Fig. 3: Classification of flipped DNA Origami data and NPC data.
From: Detecting structural heterogeneity in single-molecule localization microscopy data

a Fusion result without classification of 456 DNA-origami structures of the TUD-logo with 80% DoL. b, c The two classes resulting from the classification of a, containing 446 (normal orientation) and 10 (flipped orientation) particles per class, respectively. d–f Same as a–c, but with 50% DoL and 381 and 8 particles per class, respectively. g, h The two classes resulting from the classification of i, containing 168 and 136 particles, respectively. We used K = 4 for the 80% DoL data and K = 40 for the 50% DoL data, followed by further grouping with the eigen image approach with C = 2 (Methods). i Fusion result without classification of 304 NPCs. j–l Same as g, h, but with classification into three classes, with 133, 90 and 81 particles per class, respectively. For g–l, the ellipticity values e are defined as the ratio of the major axis over the minor axis of the ellipse fitted to the localizations, and are indicated below each class. Scale bar of a applies to b–f. Scale bar of g applies to h–l.