Related Experiment Video
Updated: Jan 17, 2026
Droplet Barcoding-Based Single Cell Transcriptomics of Adult Mammalian Tissues
Published on: January 10, 2019
Computational tracking of cell origins using CellSexID from single-cell transcriptomes
Huilin Tai1, Qian Li2, Jingtao Wang3
1Meakins-Christie Laboratories, Research Institute of McGill University Health Centre, 1001 Decarie Boulevard, Montreal, QC H4A 3J1, Canada; Department of Mathematics and Statistics, McGill University, 845 Sherbrooke Street West, Montreal, QC H3A 0G4, Canada; Department of Computer Science, Columbia University, New York, NY, USA.
CellSexID uses sex as a marker to track cells in mixed populations. This computational tool accurately distinguishes donor from recipient cells, offering a practical alternative to physical labeling in research.
Area of Science:
- Developmental Biology
- Regenerative Medicine
- Transplantation Research
Background:
- Cell tracking in chimeric models is crucial but challenging.
- Existing methods like fluorescent labeling are costly and complex.
- Dynamic tissues require innovative cell tracking solutions.
Purpose of the Study:
- To introduce CellSexID, a computational framework for inferring cell origin.
- To utilize sex as a surrogate marker for distinguishing cell populations.
- To provide a practical alternative to physical cell labeling.
Main Methods:
- Machine learning models trained on single-cell transcriptomic data.
- Ensemble feature selection to identify minimal sex-linked gene sets.
- Validation using public datasets and experimental flow sorting.
Main Results:
- CellSexID accurately predicts individual cell sex.
- Enables in silico distinction of donor and recipient cells in sex-mismatched settings.
- Demonstrated applicability in chimeric models, organ transplantation, and sample demultiplexing.
Conclusions:
- CellSexID offers a cost-effective and practical approach to cell tracking.
- Facilitates precise cell tracking in diverse biomedical applications.
- Supports the distinction of mixed cellular populations without physical labeling.
