Figure 3
From: Ice thickness monitoring for cryo-EM grids by interferometry imaging
Validation using EM for the ice thickness determination. (a) Absolute ice thickness measurement using EM. Low magnification EM images showing holes in a grid square, and acquired with (left image) and without (right image) the activation of an energy filter. The local absolute ice thickness can be determined from the different intensities in given holes (blue and red circles) (Methods). (b) Correlation between the ANNICAS grid square classification and the ice layers measurements obtained by EM. Higher row ‘Optical images’: optical interferometric images for eight representative grid squares classified by ANNICAS. Middle row ‘EM images’: corresponding low-magnification EM images (if necessary after image rotation). (b) Representative ice thicknesses for these grid square (samples in the same order than in (c)) obtained by calculation of the average ice thickness (Methods) in 10 holes randomly distributed. The standard deviations of the thickness values for the 10 positions are indicated as bars. The standard deviations are large for many samples, reflecting the strong inhomogeneity of their respective ice layers. The ice thicknesses are consistent with our classification, confirming that our ANN-based analysis reliably detects and classifies the grid squares according to the ice thickness/quality. (d–e) High-resolution cryo-EM images in a class 2 (d), and in a class 4 (e) grid squares. SbcCD particles are distinctly visible in the left insert (white arrows), whereas the particles in too thick ice on the right are much more crowded, partially forming aggregates (grey arrow), and appear in addition with a lower contrast. Scale bars 20 nms.
