fastCNV calls all of the internal functions needed to compute the putative CNV on a Seurat object or a list of Seurat objects

This function orchestrates the CNV analysis on a Seurat object (or multiple objects). It calls internal functions such as prepareCountsForCNVAnalysis, CNVAnalysis, CNVPerChromosomeArm, CNVcluster, and PlotCNVResults to compute the CNVs, perform clustering, and generate heatmaps. The results are saved in the metadata of the Seurat object(s), with options for generating and saving plots.

Usage

fastCNV(
  seuratObj,
  sampleName,
  referenceVar = NULL,
  referenceLabel = NULL,
  assay = NULL,
  prepareCounts = "default",
  aggregFactor = 15000,
  clusterResolution = 0.8,
  aggregateByVar = TRUE,
  reCluster = FALSE,
  pooledReference = TRUE,
  scaleOnReferenceLabel = TRUE,
  thresholdPercentile = 0.01,
  geneMetadata = getGenes(),
  windowSize = 150,
  windowStep = 10,
  saveGenomicWindows = FALSE,
  topNGenes = 7000,
  chrArmsToForce = NULL,
  genesToForce = NULL,
  regionToForce = NULL,
  getCNVPerChromosomeArm = TRUE,
  getCNVClusters = TRUE,
  k_clusters = NULL,
  h_clusters = NULL,
  cellTypesToCluster = NULL,
  mergeCNV = TRUE,
  mergeThreshold = 0.98,
  doPlot = TRUE,
  denoise = TRUE,
  printPlot = FALSE,
  savePath = ".",
  outputType = "png",
  clustersVar = "cnv_clusters",
  splitPlotOnVar = clustersVar,
  referencePalette = "default",
  clusters_palette = "default"
)

Arguments

seuratObj: Seurat object or list of Seurat objects to perform the CNV analysis on.
sampleName: Name of the sample or a list of names corresponding to the samples in the seuratObj.
referenceVar: The variable name of the annotations in the Seurat metadata to be used as reference.
referenceLabel: The label given to the observations you want as reference (can be any type of annotation).
assay: Name of the assay to run the CNV on. Takes the results of prepareCountsForCNVAnalysis by default if available.
prepareCounts: One of TRUE, FALSE, or "default" (default). When "default", prepareCountsForCNVAnalysis is run only if the number of cells/spots is below 60,000.
aggregFactor: The number of counts per spot desired (default = 15 000). If less than 1,000, will not run the prepareCountsForCNVAnalysis function.
clusterResolution: The resolution wanted for the Seurat clusters (default = 0.8).
aggregateByVar: If referenceVar is given, determines whether to use it to pool the observations (default = TRUE).
reCluster: Whether to re-cluster if the Seurat object given already has a seurat_clusters slot in its metadata (default = FALSE).
pooledReference: Default is TRUE. Will build a pooled reference across all samples if TRUE.
scaleOnReferenceLabel: If TRUE, scales the results depending on the normal observations (default = TRUE).
thresholdPercentile: Which quantiles to take (default 0.01). For example, 0.01 will take quantiles between 0.01-0.99. Background noise appears with higher numbers.
geneMetadata: List of genes and their metadata (default uses genes from Ensembl version 113).
windowSize: Size of the genomic windows for CNV analysis (default = 150).
windowStep: Step between the genomic windows (default = 10).
saveGenomicWindows: If TRUE, saves the information of the genomic windows in the current directory (default = FALSE).
topNGenes: Number of top expressed genes to keep (default = 7000).
chrArmsToForce: A chromosome arm (e.g., "8p", "3q") or a list of chromosome arms (e.g., c("3q", "8p", "17p")) to force into the analysis.
genesToForce: A list of genes to force into the analysis (e.g. c("FOXP3","MUC16","SAMD15")).
regionToForce: Chromosome region to force into the analysis (vector containing chr, start, end).
getCNVPerChromosomeArm: If TRUE, will save the CNV per chromosome arm into the metadata.
getCNVClusters: If TRUE, will perform clustering on the CNV scores and save them in the metadata of the Seurat object as cnv_clusters.
k_clusters: Optional. Number of clusters to cut the dendrogram into. If NULL, the optimal number of clusters is determined automatically using the elbow method.
h_clusters: Optional. The height at which to cut the dendrogram for clustering. If both k and h are provided, k takes precedence.
cellTypesToCluster: Optional. Cell type annotations to use for clustering. If NULL (default), all observations will be clustered.
mergeCNV: Logical. Whether to merge the highly correlated CNV clusters.
mergeThreshold: A numeric value between 0 and 1. Clusters with correlation greater than this threshold will be merged. Default is 0.98.
doPlot: If TRUE, will build a heatmap for each of the samples (default = TRUE).
denoise: If TRUE, the denoised data will be used in the heatmap (default = TRUE).
printPlot: If TRUE, the heatmap will be printed in the console (default = FALSE, the plot will only be saved in a PDF).
savePath: Path to save the heatmap plot. If NULL, the plot won't be saved (default = .).
outputType: Specifies the file format for saving the plot, either "png" or "pdf" (default = "png").
clustersVar: The variable name of the clusters in the Seurat metadata (default = "cnv_clusters").
splitPlotOnVar: The name of the metadata column to split the observations during the plotCNVResults step, if different from referenceVar.
referencePalette: The color palette that should be used for referenceVar (default = "default").
clusters_palette: The color palette that should be used for clustersVar (default = "default").

Value

A list of Seurat objects after all the analysis is complete. Heatmaps of the CNVs for every object in seuratObj are generated and saved in the specified path (default = current working directory).