SingleCellExperiment

SingleCellExperiment represents a standardized S4 container for single-cell genomics data that integrates raw counts, normalized expression, reduced-dimensionality representations (PCA, t-SNE, UMAP), spike-in controls, size factors, and cell- and feature-level annotations to support analysis of high-dimensional, sparse single-cell datasets.

Key Features:

• S4 class: Defines an extensible S4 class for representing single-cell genomics data within the Bioconductor ecosystem.
• Assay storage: Stores raw counts, normalized expression values, and multiple layers of derived measurements in assay-like structures.
• Dimensionality reduction: Maintains dimensionality reduction coordinates such as PCA, t-SNE, and UMAP in dedicated slots.
• Spike-ins and size factors: Provides specialized storage for spike-in controls and per-cell size factors.
• Metadata handling: Enables efficient retrieval and manipulation of per-cell and per-feature metadata.
• Bioconductor interoperability: Extends and interoperates with existing Bioconductor infrastructure for downstream analyses.

Scientific Applications:

• Quality control: Supports QC workflows that use raw counts, spike-ins, size factors, and metadata to assess cell and feature quality.
• Normalization: Facilitates normalization workflows by storing normalized assays and size factors.
• Clustering: Enables clustering analyses using assay data and reduced-dimensionality representations.
• Trajectory inference: Supports trajectory and pseudotime inference via expression matrices and dimensionality reduction coordinates.
• Differential expression: Provides assay and metadata structures suitable for differential expression testing between cell groups.
• Integrative multi-omic analyses: Accommodates integrative analyses by holding multiple derived measurements and annotations per feature and cell.

Methodology:

Stores assays (raw counts and normalized values), records spike-in controls and size factors, keeps dimensionality reduction coordinates such as PCA, t-SNE, and UMAP in dedicated slots, and exposes per-cell and per-feature metadata for retrieval and manipulation within Bioconductor S4 infrastructure.