Run scOMM label prediction

Run scOMM on shared features between a reference object and a query object, write predicted labels and class scores into query metadata, and optionally evaluate predictions against a truth label.

Usage

RunscOMM(
  srt,
  reference,
  reference_assay = NULL,
  query_assay = NULL,
  reference_label = NULL,
  features = NULL,
  prediction_prefix = "predicted_",
  evaluate = FALSE,
  truth_col = NULL,
  tool_name = "scOMM",
  rare_threshold = 0.05,
  scomm_python = NULL,
  scomm_hidden_nodes = c(128, 64),
  scomm_epochs = 10,
  scomm_batch_size = 32,
  scomm_threshold = 0.5,
  scomm_seed = 11,
  verbose = TRUE
)

Arguments

srt

A Seurat object.

reference

Reference Seurat object used for supervision.

reference_assay

Assay used in the reference object.

query_assay

Assay used in the query object.

reference_label

Metadata column in the reference used as supervision labels.

features

Shared features passed to scOMM. If NULL, reference variable features are used.

prediction_prefix

Prefix added to prediction metadata columns.

evaluate

Whether to compute prediction metrics against a truth label.

truth_col

Metadata column in srt used as the truth label when evaluate = TRUE.

tool_name

Name used to store detailed results in srt@tools.

rare_threshold

Maximum class proportion used to define rare classes when calculating rare_recall.

scomm_python

Optional Python binary used by the scOMM backend. If NULL, SCOP_SCOMM_PYTHON is consulted and reticulate defaults are used otherwise.

scomm_hidden_nodes, scomm_epochs, scomm_batch_size, scomm_threshold, scomm_seed

Parameters passed to the scOMM backend.

verbose

Whether to print the message. Default is TRUE.

Value

A Seurat object with scOMM predictions stored in metadata and tools.