Prediction of peptides binding to MHC class I and II alleles by temporal motif mining

BMC Bioinformatics

Table 1 Results of MHC-PPM in class I predictions in Peters dataset [22].

	Allele	# Peptides		ANN	ARB	SMM	MHC-PPM
		9	10
9-mers + 10-mers	HLA-A*0201	3089	1316	-	0.919	0.939	0.931
	HLA-A*0202	1447	1056	-	0.851	0.879	0.871
	HLA-A*0203	1443	1055	-	0.838	0.878	0.882
	HLA-A*0301	2094	1082	-	0.883	0.915	0.911
	HLA-A*0206	1437	1054	-	0.849	0.890	0.885
	HLA-A*1101	1985	1093	-	0.897	0.932	0.937
	HLA-A*2402	197	78	-	0.722	0.809	0.833
	HLA-A*3101	1869	1057	-	0.881	0.903	0.878
	HLA-A*3301	1140	1055	-	0.866	0.888	0.863
	HLA-A*6801	1141	1055	-	0.827	0.874	0.864
	HLA-B*0702	1262	205	-	0.925	0.952	0.954
	HLA-B*3501	736	177	-	0.833	0.886	0.866
	HLA-B*5101	244	177	-	0.782	0.875	0.886
	HLA-B*5301	254	177	-	0.758	0.854	0.847
9-mers	HLA-A*0101	1157	-	0.982	0.964	0.980	0.963
	HLA-A*2601	672	-	0.956	0.907	0.931	0.901
	HLA-A*2902	160	-	0.935	0.755	0.911	0.907
	HLA-A*6802	1434	-	0.899	0.865	0.898	0.867
	HLA-B*0801	708	-	0.955	0.936	0.943	0.926
	HLA-B*1501	978	-	0.941	0.900	0.952	0.922
	HLA-B*1801	118	-	0.838	0.573	0.853	0.906
	HLA-B*2705	969	-	0.938	0.915	0.940	0.938
	HLA-B*4002	118	-	0.754	0.541	0.842	0.891
	HLA-B*4402	119	-	0.778	0.533	0.740	0.891
	HLA-B*4403	119	-	0.763	0.461	0.770	0.847
	HLA-B*5401	255	-	0.903	0.847	0.921	0.883
	HLA-B*5701	59	-	0.826	0.428	0.871	0.929
	HLA-B*5801	988	-	0.961	0.889	0.964	0.944
	Average (All)			0.888	0.798	0.893	0.897
	Average (9mers)			0.888	0.751	0.894	0.908
	Weighted Avg			0.932	0.872	0.910	0.901

The best-performing method for each allele is underlined. The given AUC values for ARB and SMM are the weighted averages of the AUC values for 9-mers and 10-mers based on the given peptide counts for a specific allele. The alleles in the bottom part of the table were only trained & tested in 9-mers and are directly comparable.

ISSN: 1471-2105