CMDdemux: an efficient single cell demultiplexing method
2025-12-12
Preface
Multiplexing technology is widely used in single-cell experiments. It involves labeling cells from different samples with chemical tags, such as antibody-based tags (Stoeckius et al. 2018), lipid (McGinnis et al. 2019) or cholesterol tags (McGinnis et al. 2019), or fluorescent proteins (Guo et al. 2019). After labeling, cells from different samples are pooled together for sequencing. During downstream analysis, the abundance of these tags (often called hashtags) is used to determine the sample of origin for each single cell. Multiplexing has several advantages. It increases sample throughput and the number of cells that can be profiled in a single sequencing run. It also reduces batch effects and lowers experimental costs. In addition, this experimental approach is generally more robust for detecting doublets compared to methods based solely on transcriptomic features. However, multiplexing can sometimes result in low-quality data that cannot be effectively demultiplexed by many existing computational methods. Our method, CMDdemux, is designed to demultiplex both high-quality and low-quality multiplexed datasets with high accuracy and efficiency. In this tutorial, we demonstrate the use of CMDdemux on both types of data.
