Table 2.
Number of genotype matches between PCR‐SSO, NGS‐454 and NGS‐MiSeq
| Locus | Pairwise comparisons between different typing techniques | ||||||
|---|---|---|---|---|---|---|---|
| PCR‐SSO vs NGS‐454 | NGS‐454 vs NGS‐MiSeq | PCR‐SSO vs NGS‐MiSeq | |||||
| HLA‐A | No. of genotyped individuals | 196 | 87 | ||||
| No. of compared individuals | 85 | ||||||
| No. of genotype matches (%) | 83 (97.6%) | ||||||
| CI95 | 97.0–100.0% | ||||||
| HLA‐B | No. of genotyped individuals | 198 | 83 | ||||
| No. of compared individuals | 82 | ||||||
| No. of genotype matches (%) | 75 (91.5%) | ||||||
| CI95 | 89.4–95.5% | ||||||
| HLA‐C | No. of genotyped individuals | 165 | 83 | ||||
| No. of compared individuals | 66 | ||||||
| No. of genotype matches (%) | 52 (78.8%) | ||||||
| CI95 | 72.7–72.7% | ||||||
| HLA‐DRB1 | No. of genotyped individuals | 198 | 194 | 194 | 81 | 198 | 81 |
| No. of compared individuals | 188 | 78 | 77 | ||||
| No. of genotype matches (%) | 135 (71.8%) | 70 (89.7%) | 47 (61.0%) | ||||
| CI95 | 63.6–80.3% | 87.9–92.4% | 57.5–65.2% | ||||
| HLA‐DQA1 | No. of genotyped individuals | 194 | 82 | ||||
| No. of compared individuals | 66 | ||||||
| No. of genotype matches (%) | 64 (97.0%) | ||||||
| CI95 | 97.0% | ||||||
| HLA‐DQB1 | No. of genotyped individuals | 195 | 196 | 196 | 76 | 195 | 76 |
| No. of compared individuals | 188 | 74 | 72 | ||||
| No. of genotype matches (%) | 153 (81.4%) | 68 (91.9%) | 13 (18.1%) | ||||
| CI95 | 72.7–87.9% | 90.9–93.9% | 15.2–19.7% | ||||
| HLA‐DPB1 | No. of genotyped individuals | 193 | 199 | 199 | 82 | 193 | 82 |
| No. of compared individuals | 193 | 82 | 79 | ||||
| No. of genotype matches (%) | 172 (89.1%) | 72 (87.8%) | 24 (30.3%) | ||||
| CI95 | 83.3–94.0% | 84.9–90.9% | 25.8–36.4% | ||||
No., Number; CI95, 95% confidence interval of the percentage of matches estimated by considering that the number of individuals compared between 2 different techniques is as small as the smallest number of individuals actually compared (ie, 66 individuals which is the number compared both for HLA‐C between PCR‐SSO and NGS‐MiSeq and for HLA‐DQA1 between NGS‐454 and NGS‐MiSeq); these intervals were obtained by drawing 1000 random samples of 66 individuals without replacement, see section 2.