Table 9:
Accuracy of algorithms by refined lineage.
| Algorithm | Overall | EAI | East-Asian | Euro-American | Indo-Oceanic | M. africanum 1 | M. africanum 2 | M. bovis | M. caprae |
|---|---|---|---|---|---|---|---|---|---|
| MIRU-VNTRplus (*) | 49.8 [98.5] | 44.1 | 72.6 | 39.4 | 61.1 | 25.0 | 100 | 100 | 100 |
| Same, no threshold (*) | 97.4 | 91.2 | 99.1 | 99.7 | 100 | 25.0 | 100 | 100 | 100 |
| MIRU-VNTRplus (†) | 91.9 [97.0] | 93.4 | 99.7 | 95.5 | 75.2 | 45.3 | 40.0 | 100 | 50.0 |
| Same, no threshold (†) | 96.5 | 96.3 | 99.7 | 99.8 | 87.8 | 45.3 | 40.0 | 100 | 100 |
| TBminer - MIRU-VNTRplus | 97.0 | 88.2 | 100 | 100 | 100 | 25.0 | 100 | 100 | 0 |
| TBminer - Expert | 96.8 | 91.2 | 100 | 100 | 96.3 | 25.0 | 100 | 100 | 0 |
| TBminer - MIRU-VNTRplus (‡) | 97.0 | 88.2 | 100 | 100 | 100 | 25.0 | 100 | 100 | 0 |
| TBminer - Expert (‡) | 96.8 | 88.2 | 98.6 | 100 | 100 | 25.0 | 100 | 100 | 0 |
| RuleTB | 94.0 | 88.2 | 93.1 | 96.0 | 98.1 | 50.0 | 50.0 | 100 | 100 |
| StackTB | 98.0 | 91.2 | 100 | 100 | 100 | 50.0 | 50.0 | 100 | 100 |
(*)
= the MIRU-VTNRplus database is used for training;
(†)
= the training subset of the entire data is used for training;
‡
= only the 15 most discriminative VNTR loci are used for the prediction. The numbers in square brackets indicate the accuracy of only the assignments actually made (excluding NAs). Note that TB-Insight is omitted because it does not identify lineages at this level of re nement.