scUmaper: An automated framework for doublet removal and cell-type annotation in single-cell transcriptomics.

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Single-cell RNA sequencing (scRNA-seq) offers high resolution for cellular heterogeneity.
• scRNA-seq data is susceptible to technical artifacts like heterotypic doublets.
• Current preprocessing and annotation methods often demand significant user expertise.

Purpose of the Study:

• To introduce scUmaper, an R/Seurat-native workflow.
• To integrate quality control, biologically grounded doublet filtering, and marker-library-based cell-type annotation.
• To enhance the reproducibility and accessibility of scRNA-seq data analysis.

Main Methods:

• scUmaper utilizes lineage-marker incompatibility rules for doublet detection.
• It employs global clustering followed by within-lineage re-clustering.
• The workflow integrates quality control and cell-type annotation using marker libraries.

Main Results:

• scUmaper identified and removed high-confidence heterotypic doublets missed by simulation-based methods.
• Achieved annotation agreement comparable to or exceeding common R-based baselines.
• Demonstrated stable performance under simulated ambient RNA contamination and reduced sequencing depth.

Conclusions:

• scUmaper provides an interpretable and extensible framework for scRNA-seq analysis.
• The workflow effectively addresses challenges in doublet filtering and cell-type annotation.
• scUmaper lowers barriers for reproducible single-cell RNA sequencing analysis.