Although 3D reconstruction methods for EM were developed almost 40 years ago (DeRosier and Klug 1968), obtaining high-resolution information with this technique seemed impossible until 2 decades later, tte reason was related to the use of negative staining, which largely altered the relation between the sample in its native state and its images. Another limitation came from the low spatial coherence provided by the available electron sources, thermoionic electron guns, tte low coherence produced a fast decay in the signal amplitude, so much so that the images barely contained any information beyond 20-A resolution, tte qualitative jump to high resolution came with the popularization of field-emission guns as electron sources, ttese guns produce highly coherent electron beams, and consequently the images contain useful information all the way to high frequencies.
One problem that cryo-EM does not solve is the lack of contrast of biological samples, ttis becomes even worse, because images of frozen hydrated samples are taken with very low electron doses to minimize radiation damage, and the scattering powers of ice and biological material are very similar, tte result is that cryo-EM images are very noisy, with a very low signal-to-noise ratio. To overcome this difficulty, a great number of images from chemically identical specimens are averaged. Averaging requires all the images to be first taken into register, since the different individual specimens will in general lie in random positions on the specimen support. In addition, it is necessary to use robust classification techniques to ensure that the average comes from a homogeneous data set. Sophisticated image restoration procedures have to be applied to compensate for the electron microscope aberrations. Several thousands, or tens of thousands, of individual images need to be processed in this way before obtaining a satisfactory 3D reconstruction, making 3D-EM a very computationally demanding technique. Ms is why the swift evolution of computers and computational methods has been another key factor for the takeoff of EM as a high-resolution technique in biology.
Was this article helpful?