Search

Live-cell imaging experiments have opened an exciting window into the behavior of living systems. While these experiments can produce rich data, the computational analysis of these datasets is challenging. Single-cell analysis requires that cells be accurately identified in each image and subsequently tracked over time. Increasingly, deep learning is being used to interpret microscopy image with single cell resolution. In this work, we apply deep learning to the problem of tracking single cells in live-cell imaging data. Using crowdsourcing and a human-in-the-loop approach to data annotation, we constructed a dataset of over 11,000 trajectories of cell nuclei that includes lineage information. Using this dataset, we successfully trained a deep learning model to perform cell tracking within a linear programming framework. Benchmarking tests demonstrate that our method achieves state-of-the-art performance on the task of cell tracking with respect to multiple accuracy metrics. Further, we show that our deep learning-based method generalizes to perform cell tracking for both fluorescent and brightfield images of the cell cytoplasm, despite having never been trained those data types. This enables analysis of live-cell imaging data collected across imaging modalities. A persistent cloud deployment of our cell tracker is available at http://www.deepcell.org.

Deep learning is transforming the ability of life scientists to extract information from images. While these techniques have superior accuracy in comparison to conventional approaches and enable previously impossible analyses, their unique hardware and software requirements have prevented widespread adoption by life scientists. To meet this need, we have developed DeepCell 2.0, an open source library for training and delivering deep learning models with cloud computing. This library enables users to configure and manage a cloud deployment of DeepCell 2.0 on all commonly used operating systems. Using single-cell segmentation as a use case, we show that users with suitable training data can train models and analyze data with those models through a web interface. We demonstrate that by matching analysis tasks with their hardware requirements, we can efficiently use computational resources in the cloud and scale those resources to meet demand, significantly reducing the time necessary for large-scale image analysis. By reducing the barriers to entry, this work will empower life scientists to apply deep learning methods to their data. A persistent deployment is available at http://www.deepcell.org.