Convolutional neural network approach for the automated identification ofin cellulocrystals

In cellulocrystallization is a rarely occurring event in nature. Recent advances, making use of heterologous overexpression, can promote the intracellular formation of protein crystals, but new tools are required to detect and to characterize these targets in the complex cell environment. In the present work we make 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 to extract relevant information to support a semi-automated screening pipeline, with the purpose to aid in the development of the intracellular protein crystallization approach.