Cluster the given terms by the pairwise similarity.

kobas.term_cluster {KOBAS}

R Documentation

Cluster the given terms by the pairwise similarity.

Description

Usage

kobas.term_cluster(terms,
    threshold = 0.35,
    method = "kappa",
    universe = -1,
    pvalue = NULL);

Arguments

terms

a named list of the character vectors, see the terms parameter of kobas.term_similarity. [as list]

threshold

the similarity threshold of the term cluster, default 0.35. [as double]

method

the similarity method: kappa (default) or jaccard. [as string]

universe

the gene universe size of the kappa method, default -1 (auto). [as integer]

pvalue

an optional numeric vector of the p-values of the terms, it is used for scoring the term cluster (the geometric mean of -log10(p)). Default NULL means NA.

env

the R# runtime environment. [as Environment]

Details

The cluster self test of the KOBAS algorithm requires more than 3 pairwise similarities, that means a cluster should contain at least 4 terms.

Authors

gseakit

Value

a dataframe with the cluster index as the row names and the columns: size (the number of the terms in the cluster), score, members (the term ids joined by ;) and index (the term index in the input term list).

clr value class

dataframe

[Package KOBAS version 1.0.0.0 Index]