SPARK-X

SPARK-X detects spatially variable gene expression in large-scale spatial transcriptomic datasets using non-parametric modeling to provide scalable and robust identification of spatial expression patterns.

Key Features:

• Non-Parametric Methodology: Employs a non-parametric approach to model spatial expression without relying on predefined distributional assumptions.
• Scalability and Efficiency: Provides substantial computational efficiency, reducing processing time by orders of magnitude for large spatial transcriptomic datasets.
• Statistical Rigor: Controls type I error rates while maintaining high statistical power for detecting spatially expressed genes.

Scientific Applications:

• Identification of spatially expressed genes: Detects spatial expression patterns across large spatial transcriptomic datasets to discover spatially expressed genes.
• Within–cell-type spatial heterogeneity: Identifies spatially expressed genes within the same cell type to reveal intra-cell-type spatial organization.
• Biological interpretation of cellular organization: Enables insights into cellular organization and function derived from spatial gene expression patterns.

Methodology:

Applies non-parametric modeling to rapidly process large volumes of spatial transcriptomic data while addressing statistical and computational challenges in detecting spatial gene expression patterns.