Calculate Cross-Modality Correlation Fidelity — calc_cross_modality

Evaluates whether cross-modality feature relationships (e.g. peak-to-gene links, promoter accessibility vs. gene expression, or mRNA vs. surface protein) are accurately captured by the simulation method compared to the empirical reference.

Usage

calc_cross_modality_correlation(
  ref_mod1,
  ref_mod2,
  sim_mod1,
  sim_mod2,
  feature_pairs = NULL,
  method = c("spearman", "pearson")
)

Arguments

ref_mod1: Reference matrix for Modality 1 (features x cells).
ref_mod2: Reference matrix for Modality 2 (features x cells).
sim_mod1: Simulated matrix for Modality 1.
sim_mod2: Simulated matrix for Modality 2.
feature_pairs: Optional 2-column data.frame/matrix of paired feature names/indices.
method: Correlation method: "spearman" (default) or "pearson".

Value

A named list of the 7 univariate accuracy metrics comparing the reference vs. simulated cross-modality correlation distributions.

Examples

data(example_multiomics, package = "scSimEval")
m <- example_multiomics
calc_cross_modality_correlation(m$ref_multi$rna, m$ref_multi$atac,
                                m$sim_multi$rna, m$sim_multi$atac)
#> Metric: 'bhattacharyya' with unit: 'log'; comparing: 2 vectors
#> $cross_modality_cor_MAD
#> [1] 0.06236202
#> 
#> $cross_modality_cor_KS
#> [1] 0.3096934
#> 
#> $cross_modality_cor_MAE
#> [1] 0.06352891
#> 
#> $cross_modality_cor_RMSE
#> [1] 0.06626403
#> 
#> $cross_modality_cor_OV
#> [1] 0.7564079
#> 
#> $cross_modality_cor_Bhattacharyya
#> [1] 0.001537345
#> 
#> $cross_modality_cor_Wasserstein
#> [1] 0.06352891
#> 
#> $cross_modality_cor_ECDF_DiffArea
#> [1] 0.1199979
#> 
#> $cross_modality_cor_Runs_Statistic
#> [1] -3.300231
#> 
#> $cross_modality_cor_Runs_PValue
#> [1] 0.0004830262
#> 
#> $cross_modality_cor_NN_Mismatch
#> [1] 0.1166667
#> 
#> $cross_modality_cor_Between_Dataset_Silh_Global
#> [1] 0.05927015
#> 
#> $cross_modality_cor_Between_Dataset_Silh_Local
#> [1] 0.1521508
#>