Skip to contents

Integrate a merged Seurat object across batches.

integrate_seurat_object.Rd

The integrate step of the pipeline in one call. It plays the same role for a merged object that create_seurat_object() plays for a single sample.

Usage

integrate_seurat_object(
  x,
  assay = "RNA",
  int_reduction = "cca",
  num_dim,
  batch_var = "orig.ident",
  num_variable_genes = 3000,
  num_neighbors = 30,
  k_anchor = 10,
  k_weight = 100,
  log_file = NULL,
  ...
)

Arguments

x

Merged Seurat object, such as the output of merge_seurat_objects().

assay

Assay to integrate. Passed straight through to integrate_layers().

int_reduction

Integration method: "cca", "rpca", or "harmony".

num_dim

Number of dimensions to use (5-50).

batch_var

Metadata column identifying the batch.

num_variable_genes

Number of variable features per batch.

num_neighbors

Neighbors for UMAP.

k_anchor

Neighbors to use when picking anchors. Passed straight through to integrate_layers() - see there for defaults/caveats (ignored by "harmony").

k_weight

Neighbors to use when weighting the corrections. Passed straight through to integrate_layers() - see there for defaults/caveats (also ignored by "harmony").

log_file

Filename for the log file.

...

Passed straight through to integrate_layers() - extra method-specific tuning arguments (harmony's theta/lambda/sigma/..., or cca/rpca's k.filter/sample.tree/...).

Value

A Seurat object with rejoined layers, an int_reduction-named reduction ("cca", "rpca", or "harmony"), and "tsne"/"umap" reductions computed from it. Also writes the variable-gene overlap, QC distribution, integrated-embedding, and tSNE/UMAP plots to the working directory.

Details

The function runs these steps:

  1. Split the assay layers by batch and prepare them (split_layers_by_batch()).

  2. Report how much the per-batch variable genes overlap (plot_var_genes_euler(), plot_var_genes_upset()).

  3. Correct the batch effect (integrate_layers()).

  4. Compute a first-pass tSNE/UMAP on the corrected embedding.

The tSNE/UMAP from step 4 is a preview, capped at num_dim. It differs from the real one cluster_seurat_object() computes later over its own num_dim. cluster_seurat_object() has no default reduction. Pass it this function's int_reduction name explicitly. Otherwise it may use the uncorrected "pca" reduction that split_layers_by_batch() also leaves on the object.