scSpatialSIM

scSpatialSIM simulates spatial single-cell molecular data to assess the power of statistical and bioinformatic methods used to understand tissue architecture and cell interactions.

Key Features:

• Tissue-scale simulation: Simulates single-cell molecular data to mimic real tissues at scale, including clustering of cell types and co-clustering or co-localization of two or more cell types.
• Quantitative distribution modeling: Simulates quantitative distributions for positive and negative cells, such as gene expression and fluorescence intensity.
• Kernel density simulation: Simulates various kernel densities of probability distributions used to create marked point patterns with numeric or categorical features.
• Spatial statistic evaluation: Provides a platform to compare spatial clustering measures including Ripley's K(r), nearest neighbor G(r), and pair correlation g(r).

Scientific Applications:

• Spatial method benchmarking: Assesses and compares statistical and bioinformatic methods for analyzing spatial contexture of tissues.
• Spatial clustering analysis: Evaluates spatial clustering measures under varying simulation scenarios with different cell abundances and cluster sizes.
• Integration with spatial analysis pipelines: Generates simulated data compatible with other R packages such as spatialTIME for deriving spatial statistics.

Methodology:

scSpatialSIM uses kernel densities of probability distributions to create marked point patterns, with points distributed in space carrying either numeric or categorical features. The package implements four univariate spatial simulation scenarios to evaluate spatial clustering measures.