RankMap

RankMap performs reference-based cell type annotation of single-cell and spatial transcriptomics data using ranked gene expression and multinomial regression to achieve robust predictions with partial gene coverage.

Key Features:

• Rank-based representation: Transforms gene expression profiles into rank-based representations using the top expressed genes per cell, improving robustness to platform-specific biases and expression scale differences.
• Multinomial regression with elastic net regularization: Predicts cell types and associated confidence scores using a multinomial regression model trained with elastic net regularization.
• Flexible preprocessing: Compatible with Seurat, SingleCellExperiment, and SpatialExperiment objects.
• Scalable runtime: Achieves significantly faster runtime than tools like SingleR, Azimuth, and RCTD, particularly for large spatial datasets.

Scientific Applications:

• Spatial transcriptomics annotation: Annotates cell types in spatial transcriptomics datasets including Xenium, MERFISH, and Stereo-seq platforms.
• Single-cell transcriptomics annotation: Annotates cell types in single-cell transcriptomics datasets.
• Partial gene panel analysis: Provides robust predictions for platforms with partial gene coverage where full-transcriptome profiles are unavailable.

Methodology:

RankMap transforms gene expression profiles into rank-based representations using the top expressed genes per cell, then applies a multinomial regression model trained with elastic net regularization to predict cell types and associated confidence scores.