Table 2.
Comparison of CryoTransformer with CrYOLO and Topaz’s performance in terms of the resolution of 3D density maps reconstructed for six test proteins from the particles picked from a small set of micrographs in the CryoPPP.
| EMPIAR ID | Number of micrographs | Number of particles |
3D resolution (Å) |
||||
|---|---|---|---|---|---|---|---|
| CrYOLO | Topaz | CryoTransformer | CrYOLO | Topaz | CryoTransformer | ||
| 10017 (Scheres 2015) | 84 | 283 | 98 625 | 43 735 | 10.4 | 5.3 | 5.61 |
| 10081 (Lee and MacKinnon 2017) | 300 | 17 550 | 111 752 | 88 632 | 9.78 | 7.81 | 5.47 |
| 10093 (Jin et al. 2017) | 295 | 8802 | 257 490 | 151 545 | 8.8 | 6.06 | 6.86 |
| 10345 (Campbell et al. 2020) | 295 | 4095 | 93 699 | 105 739 | 10.2 | 8.12 | 6.43 |
| 10532 (Tan and Rubinstein 2020) | 300 | 12 166 | 356 222 | 148 345 | 15.69 | 5.47 | 3.92 |
| 11056 (Asami et al. 2022) | 305 | 46 582 | 144 600 | 98 193 | 10 | 8.34 | 7.42 |
Bold font denotes the best resolution of the density map reconstructed from picked particles in the three trials.