Deep Learning Strategies for Differential Expansion Microscopy

Abstract

Differential expansion microscopy (DiExM) achieves greater than 500-fold volumetric expansion of biological specimens without loss of cellular antigens. The anisotropic character of this expansion, which affects tissues, cells, organelles and even sub-organelle features, requires the application of novel machine learning approaches to extract accurate and meaningful morphological and biochemical information from DiExM images. Here we describe current strategies to achieve this goal.