Build a k-nearest neighbour graph — buildGraph • miloR

This function is borrowed from the old buildKNNGraph function in scran. Instead of returning an igraph object it populates the graph and distance slots in a Milo object. If the input is a SingleCellExperiment object or a matrix then it will return a de novo Milo object with the same slots filled.

buildGraph(
  x,
  k = 10,
  d = 50,
  transposed = FALSE,
  get.distance = FALSE,
  reduced.dim = "PCA",
  BNPARAM = KmknnParam(),
  BSPARAM = bsparam(),
  BPPARAM = SerialParam()
)

Arguments

x	A matrix, `SingleCellExperiment` or Milo object containing feature X cell gene expression data.
k	An integer scalar that specifies the number of nearest-neighbours to consider for the graph building.
d	The number of dimensions to use if the input is a matrix of cells X reduced dimensions. If this is provided, transposed should also be set=TRUE.
transposed	Logical if the input x is transposed with rows as cells.
get.distance	A logical scalar whether to compute distances during graph construction.
reduced.dim	A character scalar that refers to a specific entry in the `reduceDim` slot of the `Milo` object.
BNPARAM	refer to `buildKNNGraph` for details.
BSPARAM	refer to `buildKNNGraph` for details.
BPPARAM	refer to `buildKNNGraph` for details.

Value

A Milo object with the graph and distance slots populated.

Details

This function computes a k-nearest neighbour graph. Each graph vertex is a single-cell connected by the edges between its neighbours. Whilst a kNN-graph is strictly directed, we remove directionality by forcing all edge weights to 1; this behaviour can be overriden by providing directed=TRUE.

If you wish to use an alternative graph structure, such as a shared-NN graph I recommend you construct this separately and add to the relevant slot in the Milo object.

Examples

library(SingleCellExperiment)
ux <- matrix(rpois(12000, 5), ncol=200)
vx <- log2(ux + 1)
pca <- prcomp(t(vx))

sce <- SingleCellExperiment(assays=list(counts=ux, logcounts=vx),
                            reducedDims=SimpleList(PCA=pca$x))

milo <- Milo(sce)
milo <- buildGraph(milo, d=30, transposed=TRUE)
#> Constructing kNN graph with k:10

milo
#> class: Milo 
#> dim: 60 200 
#> metadata(0):
#> assays(2): counts logcounts
#> rownames: NULL
#> rowData names(0):
#> colnames: NULL
#> colData names(0):
#> reducedDimNames(1): PCA
#> spikeNames(0):
#> altExpNames(0):
#> nhoods dimensions(1): 0
#> nhoodCounts dimensions(2): 1 1
#> nhoodDistances dimension(1): 0
#> graph names(1): graph
#> nhoodIndex names(1): 0
#> nhoodExpression dimension(2): 1 1
#> nhoodReducedDim names(0):
#> nhoodGraph names(0):
#> nhoodAdjacency dimension(0):