Compute Normalized Correlation (approximation)

This method calculates the normalized correlation between pairs of cell types based on CCA weights and the respective kernel matrix. It uses the spectral norm of the kernel matrix for normalization.

Usage

computeNormalizedCorrelation(object, tol = 1e-04, calculationMode = "perSlide")

# S4 method for class 'CoPro'
computeNormalizedCorrelation(object, tol = 1e-04)

# S4 method for class 'CoProMulti'
computeNormalizedCorrelation(object, tol = 1e-04, calculationMode = "perSlide")

Arguments

object

A CoPro or CoProMulti object containing CCA results and kernel matrices.

tol

tolerance for approximate SVD calculation

calculationMode

(for CoProMulti only) either "perSlide" or "aggregate", for single slide analysis, it is ignored, with default value "perSlide".

Value

The object with the normalized correlation value between any pair of cell types added as a new slot, normalizedCorrelation.

See also

runSkrCCA(), computeBidirCorrelation(), computeGeneAndCellScores()

Other scores-and-correlation: computeGeneAndCellScores()