Table 3.
Sensitivity and false discovery rate (FDR) of the genotyping method.
| Marker | TP | FN | FP | Sensitivity | FDR | Detected Haplotypesa | ||
|---|---|---|---|---|---|---|---|---|
| n | ni | niia | niib | n | TP/(TP + FNi+iiab) | FP/(TP + FP) | (TP + FNi)/(TP + FNi+iiab) | |
| msp2-CE | 86 | 10 | 5 | n/ab | n/ac | 0.851 ± 0.101d | n/ac | 0.950 ± 0.061d |
| cpmp | 115 | 4 | 2 | 1 | 5 | 0.943 ± 0.066 | 0.042 ± 0.057 | 0.975 ± 0.044 |
| ama1-D2 | 109 | 3 | 0 | 1 | 1 | 0.965 ± 0.052 | 0.009 ± 0.027 | 0.991 ± 0.026 |
| ama1-D3 | 108 | 4 | 2 | 1 | 3 | 0.939 ± 0.068 | 0.027 ± 0.046 | 0.974 ± 0.045 |
Sensitivity and FDR including 95% confidence interval was estimated based on persistent clones in 48 longitudinal samples from 12 individuals. Detectability of minority clone can be increased by including missed persistent haplotypes detected below the cut-off criteria. TP, true-positive haplotypes. FNi, false-negative haplotypes detected, but below cut-off criteria. FNiiab, false-negative haplotypes with no read detected.
aDetected true-positive and false-negative haplotypes.
bNot imputed for msp2-CE as multi-locus haplotypes cannot be established.
cLength-polymorphic data generated in different laboratories do not provide replicates suited for determination of false-positive haplotype calls and estimation of FDR.
dWithout haplotypes, that were imputed based on multi-locus haplotypes at the beginning or end of an infection.