Table 2.
Comparison of allele vectors in spaces Y and Z.
| HLA | Numerical | Categorical | ||
| allele | R2 (%) | Sensitivity (%) | Specificity (%) | Likelihood Ratio |
| A1 | 27.1 | 50 | 97 | 15 |
| A2 | 71.9 | 67 | 100 | |
| A3 | 63.7 | 50 | 98 | 22 |
| A11 | 40.5 | 50 | 97 | 18 |
| A24 | 25.8 | 100 | 97 | 36 |
| A29 | 12.7 | 67 | 95 | 13 |
| A31 | 26.8 | 100 | 97 | 30 |
| A33 | 16.0 | 100 | 96 | 25 |
| B7 | 36.6 | 60 | 94 | 10 |
| B8 | 14.4 | 50 | 95 | 10 |
| B14 | 22.9 | 50 | 94 | 9 |
| B27 | 71.8 | 100 | 99 | 86 |
| B35 | 36.9 | 100 | 97 | 30 |
| B38 | 0.2 | 0 | 93 | |
| B44 | 18.3 | 25 | 96 | 6 |
| B51 | 48.6 | 100 | 98 | 45 |
| B53 | 22.4 | 67 | 97 | 20 |
| B54 | 40.9 | 67 | 98 | 30 |
| ALL | 27.7 | 64 | 96 | 18 |
'R2' stands for the coefficient of determination in the regression analysis of numerical values of the vectors in spaces Y and Z (see text) for each allele. (In all cases the p value was 0.000, except for B38, which was 0.51.) Allele vectors values in Y and Z were converted to categorical values (0, 1) using cut-off values and pairs of vectors for each allele were compared using a 2 × 2 contingency table. 'Sensitivity', 'specificity', and 'likelihood ratio' refer to the prediction of a Z vector by the corresponding Y vector, and were calculated from these 2 × 2 tables.