Machine learning approaches for the genomic prediction of rheumatoid arthritis and systemic lupus erythematosus

Table 2 Top 20 ranking genes by feature importance for predicting RA and SLE in proposed models

Model	Gradient tree boosting			Extreme gradient boosting
Rank	SNP	Gene Symbol	Feature Importance	SNP	Gene Symbol	Feature Importance
1	rs6906021	HLA-DQA1	0.1205	rs34965214	HLA-DQA1	0.0998
2	rs9271858	HLA-DRB1	0.1030	rs3104376	HLA-DQA1	0.0562
3	rs9273505	HLA-DQB1	0.0895	rs1391371	HLA-DQA1	0.0557
4	Affx-28,477,341	HLA-DRB5	0.0848	rs9271662	HLA-DRB1	0.0494
5	rs4999342	HLA-DRB5	0.0741	rs9273322	HLA-DQA1	0.0387
6	rs41269945	HLA-DQA1	0.0452	rs1049072	HLA-DQB1	0.0380
7	rs3104376	HLA-DQA1	0.0372	rs9274605	HLA-DQB1	0.0359
8	rs9274605	HLA-DQB1	0.0348	rs9273370	HLA-DQA1	0.0345
9	rs2395533	HLA-DQA1	0.0323	rs17843604	HLA-DQA1	0.0337
10	rs9274655	HLA-DQB1	0.0317	rs9271850	HLA-DRB1	0.0269
11	rs9271662	HLA-DRB1	0.0239	rs9271588	HLA-DRB1	0.0241
12	rs3830059	HLA-DQB1	0.0224	rs9271425	HLA-DRB1	0.0220
13	rs200716952	HLA-DQB2	0.0207	rs4999342	HLA-DRB5	0.0211
14	rs1003879	C6orf10	0.0201	rs9469219	HLA-DQB1	0.0210
15	rs9271489	HLA-DRB1	0.0194	rs9271858	HLA-DRB1	0.0177
16	rs9272461	HLA-DQA1	0.0166	rs9275087	HLA-DQB1	0.0173
17	Affx-28,498,545	HLA-DQB1	0.0137	rs17843619	HLA-DQA1	0.0165
18	rs2395111	NOTCH4	0.0135	rs17843605	HLA-DQA1	0.0162
19	rs1049072	HLA-DQB1	0.0124	rs9273505	HLA-DQB1	0.0159
20	Affx-28,494,632	HLA-DQA1	0.0121	rs2894249	C6orf10	0.0158

ISSN: 1756-0381