High-Throughput, Automated Image Processing for Large-Scale Fluorescence Microscopy Experiments

Novel image analysis tool for rapid screening of cell morphology in preclinical animal models of disease

The field of cell biology has seen major advances in both cellular imaging modalities and the development of automated image analysis platforms that increase rigor, reproducibility, and throughput for large imaging data sets. However, there remains a need for tools that provide accurate morphometric analysis of single cells with complex, dynamic cytoarchitecture in a high-throughput and unbiased manner. We developed a fully automated image-analysis algorithm to rapidly detect and quantify changes in cellular morphology using microglia cells, an innate immune cell within the central nervous system, as representative of cells that exhibit dynamic and complex cytoarchitectural changes. We used two preclinical animal models that exhibit robust changes in microglia morphology: (1) a rat model of acute organophosphate intoxication, which was used to generate fluorescently labeled images for algorithm development; and (2) a rat model of traumatic brain injury, which was used to validate the algorithm using cells labeled using chromogenic detection methods. All ex vivo brain sections were immunolabeled for IBA-1 using fluorescence or diaminobenzidine (DAB) labeling, images were acquired using a high content imaging system and analyzed using a custom-built algorithm. The exploratory data set revealed eight statistically significant and quantitative morphometric parameters that distinguished between phenotypically distinct groups of microglia. Manual validation of single-cell morphology was strongly correlated with the automated analysis and was further supported by a comparison with traditional stereology methods. Existing image analysis pipelines rely on high-resolution images of individual cells, which limits sample size and is subject to selection bias. However, our fully automated method integrates quantification of morphology and fluorescent/chromogenic signals in images from multiple brain regions acquired using high-content imaging. In summary, our free, customizable image analysis tool provides a high-throughput, unbiased method for accurately detecting and quantifying morphological changes in cells with complex morphologies.

Quantitative cell imaging approaches to metastatic state profiling

Genetic heterogeneity of metastatic dissemination has proven challenging to identify exploitable markers of metastasis; this bottom-up approach has caused a stalemate between advances in metastasis and the late stage of the disease. Advancements in quantitative cellular imaging have allowed the detection of morphological phenotype changes specific to metastasis, the morphological changes connected to the underlying complex signaling pathways, and a robust readout of metastatic cell state. This review focuses on the recent machine and deep learning developments to gain detailed information about the metastatic cell state using light microscopy. We describe the latest studies using quantitative cell imaging approaches to identify cell appearance-based metastatic patterns. We discuss how quantitative cancer biologists can use these frameworks to work backward toward exploitable hidden drivers in the metastatic cascade and pioneering new Frontier drug discoveries specific for metastasis.

Interpretable deep learning uncovers cellular properties in label-free live cell images that are predictive of highly metastatic melanoma

Deep learning has emerged as the technique of choice for identifying hidden patterns in cell imaging data but is often criticized as "black box." Here, we employ a generative neural network in combination with supervised machine learning to classify patient-derived melanoma xenografts as "efficient" or "inefficient" metastatic, validate predictions regarding melanoma cell lines with unknown metastatic efficiency in mouse xenografts, and use the network to generate in silico cell images that amplify the critical predictive cell properties. These exaggerated images unveiled pseudopodial extensions and increased light scattering as hallmark properties of metastatic cells. We validated this interpretation using live cells spontaneously transitioning between states indicative of low and high metastatic efficiency. This study illustrates how the application of artificial intelligence can support the identification of cellular properties that are predictive of complex phenotypes and integrated cell functions but are too subtle to be identified in the raw imagery by a human expert. A record of this paper's transparent peer review process is included in the supplemental information. VIDEO ABSTRACT.

New software for automated cilia detection in cells (ACDC)

Background

Primary cilia frequency and length are key metrics in studies of ciliogenesis and ciliopathies. Typically, quantitative cilia analysis is done manually, which is very time-consuming. While some open-source and commercial image analysis software applications can segment input data, they still require the user to optimize many parameters, suffer from user bias, and often lack rigorous performance quality assessment (e.g., false positives and false negatives). Further, optimal parameter combinations vary in detection accuracy depending on cilia reporter, cell type, and imaging modality. A good automated solution would analyze images quickly, robustly, and adaptably—across different experimental data sets—without significantly compromising the accuracy of manual analysis.

Methods

To solve this problem, we developed a new software for automated cilia detection in cells (ACDC). The software operates through four main steps: image importation, pre-processing, detection auto-optimization, and analysis. From a data set, a representative image with manually selected cilia (i.e., Ground Truth) is used for detection auto-optimization based on four parameters: signal-to-noise ratio, length, directional score, and intensity standard deviation. Millions of parameter combinations are automatically evaluated and optimized according to an accuracy ‘F1’ score, based on the amount of false positives and false negatives. Afterwards, the optimized parameter combination is used for automated detection and analysis of the entire data set.

Results

The ACDC software accurately and adaptably detected nuclei and primary cilia across different cell types (NIH3T3, RPE1), cilia reporters (AcTub, Smo-GFP, Arl13b), and image magnifications (60×, 40×). We found that false-positive and false-negative rates for Arl13b-stained cilia were 1–6%, yielding high F1 scores of 0.96–0.97 (max. = 1.00). The software detected significant differences in mean cilia length between control and cytochalasin D-treated cell populations and could monitor dynamic changes in cilia length from movie recordings. Automated analysis offered up to a 96-fold speed enhancement compared to manual analysis, requiring around 5 s/image, or nearly 18,000 cilia analyzed/hour.

Conclusion

The ACDC software is a solution for robust automated analysis of microscopic images of ciliated cells. The software is extremely adaptable, accurate, and offers immense time-savings compared to traditional manual analysis.

Electronic supplementary material

The online version of this article (10.1186/s13630-019-0061-z) contains supplementary material, which is available to authorized users.