scran

scran provides normalization and analysis methods for single-cell RNA sequencing (scRNA-seq) data to correct technical variation and enable quality control, cell-cycle phase assignment, detection of highly variable and correlated genes, clustering, and marker-gene identification.

Key Features:

• Normalization of Cell-Specific Biases: Implements normalization techniques to correct cell-specific biases in scRNA-seq gene-level count data.
• Cell Cycle Phase Assignment: Assigns cell cycle phase to individual cells to account for cell-cycle-related transcriptional variation.
• Detection of Highly Variable and Correlated Genes: Identifies genes with high variability and significant correlation across single cells for downstream analysis.
• Quality Control and Data Exploration: Provides quality-control procedures and initial data-exploration steps for assessing scRNA-seq dataset integrity.
• Clustering and Marker Gene Detection: Supports clustering of cells into subpopulations and detection of marker genes for cell-type characterization.

Scientific Applications:

• Gene-level count data analysis: Applied to gene-level count matrices from publicly available single-cell datasets.
• Haematopoietic stem cells: Demonstrated on scRNA-seq data from haematopoietic stem cell studies.
• Brain-derived cells: Demonstrated on scRNA-seq data from brain-derived cell populations.
• T-helper cells: Demonstrated on scRNA-seq data from T-helper cell studies.
• Mouse embryonic stem cells: Demonstrated on scRNA-seq data from mouse embryonic stem cells.

Methodology:

Computational workflow implemented with Bioconductor packages, including quality control, normalization to correct cell-specific biases, cell cycle phase assignment, identification of highly variable and correlated genes, clustering, and marker gene detection.