Table 11 Classification scores obtained for the C1 test set using different methods
From: Environmental metagenome classification for constructing a microbiome fingerprint
| Â | Â | NYC | OFA | PXO | SCL | Overall accuracy |
|---|---|---|---|---|---|---|
Harris et al. [32] | #correct | 0 | 5 | 10 | 5 | Â |
|  | PPV | — | 0.83 | 1.00 | 1.00 | ACC=0.667 |
| Â | TPR | 0.00 | 1.00 | 1.00 | 1.00 | Â |
Walker and Datta [22] | #correct | 4 | 1 | 9 | 2 | Â |
| Â | PPV | 0.67 | 1.00 | 0.75 | 0.50 | ACC=0.533 |
| Â | TPR | 0.40 | 0.20 | 0.90 | 0.40 | Â |
Zhu [25] | #correct | 3 | 5 | 7 | 4 | Â |
|  | PPV | — | 5.00 | 0.58 | 1.00 | ACC=0.633 |
| Â | TPR | 0.30 | 1.00 | 0.70 | 0.80 | Â |
Chierici et al. [23] | #correct | 10 | 0 | 10 | 5 | Â |
|  | PPV | 0.67 | — | 1.00 | 1.00 | ACC=0.833 |
| Â | TPR | 1.00 | 0.00 | 1.00 | 1.00 | Â |
Our method using Mash | #correct | 0 | 3 | 4 | 2 | Â |
sketch size=1000 | PPV | — | 1 | 1 | 0.5 | ACC=0.300 |
| Â | TPR | 0 | 0.6 | 0.4 | 0.4 | Â |
Our method using Mash | #correct | 0 | 3 | 6 | 5 | Â |
sketch size=10000 | PPV | — | 1 | 1 | 1 | ACC=0.467 |
| Â | TPR | 0 | 0.6 | 0.6 | 1 | Â |
Our method using Mash | #correct | 0 | 3 | 5 | 4 | Â |
sketch size=100000 | PPV | — | 1 | 1 | 1 | ACC=0.400 |
| Â | TPR | 0 | 0.6 | 0.5 | 0.8 | Â |
Our method using CoMeta | #correct | 10 | 4 | 2 | 4 | Â |
(class-level filtering) | PPV | 0.91 | 1.00 | 0.91 | 1.00 | ACC=0.667 |
| Â | TPR | 1.00 | 0.80 | 1.00 | 0.80 | Â |
Our method using CoMeta | #correct | 10 | 4 | 10 | 4 | Â |
(sample-level filtering) | PPV | 0.91 | 1.00 | 1.00 | 1.00 | ACC=0.933 |
| Â | TPR | 1.00 | 0.80 | 0.20 | 0.80 | Â |