perform.seurat.cca: Performs Seurat Integration

In connorhknight/IBRAP: Integrated Benchmarking scRNA-seq Analytical Pipeline (IBRAP)

Performs Seurat Integration

Description

Performs Seurat integration on the supplied assay names. Results are saved under integration_reductions

Usage

perform.seurat.integration(
  object,
  object.list,
  assay,
  reduction.save.suffix = NULL,
  nfeatures = 2000,
  anchors.dims = 1:30,
  l2.norm = T,
  k.anchor = 5,
  k.filter = 200,
  k.score = 30,
  max.features = 200,
  nn.method = "annoy",
  n.trees = 50,
  anchor.eps = 0,
  features = NULL,
  integrate.dims = 1:30,
  k.weight = 100,
  sd.weight = 1,
  sample.tree = NULL,
  integrate.eps = 0,
  print.variance = TRUE,
  verbose = FALSE,
  seed = 1234,
  ...
)

Arguments

object	IBRAP S4 class object
object.list	list of individual sample IBRAP S4 class objects.
assay	Character. String containing indicating which assay to use
nfeatures	Numerical. How many features should be found as integration anchors. Default = 3000
l2.norm	Logical. Perform L2 normalization on the CCA cell embeddings after dimensional reduction. Default = TRUE
k.anchor	Numerical. How many neighbors (k) to use when picking anchors. Default = 5
k.filter	Numerical. How many neighbors (k) to use when filtering anchors. Default = 200
k.score	Numerical. How many neighbors (k) to use when scoring anchors. Default = 30
nn.method	Character. Method for nearest neighbor finding. Options include: rann, annoy. Default = annoy
anchor.eps	Numerical. Error bound on the neighbor finding algorithm (from RANN/Annoy) when finding integration genes.
features	Character. Vector of features to use when computing the PCA to determine the weights. Only set if you want a different set from those used in the anchor finding process. Default = NULL
integrate.dims	Numerical. Number of dimensions to use in the anchor weighting procedure. Default = 1:30
k.weight	Numerical. Number of neighbors to consider when weighting anchors. Default = 100
sd.weight	Numerical. Controls the bandwidth of the Gaussian kernel for weighting. Default = 1
sample.tree	Character. Specify the order of integration. If NULL, will compute automatically. Default = NULL
integrate.eps	Numerical. Error bound on the neighbor finding algorithm (from RANN)
reduction.save.suffix.	Character. What should be added as a suffix to reduction name. Default = ”
max.features.	Numerical. The maximum number of features to use when specifying the neighborhood search space in the anchor filtering. Default = 200
n.tree	Numerical. More trees gives higher precision when using annoy approximate nearest neighbor search. Default = 50
save.plot	Boolean. Should the automatically genewrated plot be saved? Default = TRUE

Value

Produces a new 'methods' assay containing normalised, scaled and HVGs.

Examples

perform.seurat.cca <- function(object = object, 
                                         assay = c('SCT', 'SCRAN', 'SCANPY'), 
                                         reduction.save.suffix=NULL,
                                         nfeatures = 3000,
                                         reduction = 'cca',
                                         anchors.dims = 1:30,
                                         l2.norm = T,
                                         k.anchor = 5,
                                         k.filter = 200,
                                         k.score = 30, 
                                         max.features = 200, 
                                         nn.method = 'annoy', 
                                         n.trees = 50, 
                                         anchor.eps = 0, 
                                         features = NULL,
                                         integrate.dims = 1:30,
                                         k.weight = 100,
                                         sd.weight = 1, 
                                         sample.tree = NULL, 
                                         integrate.eps = 0)

connorhknight/IBRAP documentation built on March 9, 2023, 7:01 p.m.