runSpatialCorrelationAnalysis: Initiate gene clustering analysis based on spatial patterns

In kueckelj/SPATA2: A Toolbox for Spatial Gene Expression Analysis

Initiate gene clustering analysis based on spatial patterns

Description

This function screens a subset of genes and evaluates their spatial overlap by correlation- and subsequent clustering analysis. Results can be conveniently obtained or processed with additional functions such as clusterSpCorResults(), getGenes() or getSpCorResults().

Usage

runSpatialCorrelationAnalysis(
  object,
  of_sample = "",
  genes = 2000,
  genes_additional = NULL,
  threshold_stw = 0.5,
  threshold_stpv = 0.1,
  with_ties = TRUE,
  method_cor = "pearson",
  method_dist = "euclidean",
  mtr_name = NULL,
  verbose = TRUE
)

Arguments

object	A valid spata-object.
of_sample	This argument is currently inactive. It might be reactivated when spata-objects can store more than one sample.
genes	A numeric value (integer) or a character vector. Determines which genes are included in the correlation assessment. If specified as a numeric value the genes are sorted in a decreasing fashion corresponding to their variance across all barcode spots. Then the top n genes are included whereby n is equal to the specified numeric value. If specified as a character vector it's elements are considered to be gene names and all valid inputs are included.
genes_additional	Character vector of gene names. If genes is specified as a numeric value but you want certain genes to be included irrespective of their variance you can denote them here and they are added after the variance evaluation.
threshold_stw, threshold_stpv	Numeric values. Both values refer to the results of the shapiro-wilkinson test results for each gene. Before beeing sorted according to their variance you can use both arguments to filter for genes with a W-value bigger or equal to threshold_stw and a respective p-value lower or equal to threshold_stpv.
with_ties	Logical. If set to TRUE (the default) genes with equal variances are kept even if the total number of genes
method_dist	Character value or vector (see details for more). Denotes the distance methods (e.g. 'euclidean') to be used. Run validDistanceMethods() to obtain all valid input options.
mtr_name	Character value. The name of the matrix of interest.

Details

The overall expression matrix is filtered according to the input of argument genes, transposed and given to stats::cor(). The returned correlation matrix is given to stats::dist() to calculate the distance matrix needed for subsequent cluster analysis.

Use getGenes() and it's argument similar_to in order to get genes that feature a similar expression profile/pattern as a gene of interest.

Value

An updated spata-object.

kueckelj/SPATA2 documentation built on March 16, 2024, 10:25 a.m.