MHC2MIL: a novel multiple instance learning based method for MHC-II peptide binding prediction by considering peptide flanking region and residue positions

Table 6 Comparisons of ARB, SMM-align, CombLib, MHC2SK, NN-algin and MHC2MIL on DP and DQ. For each allele, the largest value is displayed in boldface.

allele	ARB	SMM-align	CombLib	MHC2SK	NN-align	MHC2MIL
HLA-DPA1_01-DPB1_0401	0.746	0.767	0.704	0.778	0.802	0.791
HLA-DPA1_0103-DPB1_0201	0.745	0.767	0.724	0.789	0.793	0.808
HLA-DPA1_0201-DPB1_0101	0.743	0.786	0.723	0.824	0.818	0.834
HLA-DPA1_0201-DPB1_0501	0.709	0.728	0.729	0.794	0.787	0.805
HLA-DPA1_0301-DPB1_0402	0.771	0.818	0.756	0.838	0.828	0.846
HLA-DPB1_0301-DPB1_0401	-	-	-	0.792	-	0.790
HLA-DQA1_0101-DQB1_0501	0.741	0.783	0.728	0.789	0.805	0.781
HLA-DQA1_0102-DQB1_0602	0.708	0.734	0.752	0.770	0.762	0.769
HLA-DQA1_0301-DQB1_0302	0.637	0.663	0.616	0.676	0.693	0.706
HLA-DQA1_0401-DQB1_0402	0.643	0.761	0.637	0.767	0.742	0.774
HLA-DQA1_0501-DQB1_0201	0.700	0.736	0.620	0.775	0.777	0.810
HLA-DQA1_0501-DQB1_0301	0.756	0.801	0.745	0.807	0.811	0.817
average AUC	0.718	0.759	0.703	0.783	0.783	0.794

ISSN: 1471-2164