THINKCYTE SCIENCE

Ghost Cytometry

Label-free, high throughput cell characterization and isolation based on high-resolution morphology and artificial intelligence.

Ghost Cytometry, the Science-published technology inside VisionSort, brings artificial intelligence to cytometry to enable high-throughput, data rich, and novel insights for life science R&D.

NEW ON-DEMAND WEBINAR

Label-free Use of Stem Cells with VisionSort

How Ghost Cytometry Works

A. Profile Acquisition

Cells pass through a structured illumination and morphological profiles are collected as high-dimensional temporal waveforms with a single pixel detector.

B. AI-based Classification

A trained AI model predicts the cell class based on the waveform.

C. Sorting

The classified cells are gently isolated using fluid pressure.

AI-Driven Cell Classification & Sorting

With Ghost Cytometry, harness the power of AI in your cellular analysis and sorting workflows using both supervised and unsupervised machine learning approaches. Match the approach with your downstream R&D goals and experience ultimate control with powerful algorithms and visualization tools embedded directly in VisionSort.

Supervised Machine Learning

Labeling

Morphological profiles are paired with fluorescently labeled markers for cell classes of interest.

Modeling

A machine-learning model is developed.

In Silico Prediction

The machine-learning model predicts cell classes by evaluating morphological profiles in unlabeled samples.

Validation

Isolated cell classes are validated with downstream tools such as multi-omics analyses and functional assays.

Unsupervised Machine Learning

Profiling

Morphological profiles are generated for each cell by Ghost Cytometry

Dimensional Reduction

Morphological profiles are collapsed using dimensional reduction methods such as UMAP. Users create gates to identify potentially unique subpopulations.

Model Generation & Sorting

Classification models for subpopulations are created to enable real time live cell sorting.