Three dimensional cross-modal image inference: label-free methods for subcellular structure prediction

Abstract

Fluorescence microscopy has enabled imaging of key subcellular structures in living cells; however, the use of fluorescent dyes and proteins is often expensive, time-consuming, and damaging to cells. Here, we present a tool for the prediction of fluorescently labeled structures in live cells solely from 3D brightfield microscopy images. We show the utility of this approach in predicting several structures of interest from the same static 3D brightfield image, and show that the same tool can prospectively be used to predict the spatiotemporal position of these structures from a bright-field time series. This approach could also be useful in a variety of application areas, such as cross-modal image registration, quantification of live cell imaging, and determination of cell state changes.