schard

name	schard
domain	compbio
description	R package for converting h5ad files (HDF5 AnnData format from scanpy) to Seurat or SingleCellExperiment objects. Use whenever working with single-cell data in h5ad format that needs to be converted to R. Includes support for Visium spatial transcriptomics data, raw counts, cell metadata, and dimensionality reduction embeddings.

schard - h5ad to R Converter

The schard R package converts Python scanpy h5ad files to Seurat or SingleCellExperiment objects for analysis in R.

Installation

# Install from GitHub
remotes::install_github("cellgeni/schard")

Downloading Public h5ad Files

From CZI cellxgene

download.file('https://datasets.cellxgene.cziscience.com/c5ac5c36-f60c-4680-8018-2d6cb65c0a37.h5ad', 'vis.heart.h5ad')
download.file('https://datasets.cellxgene.cziscience.com/8cc521c8-c4ff-4cba-a07b-cae67a9dcba9.h5ad', 'sn.heart.h5ad')

From Sanger Atlases

download.file('https://covid19.cog.sanger.ac.uk/baron16.processed.h5ad', 'ba16.h5ad')

Loading h5ad Files

As SingleCellExperiment

ba16.sce = schard::h5ad2sce('ba16.h5ad')

As Seurat Object

# Standard Seurat object
snhx = schard::h5ad2seurat('sn.heart.h5ad')

# Load raw counts instead of normalized data
snhr = schard::h5ad2seurat('sn.heart.h5ad', use.raw = TRUE)

Visium Spatial Data

# Load all Visium samples as single Seurat object
visx = schard::h5ad2seurat_spatial('vis.heart.h5ad')

# Load as list of Seurat objects (one per slide)
visl = schard::h5ad2seurat_spatial('vis.heart.h5ad', simplify = FALSE)

# Raw counts for Visium
visr = schard::h5ad2seurat_spatial('vis.heart.h5ad', use.raw = TRUE)

Working with Spatial Plots

# Plot total counts on tissue
Seurat::SpatialPlot(visx, features = 'total_counts')

# Plot specific sample
Seurat::SpatialPlot(visx, features = 'total_counts', images = 'HCAHeartST11702009')

# Plot from list (per-slide object)
Seurat::SpatialPlot(visl$HCAHeartST11702010, features = 'total_counts')

# Compare normalized vs raw counts
plot(colSums(visx), colSums(visr), pch = 16)
# Raw counts are different from normalized ones

Working with Dimensionality Reductions

# DimPlot works but specify reduction manually for safety
Seurat::DimPlot(snhx, group.by = 'cell_state')
# Note: reduction name is 'Xumap_' (auto-translated from scanpy to Seurat)
# Safer to specify: reduction = 'Xumap_'

Loading Cell Metadata Only

# Load obs (cell metadata)
obs = schard::h5ad2data.frame('sn.heart.h5ad', 'obs')

Loading Embeddings/UMAP

First, explore what's available in the h5ad file:

ls = rhdf5::h5ls('sn.heart.h5ad')
ls[ls$group == '/obsm',]  # Shows available embeddings like 'X_umap'

Then load the embedding:

# Load UMAP coordinates (transposed for standard orientation)
umap = t(schard::h5ad2Matrix('sn.heart.h5ad', '/obsm/X_umap'))

# Plot UMAP with cell states
plot(umap[, 1:2], pch = 16, cex = 0.4, col = factor(obs$cell_state))

Common H5ad Paths

Path	Description
/obs	Cell metadata (data.frame)
/var	Feature metadata (data.frame)
/X	Main expression matrix
/raw/X	Raw counts matrix
/obsm/X_umap	UMAP coordinates
/obsm/X_pca	PCA coordinates
/obsm/spatial	Spatial coordinates (Visium)
/uns	Unstructured annotations