Semi-supervised integration of single-cell transcriptomics data

Batch effects in single-cell RNA-seq data pose a significant challenge for comparative analyses across samples, individuals, and conditions. Although batch effect correction methods are routinely applied, data integration often leads to overcorrection and can result in the loss of biological variabi...

Full description

Saved in:
Bibliographic Details
Published in:Nature communications 2024-01, Vol.15 (1), p.872-13, Article 872
Main Authors: Andreatta, Massimo, Hérault, Léonard, Gueguen, Paul, Gfeller, David, Berenstein, Ariel J., Carmona, Santiago J.
Format: Article
Language:English
Subjects:
38/39
38/91
631/114
631/114/2397
631/337/2019
631/553
631/61/514/1949
Cancer research
Cells
Comparative analysis
Data integration
Datasets
Gene Expression Profiling - methods
Humanities and Social Sciences
Humans
Integration
Labels
Medical research
Methods
multidisciplinary
Science
Science (multidisciplinary)
Sequence Analysis, RNA - methods
Single-Cell Analysis - methods
Transcriptomics
Citations: Items that this one cites
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Batch effects in single-cell RNA-seq data pose a significant challenge for comparative analyses across samples, individuals, and conditions. Although batch effect correction methods are routinely applied, data integration often leads to overcorrection and can result in the loss of biological variability. In this work we present STACAS, a batch correction method for scRNA-seq that leverages prior knowledge on cell types to preserve biological variability upon integration. Through an open-source benchmark, we show that semi-supervised STACAS outperforms state-of-the-art unsupervised methods, as well as supervised methods such as scANVI and scGen. STACAS scales well to large datasets and is robust to incomplete and imprecise input cell type labels, which are commonly encountered in real-life integration tasks. We argue that the incorporation of prior cell type information should be a common practice in single-cell data integration, and we provide a flexible framework for semi-supervised batch effect correction. Batch effects hinder multi-sample single-cell data analyses. Here, authors present STACAS, a scalable single-cell RNA-seq data integration tool that uses prior cell type knowledge to preserve biological variability, demonstrating robustness to noisy input cell type labels.
ISSN:2041-1723
2041-1723
DOI:10.1038/s41467-024-45240-z