In-silico Knock-Out / Knock-In Analysis (single gene)

Performs a virtual knock-out or knock-in simulation for a target gene in a single-cell dataset. The function identifies cells with high or low expression of the gene, matches group sizes and cell-type composition, and performs differential expression and enrichment analysis to infer potential functional effects.

Usage

silico_ko(
  all,
  gene,
  sko_mode = c("ko", "ki"),
  cell_col = "cell_anno",
  filter_cell_threshold = 10,
  pct_threshold = 0.1,
  abs_threshold = NULL,
  deg_method = "default",
  enrichment_method = c("ORA", "GSEA"),
  enrichment_bg = c("GO", "KEGG", "MKEGG", "Reactome"),
  simplify = TRUE
)

Arguments

all: A Seurat object containing expression data and cell metadata.
gene: Character(1). The target gene symbol (must exist in the expression matrix).
sko_mode: One of c("ko", "ki"). Specifies virtual knock-out ("ko") or knock-in ("ki") mode.
cell_col: Character(1). The metadata column indicating cell type.
filter_cell_threshold: Integer. Minimum number of cells per cell type to retain.
pct_threshold: Numeric. Fraction of expressing cells to extract per group (ignored if abs_threshold is set).
abs_threshold: Integer. Absolute number of cells per group to extract; overrides pct_threshold if provided.
deg_method: Character. Differential expression method: "default" or "MAST".
enrichment_method: Character vector. Enrichment methods to run; default is c("ORA", "GSEA").
enrichment_bg: Character vector. Background databases for enrichment; default is c("GO", "KEGG", "MKEGG", "Reactome").
simplify: Logical. Whether to simplify redundant enrichment terms via clusterProfiler::simplify. Defaults to TRUE. Set to FALSE if enrichment results are sparse (e.g., small datasets) to avoid errors.

Value

Invisibly returns a list containing:

high_cells: Vector of selected high-expression cells.
low_cells: Vector of selected low-expression cells.
cell_rato: Data frame summarizing cell-type composition.
deg_results: Differential expression results (data frame).
enrichments: Enrichment analysis results.

Details

The pipeline includes:

Filtering cells expressing the target gene.
Retaining cell types with sufficient cell counts.
Selecting top-expressing cells (high group) and matched low-expressing cells (low group).
Running differential expression between the two groups using Seurat::FindMarkers.
Performing enrichment analysis via leo.basic::leo_enrich.

Examples

if (FALSE) { # \dontrun{
res <- silico_ko(
  all = srt, gene = "LRRK2",
  sko_mode = "ko", cell_col = "cell_anno",
  pct_threshold = 0.1, deg_method = "default"
)
} # }