Table 5.
Rank-based precision, sensitivity and F1-score values for the simulated reads at species level
| Reference | Method | Sensitivity (%) | Precision (%) | F1-score (%) |
|---|---|---|---|---|
| RefSeq-OLD | Centrifuge | 41.48 | 79.59 | 54.54 |
| Clark | 41.13 | 84.28 | 55.28 | |
| Diamond | 4.91 | 29.01 | 8.40 | |
| Ganon | 40.57 | 89.41 | 55.82 | |
| Kraken | 41.23 | 83.91 | 55.29 | |
| Kraken2 | 41.43 | 78.40 | 54.21 | |
| RefSeq-CG-top-3 | Centrifuge | 43.66 | 79.28 | 56.31 |
| Clark | 43.01 | 82.59 | 56.57 | |
| Diamond | 11.30 | 74.43 | 19.62 | |
| Ganon | 41.87 | 88.00 | 56.74 | |
| Kraken | 43.23 | 82.49 | 56.73 | |
| Kraken2 | 43.61 | 79.01 | 56.20 | |
| RefSeq-ALL-top-3 | Diamond | 13.10 | 88.64 | 22.82 |
| Ganon | 54.44 | 94.63 | 69.12 | |
| Kraken2 | 53.78 | 91.60 | 67.77 |
Note: Numbers in bold denote the best results for each reference set. The use of a larger reference set with RefSeq-ALL-top-3 significantly improves results. Only ganon and diamond indexed the RefSeq-ALL-top-3 in <24 h, thus, centrifuge, clark and kraken were excluded. Results for all taxonomic levels are in the Supplementary Figure S8 and Supplementary Material S2.