Image-based phenotyping of disaggregated cells using deep learning

Abstract

The ability to phenotype cells is fundamentally important in biological research and medicine. Current methods rely primarily on fluorescence labeling of specific markers. However, there are many situations where this approach is unavailable or undesirable. Machine learning has been used for image cytometry but has been limited by cell agglomeration and it is currently unclear if this approach can reliably phenotype cells that are difficult to distinguish by the human eye. Here, we show disaggregated single cells can be phenotyped with a high degree of accuracy using low-resolution bright-field and non-specific fluorescence images of the nucleus, cytoplasm, and cytoskeleton. Specifically, we trained a convolutional neural network using automatically segmented images of cells from eight standard cancer cell-lines. These cells could be identified with an average F1-score of 95.3%, tested using separately acquired images. Our results demonstrate the potential to develop an “electronic eye” to phenotype cells directly from microscopy images.

Deep Learning Image Classification of Red Blood Cell Deformability

“…Our application of machine learning in RBC deformability measurement deviates from these previous efforts because cellular features corresponding to deformability are beyond human perception. This result expands on our previous study using deep learning to distinguish between cell lines that lack readily differentiable features to a human observer, 54 further supporting our belief that imperceivable cellular parameters, such as changes in biophysical or metabolic cell state, may be measurable using deep learning. Potential future applications of this work include assessment of RBC units prior to transfusion in order to preferentially allocate stored blood bags to the most appropriate recipients.…”

“…We recently developed a microfluidic process for deformability-based sorting of RBCs [32][33][34] as well as a deep learning method to distinguish cell lines based on feature differences imperceptible to human cognition. 54 We hypothesized that a combination of these advances could enable the direct measurement of RBC deformability by cell imaging using optical microscopy.…”

ImmuNet: A Segmentation-Free Machine Learning Pipeline for Immune Landscape Phenotyping in Tumors by Muliplex Imaging

“…Cell segmentation is a notoriously difficult problem in biomedical image processing, especially in dense tissue specimens [6]. The biomedical imaging community has devoted significant efforts to cell segmentation, which has been the subject of several benchmark datasets and algorithm competitions.…”

Hydrothermal polymerization of porous aromatic polyimide networks and machine learning-assisted computational morphology evolution interpretation

“…Despite pseudo-features not containing any physical meaning, the input of two similar images should result in similar pseudo-feature values, which allows for clustering the input images as recently shown for the classication of cell microscopy images. 66 The employed image treatment and analysis pipeline is displayed in Fig. 7A.…”