FindClusters: Cluster Determination
In paodan/studySeu: Seurat : R toolkit for single cell genomics

Description Usage Arguments Value

Identify clusters of cells by a shared nearest neighbor (SNN) quasi-clique based clustering algorithm. First calculate k-nearest neighbors and construct the SNN graph. Then determine the quasi-cliques associated with each cell. Finally, merge the quasi-cliques into clusters. For a full description of the algorithm, see Xu and Su (2015) Bioinformatics.

FindClusters(object, genes.use = NULL, pc.use = NULL, k.param = 10,
  k.scale = 10, plot.SNN = FALSE, prune.SNN = 0.1, save.SNN = FALSE,
  r.param = 0.7, m.param = NULL, q = 0.1, qup = 0.1, update = 0.25,
  min.cluster.size = 1, do.sparse = FALSE, do.modularity = TRUE,
  modularity = 1, resolution = 0.8, algorithm = 1, n.start = 100,
  n.iter = 10, random.seed = 0, print.output = 1)

`object`	Seurat object
`genes.use`	Gene expression data
`pc.use`	Which PCs to use for construction of the SNN graph
`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`	Stringency of pruning for the SNN graph (0 - no pruning, 1 - prune everything)
`save.SNN`	Whether to return the SNN matrix or not. If true, returns a list with the object as the first item and the SNN matrix as the second item.
`r.param`	r defines the connectivity for the quasi-cliques. Higher r gives a more compact subgraph
`m.param`	m is the threshold for merging two quasi-cliques. Higher m results in less merging
`q`	Defines the percentage of quasi-cliques to examine for merging each iteration
`qup`	Determines how to change q once all possible merges have been made
`update`	Adjust how verbose the output is
`min.cluster.size`	Smallest allowed size for a cluster
`do.sparse`	Option to store and use SNN matrix as a sparse matrix. May be necessary datasets containing a large number of cells.
`do.modularity`	Option to use modularity optimization for single cell clustering.
`modularity`	Modularity function (1 = standard; 2 = alternative).
`resolution`	Value of the resolution parameter, use a value above (below) 1.0 if you want to obtain a larger (smaller) number of communities.
`algorithm`	Algorithm for modularity optimization (1 = original Louvain algorithm; 2 = Louvain algorithm with multilevel refinement; 3 = SLM algorithm).
`n.start`	Number of random starts.
`n.iter`	Maximal number of iterations per random start.
`random.seed`	Seed of the random number generator.
`print.output`	Whether or not to print output to the console (0 = no; 1 = yes).