SpaCCI (version 1.0.4)

run_SpaCCI: Run SpaCCI Analysis

Description

This function runs the SpaCCI analysis to infer cell-cell interactions based on ligand-receptor pairs at global, regional, or local spatial scales. It integrates gene expression data, cell type proportions, and spatial coordinates with a user-specified ligand-receptor database.

Usage

run_SpaCCI(
  gene_spot_expression_dataframe,
  spot_cell_proportion_dataframe,
  spatial_coordinates_dataframe,
  LR_database_list,
  specific_LR_pair = NULL,
  analysis_scale,
  region_spot_IDs = NULL,
  local_scale_proportion = 1,
  neighborhood_radius = 2.5
)

Value

A list containing:

If analysis_scale is "local":

A list containing:

dataframelist

A list of data frames, each representing the inferred interactions for a specific center spot. Each data frame includes information on ligand and receptor cell types, P-values, and adjusted P-values.

RegionIDs_matrix

A list of matrices, each containing the IDs of the spots within the specified radius of each center spot.

If analysis_scale is "regional" or "global":

A list containing:

pvalue_df

A data frame of inferred interactions within the specified region or globally, including information on ligand and receptor cell types, P-values, and adjusted P-values.

Arguments

gene_spot_expression_dataframe

A data frame of gene expression values, where row names are genes and column names are spot IDs.

spot_cell_proportion_dataframe

A data frame of cell type proportions, where row names are spot IDs and column names are cell types.

spatial_coordinates_dataframe

A data frame containing the spatial coordinates of the spots. The columns should include "Spot_ID", "imagerow", and "imagecol". And the row names must be the names of "Spot_ID".

LR_database_list

A list containing ligand-receptor pairs and additional information, generated by functions using `LR_database()`.

specific_LR_pair

Required if analysis_scale is "local". A vector of ligand-receptor pair names for localized analysis. The names should match those in row names in `LR_database_list$possible_LR_pairs_info`.

analysis_scale

A string specifying the scale of analysis: "global", "regional", or "local".

region_spot_IDs

Required if analysis_scale is "regional". A vector of spot IDs defining the region for regional analysis.

local_scale_proportion

Optional. A numeric value ranging from 0 to 1, (0,1] specifying the proportion of spots to use for localized analysis. Default is 1, meaning using 100% proportion of spots. One could modified if want to reducing computing time.

neighborhood_radius

Optional. A numeric value specifying the radius of the neighborhood for localized analysis. Default is 2.5, according to the 10X Visium ST data accounting for 200-250 \(\mu\)m interacting distance.

Details

The function supports three scales of analysis:

global

Analyzes interactions across the entire dataset.

regional

Analyzes interactions within a specified region of spots. Requires region_spot_IDs.

local

Analyzes localized hotspot of interactions for specific ligand-receptor pairs on the entire slides. Requires specific_LR_pair.

Examples

library(SpaCCI)
library(nnls)
library(FNN)
library(Matrix)
#Load the example data
data(test_data)
gene_spot_df <- test_data$gene_spot_df
cell_prop_df <- test_data$cell_prop_df
spatial_coords_df <- test_data$spatial_coords_df

result <- LR_database(species = "Human",
                      database_name = "CellChat",
                      gene_spot_expression_dataframe = gene_spot_df)
# global
result_global <- run_SpaCCI(gene_spot_expression_dataframe = gene_spot_df,
                            spot_cell_proportion_dataframe = cell_prop_df,
                            spatial_coordinates_dataframe = spatial_coords_df,
                            LR_database_list = result,
                            analysis_scale = "global")

# local
result_local <- run_SpaCCI(gene_spot_expression_dataframe = gene_spot_df,
                          spot_cell_proportion_dataframe = cell_prop_df,
                          spatial_coordinates_dataframe = spatial_coords_df,
                          LR_database_list = result,
                          specific_LR_pair = "EDN1_EDNRA",
                          analysis_scale = "local",
                          local_scale_proportion = 0.1,
                          neighborhood_radius = 2.5)