Table 4.
Approximate computational time of AlphaCryo4D (grey boxes) on the simulated 2-million image dataset with the SNR of 0.01 and three experimental datasets.
| Particle Number | Consensus Alignment by RELION (h) | 3D Volumes Resampling (h) | Autoencoder Training (h) | Manifold Learning (h) | High-Resolution Refinement by RELION (h) | |
|---|---|---|---|---|---|---|
| Simulated dataset with the SNR of 0.01 | 2,000,000 | 75 | 10 | 3 | 15 | 110 |
| Yeast Spliceosome (EMPIAR-10180) |
327,490 | \ | 48 | 4 | 0.1 | 30 |
|
Pf80S Ribosome (EMPIAR-10028) |
105,247 | 80 | 30 | 2 | 0.1 | 70 |
| Bacterial Ribosome (EMPIAR-10076) |
131,899 | 160 | 38 | 4 | 0.2 | 105 |