Merge Multiple Screening Analysis Results

Combines results from multiple single-cell screening analyses (e.g., Scissor, scPAS, scPP, scAB) by merging their metadata, slots, and miscellaneous information while preserving the original expression data from the first input. Performs an inner join on cell barcodes to ensure only cells present in all inputs are retained.

Usage

MergeResult(
  ...,
  weights = NULL,
  ties.method = c("random", "first", "last"),
  verbose = SigBridgeRUtils::getFuncOption("verbose")
)

Arguments

...

Input objects to merge. Can be:

• Seurat objects

• Lists containing scRNA_data (Seurat objects)

• Mixed combinations of the above

The first object serves as the base for merging. When duplicate metadata columns are found, priority is given to the first occurrence.

weights

Named numeric vector specifying weights for voting when multiple screening methods produce categorical labels. Names must match screening method names (case-insensitive). If NULL (default), no voting is performed.

ties.method

Character. Method for resolving ties in weighted voting:

"first"

Select the first category alphabetically (default)

"last"

Select the last category alphabetically

"random"

Randomly select among tied categories

verbose

Logical. Whether to print progress messages. Default: inherits from getOption("SigBridgeR.verbose").

Value

A merged Seurat object containing:

• Expression data from the first input object

• Combined metadata from all input objects (inner join on cell IDs)

• Merged slots (assays, reductions, graphs, images) from all objects

• Combined miscellaneous information from all objects

• Optional vote column in metadata if weight is provided

Processing Workflow

1. Input Validation: Extracts Seurat objects from inputs (skips invalid objects with warning)

2. Metadata Extraction: Converts metadata to data.table format for efficient merging

3. Cell Intersection: Retains only cells present in all datasets (inner join)

4. Weighted Voting: If weight provided, performs weighted voting across screening methods

5. Slot Merging: Combines assays, reductions, graphs, and images from all objects

6. Misc Merging: Aggregates miscellaneous information from all objects

Weighted Voting

When weight is provided, the function:

1. Identifies columns in metadata matching registered screening method names

2. Extracts vote data (must contain "Positive", "Negative", "Neutral", or "Other")

3. Applies WeightedVote with specified weights

4. Adds vote column to merged metadata with final aggregated labels

Notes

• Only cells present in all input objects are retained (inner join behavior)

• Duplicate metadata columns are resolved by keeping the first occurrence

• For integrating heterogeneous single-cell datasets (different samples), use SCIntegrate instead

• Weights should be non-negative; higher values indicate greater influence in voting

See also

WeightedVote, SCIntegrate, AddMisc

Examples

if (FALSE) { # \dontrun{
# Example 1: Merge results from different screening methods
scissor_out <- RunScissor(...)
scAB_out <- RunscAB(...)
scPP_out <- RunscPP(...)

merged <- MergeResult(scissor_out, scAB_out, scPP_out)

# Check merged metadata
head(merged[[]])

# Example 2: Weighted voting across methods
weights <- c(
  Scissor = 3,    # Highest confidence
  scAB = 2,       # Medium confidence
  scPP = 1        # Lower confidence
)

merged <- MergeResult(
  scissor_out,
  scAB_out,
  scPP_out,
  weight = weights,
  ties.method = "first"
)

# View voting results
table(merged$vote)

# Example 3: Merge list-containing objects
result_list1 <- list(scRNA_data = seurat_obj1, other_info = "...")
result_list2 <- list(scRNA_data = seurat_obj2, other_info = "...")

merged <- MergeResult(result_list1, result_list2)

# Example 4: Check which cells were retained
original_cells <- ncol(seurat_obj1)
merged_cells <- ncol(merged)
cat("Retained", merged_cells, "of", original_cells, "cells\n")

# Example 5: Access merged slots
# All unique assays from all inputs
names(merged@assays)

# All unique reductions
names(merged@reductions)
} # }