scCube

scCube simulates spatially resolved transcriptomics data with diverse variability to enable benchmarking of computational methods while preserving spatial expression patterns.

Key Features:

• Reference-based simulation: Preserves spatial expression patterns of genes from reference data.
• Reference-free simulation: Generates simulated data with different spatial variability covering spatial pattern type, resolution, spot arrangement, targeted gene type, and tissue slice dimension.
• Technology-diverse simulation: Supports simulation across different spatially resolved transcriptomics technologies.
• Independent and reproducible simulation: Provides reproducible generation of simulated datasets as a Python package.

Scientific Applications:

• Spot deconvolution benchmarking: Evaluates computational methods for deconvolving mixed spot-level transcriptomic signals.
• Gene imputation benchmarking: Assesses methods for imputing missing or unmeasured gene expression values.
• Resolution enhancement benchmarking: Tests methods for increasing the spatial resolution of transcriptomics data.
• Simulator comparison: Enables comprehensive benchmarking against existing single-cell or spatially resolved transcriptomics simulators.

Methodology:

scCube implements reference-based simulations that preserve spatial expression patterns of genes, and reference-free simulations that generate data with controlled spatial variability across multiple parameters including spatial pattern type, resolution, spot arrangement, targeted gene type, and tissue slice dimension.