view article
Figure 11
Impact of the number of predicted components. We report cumulative curves for LS error (a, b) and magnitude error (c, d). For each protein, we computed the error either for the best-matching pair of predicted and ground-truth vectors (a, c) or for the best combination of all pairs of predicted and ground-truth vectors using optimal linear assignment (b, d). We compare models trained to predict different numbers of components (modes): 1, 2, 4 or 8, using only the LS loss.

Journal logoSTRUCTURAL
BIOLOGY
ISSN: 2059-7983
Volume 82| Part 8| August 2026|
https://doi.org/10.1107/S2059798326006054
Follow Acta Cryst. D
Sign up for e-alerts
E-alerts
Follow Acta Cryst. on Twitter
Twitter
Follow us on facebook
Facebook
Sign up for RSS feeds
RSS