ArchRWrappers

Utility wrappers that smooth common ArchR single-cell ATAC-seq workflows, including reduced-dimension management and group-level summaries to Seurat bridge integration and plotting helpers.

Feature Highlights

• Dimensionality helpers: addReducedMNN, addAnyReducedMtrx, and plotRedDim automate batch correction, custom embeddings, and multi-panel visualization.
• Cross-platform data export: ArchR2sce converts ArchR projects into SingleCellExperiment objects for Bioconductor pipelines.
• Matrix summaries: getMeanMtrx, getSumMtrx, and getNonZeroProp compute per-group statistics that feed marker discovery or downstream plotting.
• Visualization utilities: bubblePlot and my_theme quickly communicate marker trends with ComplexHeatmap-backed bubbles and publication styling.
• Integration bridge: addGeneBridgeIntegrationMatrix drives Seurat/Signac bridge workflows, saving predicted RNA labels and scores back to Arrow files.
• Convenience helpers: %&%, imessage, and other small utilities keep scripts concise and consistent.

Documentation built from pkgdown lives under docs/ (for example the reference index at docs/reference/index.html).

Installation

ArchRWrappers expects a working ArchR setup (Bioconductor ≥ 3.16, R ≥ 4.2) plus the packages used inside the helpers (ArchR, SingleCellExperiment, SummarizedExperiment, batchelor, Seurat, Signac, ComplexHeatmap, etc.). Install the Bioconductor stack first, then pull ArchRWrappers from GitHub:

# Install base dependencies once
install.packages(c("remotes", "BiocManager"))
BiocManager::install(c("ArchR", "batchelor", "SingleCellExperiment", "SummarizedExperiment"))

# Install the wrapper package
remotes::install_github("Zepeng-Mu/ArchRWrappers")

If you build from source, enable Arrow writing support by installing hdf5/curl system libraries as required by ArchR.

Getting Started

1. Refine reduced dimensions and visualize

# Batch-correct an existing embedding with batchelor and store it on the project
proj <- addReducedMNN(
	ArchRProj = proj,
	reducedDims = "IterativeLSI",
	name = "MNN",
	groupBy = "Sample"
)

# Compare embeddings side-by-side with consistent aesthetics
plotRedDim(
	ArchRProj = proj,
	reducedDims = c("IterativeLSI", "MNN"),
	colorBy = c("Sample", "Clusters"),
	savePlot = FALSE
)

2. Convert to SingleCellExperiment for Bioconductor tools

sce <- ArchR2sce(
	ArchRProj = proj,
	featureSlot = "GeneScoreMatrix",
	reducedDims = c("UMAP", "MNN"),
	metadataFields = c("Sample", "Clusters")
)

3. Summarize markers with bubble plots

markerGenes <- c("GATA1", "SPI1", "PAX5", "MS4A1")
meanMat <- getMeanMtrx(
	ArchRProj = proj,
	useMatrix = "GeneScoreMatrix",
	groupBy = "Clusters",
	features = markerGenes
)
propMat <- getNonZeroProp(
	ArchRProj = proj,
	useMatrix = "GeneScoreMatrix",
	groupBy = "Clusters",
	features = markerGenes
)

bubblePlot(
	meanMtx = meanMat,
	propMtx = propMat,
	name = "GeneScoreMarkers",
	rowOrder = markerGenes,
	columnOrder = colnames(meanMat)
)

4. Bridge to RNA references with Seurat

proj <- addGeneBridgeIntegrationMatrix(
	ArchRProj = proj,
	matrixName = "GeneScoreMatrix",
	reducedDims = "IterativeLSI",
	seRNA = reference_rna_object,
	bridgeRNA = bridge_multiome,
	groupBy = "predictedGroup"
)

The call stores predicted labels (predictedCell, predictedGroup) and scores back into the ArchR project, enabling follow-up plotting or filtering without leaving ArchR.

Contributing & Support

• Issues and pull requests are welcome on GitHub.
• See LICENSE.md for the MIT license terms.
• For ArchR-specific questions, check the ArchR forum and the package documentation under docs/.