logo
ResearchBunny Logo
Transformer for One-Stop Interpretable Cell Type Annotation

Biology

Transformer for One-Stop Interpretable Cell Type Annotation

J. Chen, H. Xu, et al.

TOSICA, developed by Jiawei Chen and colleagues, revolutionizes cell type annotation in single-cell research with its Transformer-based model. This innovative approach not only ensures fast and accurate identification but also enhances interpretability, shedding light on rare cell types and their behavior, especially in tumor-infiltrating immune cells and COVID-19 monocytes.

00:00
00:00
Playback language: English
Abstract
Consistent annotation transfer from reference to query datasets is crucial for single-cell research reproducibility. Deep learning offers faster, more automated annotation than traditional methods, but existing autoencoder-based tools struggle with interpretability. This paper introduces TOSICA, a Transformer-based model enabling interpretable cell type annotation using pathways or regulons. TOSICA achieves fast, accurate, batch-insensitive integration and annotation, providing biological insights into cellular behavior. Applications to tumor-infiltrating immune cells and COVID-19 monocytes demonstrate TOSICA's ability to reveal rare cell types and heterogeneity.
Publisher
Nature Communications
Published On
Jan 14, 2023
Authors
Jiawei Chen, Hao Xu, Wanyu Tao, Zhaoxiong Chen, Yuxuan Zhao, Jing-Dong J. Han
Tags
single-cell research
deep learning
cell type annotation
interpretable model
TOSICA
biological insights
tumor-infiltrating immune cells
Listen, Learn & Level Up
Over 10,000 hours of research content in 25+ fields, available in 12+ languages.
No more digging through PDFs, just hit play and absorb the world's latest research in your language, on your time.
listen to research audio papers with researchbunny