Robust Integration of Single-cell Protein Measurement

Overview: Our project introduces ADTnorm, a comprehensive computational framework for normalizing and integrating antibody-derived tag (ADT) data from CITE-seq experiments. CITE-seq technology enables simultaneous measurement of surface protein and mRNA expression in single cells, but batch effects and technical variations have posed significant challenges in data interpretation and cross-study analyses.

Motivation: Single-cell multimodal profiling has revolutionized our understanding of cellular heterogeneity, but technical variations in antibody staining can obscure biological signals. Our research addresses these challenges through:

1. Development of a robust normalization method specifically designed for ADT abundance data
2. Implementation of automated threshold-gating and quality assessment tools
3. Creation of utilities for antibody panel optimization and experimental design

We collaborate with multiple research groups, combining expertise in statistics, computational biology, and single-cell genomics. Our method has been extensively validated across 13 public datasets, demonstrating superior performance in aligning cell populations and removing batch effects compared to existing methods.

The project’s success enables more accurate integration of public CITE-seq datasets, facilitating atlas-level analyses and broader biological applications. Our application to COVID-19 CITE-seq data has already revealed novel disease-associated markers, highlighting the method’s potential impact on biomedical research. For specific information, please visit our ADTnorm.

Publications

⁺: co-corresponding author *: co-first author

1. Zheng Y*, Caron D*, Kim J, Jun S, Tian Y, Florian M, Stuart K, Sims P, Gottardo R. ADTnorm: Robust Integration of Single-cell Protein Measurement across CITE-seq Datasets. Revised for Nature Communications.(2024).