BuildTopicSNN: Build SNN graph

In ludvigla/spaceST: spaceST

Description

This function is a modified version of the BuildSNN function in the single cell genomics tool Seurat (Version: 2.1.0, date: 2017-10-11).

Usage

BuildTopicSNN(object, k.param = 30, k.scale = 25, plot.SNN = FALSE,
  prune.SNN = 1/15, print.output = TRUE, clusters = NULL, seed = 0,
  cluster.method = "SNN", perplexity = 15, color.by.sample = T,
  tsne = NULL, layout = NULL, algorithm = 1, ...)

Arguments

object	Topic matrix obtained with compute.lda() from cellTree package
k.param	Defines k for the k-nearest neighbor algorithm
k.scale	Granularity option for k.param
plot.SNN	Plot the SNN graph
prune.SNN	Sets the cutoff for acceptable Jaccard distances when computing the neighborhood overlap for the SNN construction. Any edges with values less than or equal to this will be set to 0 and removed from the SNN graph. Essentially sets the strigency of pruning (0 — no pruning, 1 — prune everything).
print.output	Whether or not to print output to the console
clusters	Clusters used to color vertices in SNN ntwork graph
seed	Set random seed for t-SNE algorithm
cluster.method	Chose method to use for clustering. "topic" will run clusterST() and cluster features based on topic porportion using hierarchical clustering with euclidean distances and ward.D2. Clusters are chosen using the unsupervised cutreeDynamic function from dynamicTreeCut package. "SNN" will run a network graph based algorithm using methods from the Seurat package.
perplexity	Set perplexity for tsne if plot.SNN is TRUE
color.by.sample	Color vertices by sample in network graph
tsne	Two column matrix with tsne results, alternatively run tsne using preset parameters
layout	Set custom layout of vertices in network graph. Will override t-SNE results
algorithm	Algorithm for modularity optimization (1 = original Louvain algorithm; 2 = Louvain algorithm with multilevel refinement; 3 = SLM algorithm)
...	Parameters passed to RunModularityClusteringspaceST() function used for clustering

Value

SNN matrix

ludvigla/spaceST documentation built on May 29, 2019, 3:43 a.m.