Table 1.
Reported performance of m6A modification identification achieved by existing works
| Method | AUC-ROC | |||
|---|---|---|---|---|
| Read-levela | Site-levela | Yeast KOb | Humanc | |
| EpiNano (2019) (Liu et al. 2019) | – | 0.90 | 0.680 | – |
| ELIGOS (2021) (Jenjaroenpun et al. 2021) | – | 0.756 | 0.287 (F1) | – |
| Nanocompore (2021) (Leger et al. 2021) | – | – | 0.18 (F1) | – |
| Nanom6A (2021) (Gao et al. 2021) | – | 0.97 | 0.71 | – |
| CHEUI (2022) (Acera Mateos et al. 2024) | 0.806 | 0.92 | – | – |
| m6Anet (2022) (Hendra et al. 2022) | 0.90 | 0.94 | – | 0.83 |
| Xron (this work) | 0.93 | >0.99 | 0.90 | 0.91 |
-
Bold values: P < 0.001 (***). aThese results were reported on the IVT data set (Liu et al. 2019), in which single-read m6A modifications were known.
-
bYeast ime4Δ knockout data set from Liu et al. (2019).
-
cHuman HEK293T cell data set from Chen et al. 2021.
