Fig. 6: Evaluation of the performance of CASSPER cross model. | Communications Biology

Fig. 6: Evaluation of the performance of CASSPER cross model.

From: CASSPER is a semantic segmentation-based particle picking algorithm for single-particle cryo-electron microscopy

Fig. 6

Panels a–d show representative unlabeled micrograph (scale bar, 50 nm), semantically segmented and labeled micrograph using CASSPER cross model and AUC curves for proteins; a TcdA1 (EMPIAR 10089), b Plasmodium falciparum 80S ribosome bound to the anti-protozoan drug emetine (EMPIAR 10028), c afTMEM16/nanodisc complex (EMPIAR 10240), and d human Bact spliceosome (EMPIAR 10160), respectively. The high values of AUC score indicate the generalization ability of CASSPER. The PR curve is obtained by comparing the coordinates with manually picked ground truth labels.

Back to article page