Related Experiment Video
Updated: Aug 15, 2025
03:37
Author Spotlight: Impact of Intergenic Interactions on Disease-Identifying Dark Biomarkers
Published on: March 1, 2024
837
Robotic data acquisition with deep learning enables cell image-based prediction of transcriptomic phenotypes
Jianshi Jin1, Taisaku Ogawa1, Nozomi Hojo1
1Laboratory for Prediction of Cell Systems Dynamics, RIKEN Center for Biosystems Dynamics Research (BDR), Suita, Osaka 565-0874, Japan.
Summary
Researchers developed an automated live imaging and cell picking system (ALPS) for real-time, noninvasive single-cell whole-transcriptome analysis. This system links cell images to transcriptomes, enabling phenotype prediction from images using deep learning.
Area of Science:
- Cell Biology
- Genomics
- Robotics
Background:
- Single-cell whole-transcriptome analysis is crucial for defining cell phenotypes.
- Current methods are incompatible with dynamic measurements like live-cell imaging.
Purpose of the Study:
- To develop a system for real-time, noninvasive single-cell RNA sequencing compatible with live-cell imaging.
- To enable prediction of cell phenotypes from microscopic images.
Main Methods:
- Development of the automated live imaging and cell picking system (ALPS).
- Integration of multiple imaging modes with single-cell RNA sequencing.
- Application of cell image-based deep learning to predict transcriptome-defined phenotypes.
Main Results:
- Successful linkage of cell images with whole-transcriptome data.
- Noninvasive prediction of transcriptome-defined cell phenotypes using deep learning on cell images.
- Demonstration of real-time whole-transcriptome analysis for live cells.
Conclusions:
- The ALPS system enables noninvasive, real-time transcriptome analysis of live cells.
- This data-driven approach provides a proof of concept for predicting cell phenotypes from images.
- The methodology is adaptable for various cell types and transcriptome-defined phenotypes.
