Load a Seurat object from an h5Seurat file

Load a previously saved Seurat object from an h5Seurat file. This function supports flexible loading options, allowing you to load only the components you need (e.g., specific assays, reductions) to minimize memory usage on large datasets.

Usage

LoadH5Seurat(file, ...)

# S3 method for class 'character'
LoadH5Seurat(
  file,
  assays = NULL,
  reductions = NULL,
  graphs = NULL,
  neighbors = NULL,
  images = NULL,
  meta.data = TRUE,
  commands = TRUE,
  misc = is.null(x = assays),
  tools = is.null(x = assays),
  verbose = TRUE,
  ...
)

# S3 method for class 'H5File'
LoadH5Seurat(
  file,
  assays = NULL,
  reductions = NULL,
  graphs = NULL,
  neighbors = NULL,
  images = NULL,
  meta.data = TRUE,
  commands = TRUE,
  misc = is.null(x = assays),
  tools = is.null(x = assays),
  verbose = TRUE,
  ...
)

# S3 method for class 'h5Seurat'
LoadH5Seurat(
  file,
  assays = NULL,
  reductions = NULL,
  graphs = NULL,
  neighbors = NULL,
  images = NULL,
  meta.data = TRUE,
  commands = TRUE,
  misc = is.null(x = assays),
  tools = is.null(x = assays),
  verbose = TRUE,
  ...
)

# S3 method for class 'h5Seurat'
as.Seurat(
  x,
  assays = NULL,
  reductions = NULL,
  graphs = NULL,
  neighbors = NULL,
  images = NULL,
  meta.data = TRUE,
  commands = TRUE,
  misc = TRUE,
  tools = TRUE,
  verbose = TRUE,
  ...
)

Arguments

file, x

Name of an h5Seurat file path (character) or connected h5Seurat file to load

...

Arguments passed to other methods

assays

One of:

• NULL (default): Load all assays

• A character vector with names of assays to load (e.g., c("RNA", "ADT"))

• A character vector specifying which data layers to load for all assays: c("counts", "data") loads only the counts and data layers, skipping scale.data

• A named list for fine-grained control, e.g., list(RNA = "data", ADT = c("data", "scale.data"))

reductions

One of:

• NULL (default): Load all reductions (PCA, UMAP, etc.)

• A character vector with names of specific reductions (e.g., c("pca", "umap"))

• NA: Load only global (assay-independent) reductions

• FALSE: Skip loading all reductions

Note: Only reductions associated with a loaded assay or marked as global will be loaded.

graphs

One of:

• NULL (default): Load all graphs

• A character vector with specific graph names (e.g., c("RNA_snn", "ADT_snn"))

• FALSE: Skip loading graphs

Note: Only graphs associated with loaded assays will be available.

neighbors

One of:

• NULL (default): Load all neighbor information

• A character vector with neighbor names

• FALSE: Skip neighbors

images

One of:

• NULL (default): Load all images (for spatial experiments)

• A character vector with image names

• NA: Load only global images

• FALSE: Skip images

meta.data

Logical; if TRUE (default), load cell metadata

commands

Logical; if TRUE (default), load command history. Commands are only loaded if their associated assays are loaded.

misc

Logical; if TRUE (default when all assays loaded), load miscellaneous data

tools

Logical; if TRUE (default when all assays loaded), load tool-specific information

verbose

Logical; if TRUE (default), show progress messages

Value

A Seurat object containing the requested components

Details

The h5Seurat format is highly flexible for selective loading. This is particularly useful when:

• Working with very large datasets where loading everything would exceed memory

• You only need specific assays or reductions for downstream analysis

• You want to quickly inspect object structure without full data loading

Seurat V5 Layer Support

For Seurat V5 objects with multiple layers, you can selectively load layers per assay. For example, use assays = list(RNA = "data") to load only the normalized expression layer, skipping raw counts and scaled data.

See also

SaveH5Seurat to save a Seurat object to h5Seurat format Convert to convert to other formats

Examples

if (FALSE) { # \dontrun{
library(SeuratDisk)

# Load entire h5Seurat file
seurat_obj <- LoadH5Seurat("data.h5seurat")

# Load only specific assays
seurat_obj <- LoadH5Seurat("data.h5seurat", assays = c("RNA", "ADT"))

# Load only specific data layers (memory-efficient for large files)
seurat_obj <- LoadH5Seurat("data.h5seurat", assays = c("data"))  # Only normalized expression

# Load specific assays with different layers
seurat_obj <- LoadH5Seurat(
  "data.h5seurat",
  assays = list(RNA = c("data", "scale.data"), ADT = "data")
)

# Load without reductions (faster)
seurat_obj <- LoadH5Seurat("data.h5seurat", reductions = FALSE)

# Load UMAP and PCA reductions only
seurat_obj <- LoadH5Seurat("data.h5seurat", reductions = c("umap", "pca"))

# Load spatial data without graphs (for Visium experiments)
seurat_obj <- LoadH5Seurat("visium.h5seurat", images = TRUE, graphs = FALSE)
} # }