Related Experiment Video
Updated: Oct 9, 2025
09:31
Visualization of Gut Microbiota-host Interactions via Fluorescence In Situ Hybridization, Lectin Staining, and Imaging
Published on: July 9, 2021
8.8K
Decoding gut microbiota by imaging analysis of fecal samples.
Chikara Furusawa1,2, Kumi Tanabe1, Chiharu Ishii3
1Center for Biosystem Dynamics Research, RIKEN, Suita, Japan.
Iscience
|December 20, 2021
Summary
Monitoring gut microbiota dynamics is key for health. This study uses a deep convolution network on fecal images to predict microbial abundances, offering a simple, inexpensive analysis method.
Area of Science:
- Microbiology and Bioinformatics
- Artificial Intelligence in Healthcare
Background:
- The gut microbiota is essential for host health.
- Monitoring microbial population dynamics is crucial.
- Current methods can be complex and costly.
Purpose of the Study:
- To develop a novel method for characterizing gut microbiota dynamics.
- To utilize deep learning on low-resolution fecal images for microbial analysis.
- To quantitatively predict microbial relative abundances.
Main Methods:
- Development of a deep convolution network (CNN).
- Application of the CNN to low-resolution images of fecal samples.
- Validation of CNN predictions against 16S rRNA amplicon sequencing data.
Main Results:
- The deep convolution network successfully characterized dynamic changes in the gut microbiota.
- Microbial relative abundances were quantitatively predicted by the neural network.
- The method demonstrated a correlation with 16S rRNA amplicon sequencing results.
Conclusions:
- Deep learning on fecal images offers a viable approach for gut microbiota analysis.
- This method provides a simple and inexpensive alternative for monitoring microbial communities.
- The findings support the potential of AI in non-invasive health diagnostics.
