Deep Learning Automates the Quantitative Analysis of Individual Cells in Live-Cell Imaging Experiments
| Autor: | Nicolas Quach, Markus W. Covert, Takamasa Kudo, David Van Valen, Inbal Maayan, Euan A. Ashley, Keara Michelle Lane, Yu Tanouchi, Derek N. Macklin, Mialy M. DeFelice |
|---|---|
| Rok vydání: | 2016 |
| Předmět: |
0301 basic medicine
Fluorescence-lifetime imaging microscopy Cytoplasm Intravital Microscopy Convolutional neural network Pattern Recognition Automated Diagnostic Radiology Machine Learning 0302 clinical medicine Fluorescence Microscopy Medicine and Health Sciences Segmentation Computer vision lcsh:QH301-705.5 Microscopy Ecology Artificial neural network Radiology and Imaging Light Microscopy Bone Imaging Living systems In Vivo Imaging Computational Theory and Mathematics Cell Tracking Modeling and Simulation Cellular Structures and Organelles Research Article Computer and Information Sciences Neural Networks Imaging Techniques Biology Research and Analysis Methods Sensitivity and Specificity 03 medical and health sciences Cellular and Molecular Neuroscience Live cell imaging Diagnostic Medicine Image Interpretation Computer-Assisted Fluorescence Imaging Genetics Molecular Biology Ecology Evolution Behavior and Systematics Bacteria business.industry Deep learning Organisms Reproducibility of Results Biology and Life Sciences Pattern recognition Image segmentation Cell Biology Image Enhancement 030104 developmental biology lcsh:Biology (General) Artificial intelligence Neural Networks Computer business 030217 neurology & neurosurgery Neuroscience |
| Zdroj: | PLoS Computational Biology PLoS Computational Biology, Vol 12, Iss 11, p e1005177 (2016) |
| ISSN: | 1553-7358 |
| Popis: | Live-cell imaging has opened an exciting window into the role cellular heterogeneity plays in dynamic, living systems. A major critical challenge for this class of experiments is the problem of image segmentation, or determining which parts of a microscope image correspond to which individual cells. Current approaches require many hours of manual curation and depend on approaches that are difficult to share between labs. They are also unable to robustly segment the cytoplasms of mammalian cells. Here, we show that deep convolutional neural networks, a supervised machine learning method, can solve this challenge for multiple cell types across the domains of life. We demonstrate that this approach can robustly segment fluorescent images of cell nuclei as well as phase images of the cytoplasms of individual bacterial and mammalian cells from phase contrast images without the need for a fluorescent cytoplasmic marker. These networks also enable the simultaneous segmentation and identification of different mammalian cell types grown in co-culture. A quantitative comparison with prior methods demonstrates that convolutional neural networks have improved accuracy and lead to a significant reduction in curation time. We relay our experience in designing and optimizing deep convolutional neural networks for this task and outline several design rules that we found led to robust performance. We conclude that deep convolutional neural networks are an accurate method that require less curation time, are generalizable to a multiplicity of cell types, from bacteria to mammalian cells, and expand live-cell imaging capabilities to include multi-cell type systems. Author Summary Dynamic live-cell imaging experiments are a powerful tool to interrogate biological systems with single cell resolution. The key barrier to analyzing data generated by these measurements is image segmentation—identifying which parts of an image belong to which individual cells. Here we show that deep learning is a natural technology to solve this problem for these experiments. We show that deep learning is more accurate, requires less time to curate segmentation results, can segment multiple cell types, and can distinguish between different cell lines present in the same image. We highlight specific design rules that enable us to achieve high segmentation accuracy even with a small number of manually annotated images (~100 cells). We expect that our work will enable new experiments that were previously impossible, as well as reduce the computational barrier for new labs to join the live-cell imaging space. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|