
VCAN	1	23	18	4092	28	1	1	0	Lung	18.18
NR3C1	2	10	14	673	65	4	4	1	Blood	25.91
EPHA2	3	19	23	582	113	2	4	4	pan-cancer_adult	31.82
LRP1	4	49	58	2544	54	3	5	1	Blood	14.13
STAT1	5	22	29	604	346	5	21	4		27.54
FOS	6	44	21	579	195	2	9	2		23.94
RASA1	7	42	24	163	20	1	0	1	head_and_neck, lung, stomach, multiple	45.05
HNF4A	8	27	25	305	37	1	5	2	Hepatobiliary	31.78
SP1	9	28	10	566	291	3	10	2		33.61
CCL11	10	90	106	8929	18	1	23	2		8.70
KLF5	11	80	64	345	64	1	3	0	Multiple, bladder, colorectal, bladder	42.31
CCL5	12	66	122	4948	111	1	18	0		13.95
LRP6	13	82	137	2194	65	1	2	3		13.33
ATF3	14	35	34	410	83	1	8	2		25.00
FLT1	15	152	105	78	162	1	57	3	Brain	30.30
HSPG2	16	43	55	998	12	3	5	2	Esophagus	20.51
RUNX2	17	68	97	390	93	3	9	2		37.70
IRAK1	18	39	104	1647	9	1	3	0		18.75
SHC1	19	11	4	365	122	4	9	2		26.79
EPHA4	20	337	169	633	11	1	0	1	Skin	30.30
TP53BP1	21	164	139	1269	76	1	2	0	Multiple	20.00
FN1	22	5	11	2971	76	1	16	0		12.11
APOB	23	29	39	4239	22	1	12	1	Multiple, hepatobiliary, brain, stomach	15.25
MACF1	24	25	32	3447	4	0	0	0	Pancreas, multi, stomach	18.33
HCN1	25	1265	388	9816	3	1	0	1	Lung	0.00
INHBA	26	101	80	2832	39	1	4	0	Lung	25.00
LYN	27	8	8	344	55	1	6	2	Blood	21.17
PLCG2	28	102	51	56	10	1	2	0	Multiple, blood	41.67
HOXA10	29	216	450	145	30	1	2	0	Multiple	20.00
NRP1	30	142	186	1223	55	1	1	1		16.95

The second to fifth columns of the table show the ranking of these thirty genes in other methods, and the sixth to ninth columns show the co- citation times of these genes and the keyword cancer, driver, biomarker and drug target, in_ NCG indicates whether these genes are in the NCG database. The last column indicates the proportion of the edges connected to these genes and the driver genes in all edges

ISSN: 1471-2105