In silico-labeled ghost cytometry

Abstract

Characterization and isolation of a large population of cells are indispensable procedures in biological sciences. Flow cytometry is one of the standards that offers a method to characterize and isolate cells at high throughput. When performing flow cytometry, cells are molecularly stained with fluorescent labels to adopt biomolecular specificity which is essential for characterizing cells. However, molecular staining is costly and its chemical toxicity can cause side effects to the cells which becomes a critical issue when the cells are used downstream as medical products or for further analysis. Here, we introduce a high-throughput stain-free flow cytometry called in silico-labeled ghost cytometry which characterizes and sorts cells using machine-predicted labels. Instead of detecting molecular stains, we use machine learning to derive the molecular labels from compressive data obtained with diffractive and scattering imaging methods. By directly using the compressive ‘imaging’ data, our system can accurately assign the designated label to each cell in real time and perform sorting based on this judgment. With this method, we were able to distinguish different cell states, cell types derived from human induced pluripotent stem (iPS) cells, and subtypes of peripheral white blood cells using only stain-free modalities. Our method will find applications in cell manufacturing for regenerative medicine as well as in cell-based medical diagnostic assays in which fluorescence labeling of the cells is undesirable.

Tracking defined microbial communities by multicolor flow cytometry reveals tradeoffs between productivity and diversity

“…Our flow cytometry approach can also be expanded to track more than four species. Recent dramatic increases in the number of fluorochromes, the number of fluorescence detectors, and embedded image-based analysis have facilitated flow cytometry experiments with dozens of parameters ( McKinnon, 2018 ; Ugawa et al, 2021 ; Schraivogel et al, 2022 ). Our approach can thus be easily scaled to include additional species, evaluate more complex environments, and assay response of communities to a variety of disturbances.…”

Large‐scale investigation of zoonotic viruses in the era of high‐throughput sequencing

“…For example, "Ghost cytometry" reportedly sorts targeted cells in microseconds by machine learning using fluorescence signals or fluorescence label-free waveforms obtained from cells as input. 97,98 As another example, previous studies using prokaryotes reported that it is possible to discriminate gene expression patterns or species using Raman spectra signature. [99][100][101][102] Interestingly, a study using virus particles as samples has reported Raman spectra signatures specific for virus species or strains.…”

Pooled CRISPR screening of high-content cellular phenotypes by ghost cytometry

“…On the other hand, when suitable biomarkers or staining methods are not available and cell phenotypes can only be assessed without labeling, image analysis based on human-recognizable features can become challenging. To solve this issue, machine learning-based analysis of the label-free high-content cell phenotypes is an emerging, promising approach beyond the bias of human recognition ( 10 ) ( 11 ).…”

“…Here, we present a versatile approach to genome-wide pooled CRISPR screenings for not only fluorescence but also label-free high-content cell phenotypes using a machine vision-based cell sorter (MViCS), which we newly developed based on our ghost cytometry (GC) technologies. The heart of GC is a machine learning-based direct and integrative analysis of cellular morphological information without image production ( 11 ) ( 12 ). The high-content information is obtained by flowing cells through a structured light illumination and simultaneously detecting multiple different ghost motion imaging (GMI) waveforms in a temporal domain through different optical paths: fluorescence GMI (flGMI) waveforms excited by a 488 nm laser as well as forward scattering GMI (fsGMI), backscattering GMI (bsGMI), diffractive GMI (dGMI), and bright field GMI (bfGMI) waveforms are generated by a 405 nm laser are measured as analogs to their corresponding microscopic images (Fig.…”