TY  - JOUR
A1  - Zhou, Yang
A1  - Sheng, Qiongyu
A1  - Qi, Jing
A1  - Hua, Jiao
A1  - Yang, Bo
A1  - Wan, Lei
A1  - Jin, Shuilin
T1  - Accurate integration of multiple heterogeneous single-cell RNA-seq data sets by learning contrastive biological variation
Y1  - 2023/05/01 
JF  - Genome Research 
JO  - Genome Research 
SP  - 750 
EP  - 762 
DO  - 10.1101/gr.277522.122 
VL  - 33 
IS  - 5 
UR  - http://genome.cshlp.org/content/33/5/750.abstract 
N2  - For most biological and medical applications of single-cell transcriptomics, an integrative study of multiple heterogeneous single-cell RNA sequencing (scRNA-seq) data sets is crucial. However, present approaches are unable to integrate diverse data sets from various biological conditions effectively because of the confounding effects of biological and technical differences. We introduce single-cell integration (scInt), an integration method based on accurate, robust cell–cell similarity construction and unified contrastive biological variation learning from multiple scRNA-seq data sets. scInt provides a flexible and effective approach to transfer knowledge from the already integrated reference to the query. We show that scInt outperforms 10 other cutting-edge approaches using both simulated and real data sets, particularly in the case of complex experimental designs. Application of scInt to mouse developing tracheal epithelial data shows its ability to integrate development trajectories from different developmental stages. Furthermore, scInt successfully identifies functionally distinct condition-specific cell subpopulations in single-cell heterogeneous samples from a variety of biological conditions. 
ER  -