A modular platform for automated cryo-FIB workflows

Abstract

Lamella micromachining by focused ion beam milling at cryogenic temperature (cryo-FIB) has matured into a preparation method widely used for cellular cryo-electron tomography. Due to the limited ablation rates of low Ga+ ion beam currents required to maintain the structural integrity of vitreous specimens, common preparation protocols are time-consuming and labor intensive. The improved stability of new generation cryo-FIB instruments now enables automated operations. Here, we present an open-source software tool, SerialFIB, for creating automated and customizable cryo-FIB preparation protocols. The software encompasses a graphical user interface for easy execution of routine lamellae preparations, a scripting module compatible with available Python packages, and interfaces with 3-dimensional correlative light and electron microscopy (CLEM) tools. SerialFIB enables the streamlining of advanced cryo-FIB protocols such as multi-modal imaging, CLEM-guided lamella preparation and in situ lamella lift-out procedures. Our software therefore provides a foundation for further development of advanced cryogenic imaging and sample preparation protocols.

Label-free visual proteomics: Coupling MS- and EM-based approaches in structural biology

“…To address cryo-FIB-milling throughput, software automation has recently been introduced both commercially and as open source packages (Klumpe et al, 2021), new FIB-milling modalities have been developed to increase the size of the lamellae (Schaffer et al, 2019;Kelley et al, 2020), and commercial companies are currently exploring using plasma sources for FIB-milling instead of gallium, which is expected to increase throughput by several factors. Faster and more robust lamellae production will require an increase in cryo-ET tilt-series collection throughput, which may be overcome by the BISECT method (Bouvette et al, 2021) for parallel tilt-series collection.…”

“…Cryo-FIB-milling can create lamellae from almost any type of biological sample, including whole cells, vitrified organisms, and tissues. While practically difficult, cryo-FIB-milling is being rapidly developed with the introduction of automated milling software solutions and new instrumentation (Klumpe et al, 2021;Tacke et al, 2021) as well as advanced vitrification and milling techniques (Schaffer et al, 2019;Kelley et al, 2020). Due to its ability to create near-native cross-sections of biological specimens that allow for the highest-resolution in situ 3D visualization by cryo-ET tomograms, cryo-FIB-milling is rapidly being turned to by biologists and biomedical scientists.…”

Convolutional networks for supervised mining of molecular patterns within cellular context

“…k-n . Overlay of ribosomes within 25 nm distance to the ER (orange) with ER-bound ribosomes from DeePiCt predictions in DEF ( k , gray), from VPP gt data ( l , gray), with the published density of S. cerevisiae ribosomes derived from extracted ER 10 ( k , gray, CC 0.8972, EMD-3764), and the ER-bound HeLa ribosome 11 ( n , gray, CC 0.9105, EMD-8056). All densities show the same interface between ribosome and ER membrane.…”

“… a . Performance results of five DeePiCt networks trained on a variety of cell types for the segmentation of different cellular structures used for predictions in a HeLa cell dataset 11 . Left: plot of performance scores when employing the spectrum equalization (+EQ) filter in both the training and on the HeLa cell data.…”

“…However, fine structural details that provide insights into functional configurations of molecular complexes are lost or cannot be recovered in the final reconstructions of VPP tomography data 9 . This has been suggested to be caused by inaccurate weighing 9 and a signal loss at high spatial frequencies 11 . The preservation of high-resolution information in DEF in comparison to VPP becomes especially apparent when looking at fine structural details in the subtomogram averages.…”