scmap

scmap maps cells from query single-cell RNA sequencing (scRNA-seq) datasets onto reference cell types or individual cells to enable cross-experiment comparison of cellular identities.

Key Features:

• Data integration: Projects single-cell transcriptomic profiles from disparate scRNA-seq studies onto a reference to align datasets.
• Cell type mapping: Leverages unsupervised clustering techniques to map query cells to predefined cell types in a labeled reference dataset.
• Projection to individual cells: Supports mapping to individual reference cells as well as to labeled cell-type entries in the reference.
• Similarity metrics: Uses gene expression similarity metrics to identify matches between query and reference cells.
• Cross-experiment comparison: Addresses variability introduced by different experimental methods and computational pipelines to enable more reliable comparisons between studies.

Scientific Applications:

• Comparative studies: Compare cellular responses across different conditions or time points by projecting query scRNA-seq data onto a reference.
• Validation of cell types: Validate and refine cell-type annotations in new datasets by comparison with labeled reference datasets.
• Discovery of novel cell states: Reveal cell states not apparent within a single dataset by projecting onto a diverse reference set.

Methodology:

scmap identifies similarities between gene expression profiles of cells across experiments and projects query cells onto a labeled reference dataset using gene expression similarity metrics; reference dataset preparation requires labeled cell types or individual cells, and projection results are used to assign query cells to reference cell types or individual reference cells.