Table 1.
Comparison of the prediction accuracy between the software programs studied for each dataset and HLA gene region
| Dataset | Gene (n) | HLA-VBSeq, % | HLA-VBSeq v2, % | HLA*PRG:LA, % | |
|---|---|---|---|---|---|
| SJS | HLA-A | 228 | 98.2 | 99.1 | 98.2 |
| HLA-B | 228 | 92.5 | 97.8 | 100 | |
| HLA-C | 228 | 90.4 | 100.0 | 98.7 | |
| THC | HLA-A | 836 | 98.4 | 99.0 | 98.1 |
| HLA-B | 836 | 91.3 | 98.3 | 99.6 | |
| HLA-C | 818 | 92.1 | 99.8 | 98.4 | |
n corresponds to the number of alleles in each dataset
THC Tokyo Healthy Controls data set, SJS Stevens–Johnson syndrome data set
Bold font corresponds to the highest prediction accuracy for each HLA gene