Clustexp function

Clustering of single-cell transcriptome data

This functions performs the initial clustering of the RaceID algorithm. methods

Clustexp(
  object,
  clustnr = 3,
  bootnr = 50,
  metric = "pearson",
  do.gap = TRUE,
  SE.method = "Tibs2001SEmax",
  SE.factor = 0.25,
  B.gap = 50,
  cln = 0,
  rseed = NULL,
  quiet = FALSE
)

## S4 method for signature 'DISCBIO'
Clustexp(
  object,
  clustnr = 3,
  bootnr = 50,
  metric = "pearson",
  do.gap = TRUE,
  SE.method = "Tibs2001SEmax",
  SE.factor = 0.25,
  B.gap = 50,
  cln = 0,
  rseed = NULL,
  quiet = FALSE
)

Arguments

• object: DISCBIO class object.
• clustnr: Maximum number of clusters for the derivation of the cluster number by the saturation of mean within-cluster-dispersion. Default is 20.
• bootnr: A numeric value of booststrapping runs for clusterboot. Default is 50.
• metric: Is the method to transform the input data to a distance object. Metric has to be one of the following: ["spearman", "pearson", "kendall", "euclidean", "maximum", "manhattan", "canberra", "binary", "minkowski"].
• do.gap: A logical vector that allows generating the number of clusters based on the gap statistics. Default is TRUE.
• SE.method: The SE.method determines the first local maximum of the gap statistics. The SE.method has to be one of the following:["firstSEmax", "Tibs2001SEmax", "globalSEmax", "firstmax", "globalmax"]. Default is "Tibs2001SEmax"
• SE.factor: A numeric value of the fraction of the standard deviation by which the local maximum is required to differ from the neighboring points it is compared to. Default is 0.25.
• B.gap: Number of bootstrap runs for the calculation of the gap statistics. Default is 50
• cln: Number of clusters to be used. Default is NULL and the cluster number is inferred by the saturation criterion.
• rseed: Random integer to enforce reproducible clustering results.
• quiet: if TRUE, intermediate output is suppressed

Returns

The DISCBIO-class object input with the cpart slot filled.

Examples

sc <- DISCBIO(valuesG1msTest) # changes signature of data
sc <- Clustexp(sc, cln = 2)