aPEAR.methods {aPEAR} | R Documentation |
Default method configuration for aPEAR
Description
A list with parameters for customizing how the clusters within the enrichment data are calculated.
Usage
aPEAR.methods
Format
An object of class aPEAR.methods.config
of length 5.
Details
similarity: method for calculating similarity matrix between the pathways. Available methods: 'jaccard', 'cosine' and 'correlation'
cluster: method for detecting pathway clusters. Available methods: 'markov', 'hier'
and 'spectral'. Using 'spectral' method requires that you have the Spectrum
package installed
clusterName: method for selecting cluster names. Available methods: 'pagerank',
'hits', 'nes' and 'pval'. The 'pagerank' and 'hits' algorithms analyse the connectivity
within the cluster to detect the most important node. The 'nes' and 'pval' methods
use enrichment results to determine the most important node within the cluster: the 'nes'
method will choose the node with the maximum absolute enrichment score value and the
'pval' method will choose the node with the lowest p-value. When using the 'nes' and
'pval' methods, please specify which column in the data to use with the clusterNameColumn
parameter
clusterNameColumn: which column in the dataset should be used to select the cluster
title. Required when clusterName = 'nes'
and clusterName = 'pval'
minClusterSize: minimum cluster size (default: 2). Clusters with less elements than specified will be dropped
Value
an object of class aPEAR.methods.config
Examples
# Display all default methods used by aPEAR
aPEAR.methods
# Update methods to use different similarity metric
settings <- aPEAR.methods
settings$similarity <- 'cosine'
settings