Related Experiment Video
Updated: Oct 2, 2025
05:45
Author Spotlight: Integrating Organoid Models with Single-Cell and Spatial Transcriptomics Technologies
Published on: March 29, 2024
2.7K
CellDART: cell type inference by domain adaptation of single-cell and spatial transcriptomic data.
Sungwoo Bae1,2, Kwon Joong Na3,4, Jaemoon Koh5
1Department of Molecular Medicine and Biopharmaceutical Sciences, Graduate School of Convergence Science and Technology, Seoul National University, Seoul, Republic of Korea.
Nucleic Acids Research
|February 22, 2022
Summary
CellDART, a new computational method, accurately maps cell types in tissues using spatial transcriptomics. This approach enhances understanding of cellular composition and spatial interactions within complex biological systems.
Area of Science:
- Computational Biology
- Genomics
- Neuroscience
Background:
- Understanding cellular composition in spatial transcriptomics is crucial for tissue context.
- Existing methods face challenges in accurately predicting cell distribution.
Purpose of the Study:
- To develop CellDART, a novel method for estimating cell spatial distribution from single-cell data.
- To apply CellDART for spatial mapping in human lung and other tissues.
Main Methods:
- Developed CellDART using domain adaptation of neural networks.
- Trained a neural network to predict cell proportions in pseudospots.
- Applied the method to mouse brain, human prefrontal cortex, and human lung tissues.
Main Results:
- CellDART accurately identified layer-specific cell distributions in brain tissues.
- Demonstrated higher accuracy and stability than other methods for excitatory neuron prediction.
- Successfully decomposed cellular proportions in mouse hippocampus and elucidated cell type predominance in human lung.
Conclusions:
- CellDART provides a robust and efficient tool for spatial cell type deconvolution.
- The method aids in understanding tissue heterogeneity and cell-cell interactions.
- Expected to advance research in various tissue-based biological studies.
