SCOPE

SCOPE performs normalization and copy number estimation for whole genome single-cell DNA sequencing (scDNA-seq) data to address sparsity, noise, and variability introduced during library preparation and sequencing.

Key Features:

• Cell-specific Gini coefficients: Utilizes Gini coefficients for quality controls and identification of normal/diploid cells, which serve as negative control samples in a Poisson latent factor model for normalization.
• GC content bias modeling: Models GC content bias using an expectation-maximization algorithm embedded in Poisson generalized linear models, accounting for different copy number states along the genome.
• Cross-sample iterative segmentation: Identifies breakpoints shared across cells from the same genetic background through a cross-sample iterative segmentation procedure.

Scientific Applications:

• Cancer subclone deconvolution: Enables characterization of copy number profiles at the cellular level to deconvolve cancer subclones with increased resolution and decreased ambiguity compared to bulk-tissue sequencing.
• Cancer evolutionary history: Elucidates cancer evolutionary history through single-cell resolution of copy number alterations.

Methodology:

SCOPE employs a Poisson latent factor model for normalization using normal/diploid cells as negative controls, an expectation-maximization algorithm embedded in Poisson generalized linear models for GC content bias correction, and a cross-sample iterative segmentation procedure for breakpoint detection.