Perform PIPET Screening Analysis

Predicts cell subpopulations in single-cell data by matching expression profiles to predefined marker gene templates using various distance/similarity metrics. This function implements a template-based classification approach with permutation testing for significance assessment.

Usage

DoPIPET(
  matched_bulk,
  sc_data,
  phenotype,
  phenotype_class = c("binary", "continuous", "survival"),
  group = NULL,
  discretize_method = c("kmeans", "median", "custom"),
  cutoff = NULL,
  label_type = "PIPET",
  marker_finder = c("limma", "DESeq2"),
  log2FC = 1L,
  p_adjust = 0.05,
  show_log2FC = TRUE,
  freq_counts = NULL,
  normalize = TRUE,
  scale = TRUE,
  nPerm = 1000L,
  distance = c("cosine", "pearson", "spearman", "kendall", "euclidean", "maximum"),
  ...
)

Arguments

matched_bulk

Normalized bulk expression matrix (features × samples). Column names must match phenotype identifiers.

sc_data

Seurat object containing single-cell RNA-seq data.

phenotype

Clinical outcome data. Can be: - Vector: named with sample IDs - Data frame: with row names matching bulk columns

phenotype_class

Analysis mode: - "binary": Case-control design (e.g., responder/non-responder) - "continuous": Continuous outcome (e.g., age, size) - "survival": Patient survival

group

A character, name of one metadata column to group cells by (for example, orig.ident). The default value is NULL. In this case, screening will be performed on each group separately.

discretize_method

c("median", "kmeans", "custom"). Discretization strategy for continuous phenotypes. Note: "median" is mapped internally to "quantile" (2-group quantile split). Default: "kmeans".

cutoff

Numeric vector of length n_group - 1. Required only when discretize_method = "custom". Defines interior breakpoints on the normalized, log2-transformed scale (i.e., after scale(log2(x + 1))). Must be sorted in ascending order.

label_type

Character specifying phenotype label type (e.g., "SBS1", "time"), stored in scRNA_data@misc

marker_finder

A character, the marker finder method. The default value is "limma".

log2FC

In the DESeq differential expression analysis results, the cutoff value of log2FC. The default value is 1L.

p_adjust

In the DESeq differential expression analysis results, the cutoff value of adjust P. The default value is 0.05.

show_log2FC

Select whether to show log2 fold changes. The default value is TRUE.

freq_counts

An integer, keep genes expressed in more than a certain number of cells. The default value is NULL, which means no filtering.

normalize

Select whether to perform normalization of count data. The default value is TRUE.

scale

Select whether to scale and center features in the dataset. The default value is TRUE.

nPerm

An integer, number of permutations to do. The default value is 1000L.

distance

A character, the distance algorithm must be included in "cosine", "pearson", "spearman", "kendall","euclidean","maximum". default value is NULL, which means "cosine".

...

Additional arguments to be passed to PIPET.optimized.

seed: Random seed for reproducibility
verbose: Whether to show progress messages
parallel: Whether to use parallel processing, default is FALSE. future::plan() must be set before calling this function.
assay: The assay to use, default is "RNA"

Usage

Arguments

See also