BinPostHoc: Perform post-hoc analysis of perturbed bins

In BlishLab/scriabin: Single-cell resolved interaction analysis through binning

Perform post-hoc analysis of perturbed bins

Description

Perform post-hoc analysis of perturbed bins

Usage

BinPostHoc(
  seu,
  bin_pvals,
  interaction_graphs,
  split.by = "orig.ident",
  cell.type.calls = "celltype",
  kw_p.value = 0.001,
  bin_sd.quantile = 0
)

Arguments

seu	A binned Seurat object with bin identities in the "bin" column of meta.data
bin_pvals	Bin perturbation signifance data, ie. the output of PerturbedBins
interaction_graphs	List of summarized interaction graphs built from BuildPriorInteraction or BuiltWeightedInteraction
split.by	Meta.data column name indicating how data was split for interaction graph generation
cell.type.calls	Meta.data slot column corresponding to cell type annotations for summarization
kw_p.value	Bin-bin combinations with a KW p value above this threshold will be discarded. Default: 0.001
bin_sd.quantile	Bin-bin combinations with a summarized interaction standard deviation below this quantile will be discarded. Ensures that bin-bin combinations displaying both statistical significance and effect size of perturbation are analyzed. Default: 0.9.

Value

Performs Dunn's Kruskal-Wallis multiple comparison post-hoc test to evaluate which samples within perturbed bins are significantly perturbed

BlishLab/scriabin documentation built on Sept. 16, 2024, 1:19 a.m.