Related Experiment Video
Updated: Oct 31, 2025
Multiplexed Analysis of Retinal Gene Expression and Chromatin Accessibility Using scRNA-Seq and scATAC-Seq
Published on: March 12, 2021
Simultaneous deep generative modeling and clustering of single cell genomic data.
Qiao Liu1,2, Shengquan Chen1, Rui Jiang1
1Ministry of Education Key Laboratory of Bioinformatics, Research Department of Bioinformatics at the Beijing National Research Center for Information Science and Technology, Center for Synthetic and Systems Biology, Department of Automation, Tsinghua University, Beijing 100084, China.
scDEC, a novel deep learning tool, enhances single-cell ATAC-seq analysis by addressing data challenges. It effectively infers cell types and variations, aiding in understanding cell differentiation and multi-modal data integration.
Area of Science:
- Genomics and Bioinformatics
- Computational Biology
- Epigenetics
Background:
- Single-cell technologies like single-cell ATAC-seq (scATAC-seq) offer chromatin accessibility profiling at an unprecedented resolution.
- High sparsity and dimensionality in scATAC-seq data present significant computational challenges for analysis.
- Existing computational tools struggle to fully leverage the rich information within scATAC-seq datasets.
Purpose of the Study:
- To introduce scDEC, a computational tool utilizing deep generative neural networks for single-cell ATAC-seq data analysis.
- To develop a method capable of simultaneously learning latent representations and inferring cell labels from scATAC-seq data.
- To demonstrate the utility of scDEC in downstream applications, including trajectory inference and multi-modal data integration.
Main Methods:
- Development of scDEC, a tool based on a pair of generative adversarial networks (GANs).
- Application of scDEC to multiple scATAC-seq datasets across various experimental conditions.
- Evaluation of scDEC's performance against existing computational tools for scATAC-seq analysis.
Main Results:
- scDEC demonstrates superior performance compared to other tools in scATAC-seq analysis.
- The generative capabilities of scDEC facilitate the inference of cell differentiation trajectories and intermediate cell states.
- Latent features learned by scDEC effectively reveal distinct biological cell types and intra-cell-type variations.
Conclusions:
- scDEC provides a robust and effective computational framework for analyzing challenging scATAC-seq data.
- The tool's ability to capture latent representations aids in understanding cellular heterogeneity and dynamics.
- scDEC is extendable for integrative analysis of multi-modal single-cell data, opening new avenues for research.
