Skip to main content
. 2020 Jan 1;209(1):107402. doi: 10.1016/j.jsb.2019.10.004

Fig. 5.

Fig. 5

Seam check via supervised 3D classification. a) Supervised 3D classification strategy. Example simulated 13 PF CKK density references are shown with rotations around and translations along the helical axis of −2, −1, 0, +1 and +2 times the helical twist and rise. The resulting class reconstructions from a single classification iteration to these references are shown. b) Class occupancy for all 26 supervised classes used for the 13 PF CKK-MT dataset. i) Shows classes representing rotations around and translations along the helical axis (angle) of −6 to +6, whereas ii) shows classes with the same rotations and translations plus an additional translation of the monomer repeat distance (41 Å). For clarity, class occupancies as a % are indicated above bars representing classes. c) 4 × binned reconstructions of CKK-MT data after application of MiRP and its seam finding procedure or instead using standard helical processing in RELION without intervention. When using MiRP, CKK density is clearly absent from the seam and from 41 Å translated locations along the helical axis, while aberrant density is found at the seam and 41 Å translated locations when using standard helical processing, indicating poor MT Rot angle and αβ-tubulin register determination.