Convolutional neural network approach for the automated identification of in cellulo crystals.

In cellulo crystallization is a rare event in nature. Recent advances that have made use of heterologous overexpression can promote the intracellular formation of protein crystals, but new tools are required to detect and characterize these targets in the complex cell environment. The present work makes use of Mask R-CNN, a convolutional neural network (CNN)-based instance segmentation method, for the identification of either single or multi-shaped crystals growing in living insect cells, using conventional bright field images. The algorithm can be rapidly adapted to recognize different targets, with the aim of extracting relevant information to support a semi-automated screening pipeline, in order to aid the development of the intracellular protein crystallization approach.

Development towards high-resolution kHz-speed rotation-free volumetric imaging.

X-ray multi-projection imaging (XMPI) has the potential to provide rotation-free 3D movies of optically opaque samples. The absence of rotation enables superior imaging speed and preserves fragile sample dynamics by avoiding the centrifugal forces introduced by conventional rotary tomography. Here, we present our XMPI observations at the ID19 beamline (ESRF, France) of 3D dynamics in melted aluminum with 1000 frames per second and 8 µm resolution per projection using the full dynamical range of our detectors. Since XMPI is a method under development, we also provide different tests for the instrumentation of up to 3000 frames per second. As the high-brilliance of 4th generation light-sources becomes more available, XMPI is a promising technique for current and future X-ray imaging instruments.