Related Experiment Video
Updated: Jun 26, 2025
Adhesion Frequency Assay for In Situ Kinetics Analysis of Cross-Junctional Molecular Interactions at the Cell-Cell Interface
Published on: November 2, 2011
Single-cell classification based on label-free high-resolution optical data of cell adhesion kinetics
Kinga Dora Kovacs1,2, Balint Beres1,3, Nicolett Kanyo1
1Nanobiosensorics Laboratory, Institute of Technical Physics and Materials Science MFA, HUN-REN Centre for Energy Research, Konkoly-Thege út 29-33, 1121, Budapest, Hungary.
This study introduces a label-free optical biosensor method to classify healthy and cancerous cells by analyzing single-cell adhesion kinetics. The non-invasive technique achieves high accuracy using deep learning, enabling applications in cell therapy and cancer research.
Area of Science:
- Biomedical Engineering
- Cell Biology
- Computational Biology
Background:
- Cell selection is crucial for therapies and research but often involves invasive labeling methods.
- Current techniques can compromise cell viability and functionality.
- A need exists for non-invasive, high-accuracy cell classification methods.
Purpose of the Study:
- To develop and validate a label-free optical biosensor for classifying healthy and cancerous mammalian cells.
- To assess the performance of deep learning models in analyzing single-cell adhesion kinetics for cell type identification.
- To gain interpretable biological insights from the classification models.
Main Methods:
- Utilized an optical biosensor to monitor single-cell adhesion kinetics in a label-free manner.
- Collected a benchmark database of ~4500 single-cell adhesion measurements from various cell types.
- Applied deep learning models, including grad-CAM visualization, for classification and insight generation.
Main Results:
- Achieved high classification accuracy (70-80%) between different cell types based on adhesion kinetics.
- Demonstrated the capability of deep neural networks to classify cells using solely label-free kinetic data.
- Identified key features driving classification decisions through model visualization.
Conclusions:
- The label-free optical biosensor method is a successful proof-of-concept for non-invasive single-cell classification.
- This approach offers a promising tool for cell profiling and in vitro cancer research.
- The non-invasive nature makes it suitable for downstream applications like cell therapy and regenerative medicine.
