A repository with base code available for performing large scale three dimensional spatial (RDF) based analysis of Xenium transcriptomic data
Start by utilising the preprocessing utility scripts for downloading data from the HEST1k dataset, and saving locally as parquet files
Then run scripts found in individual_gene_pipeline or pairwise_gene_pipeline to evaluate the RDF with inhomogeneous null correction for spatially varying intensity functions via monte carlo simulation. Note that whilst the individual_gene_pipeline can run locally the pairwise evaluation will likely need extensive compute. For 422 genes across just two samples from HEST1k we required ~10HPC nodes with 120 cores and >400GB of RAM running for approximately 24 hours per node to complete the analysis.