cluster_links: cluster_links

In OmicInt: Omics Network Exploration

Description

Function to select an optimal number of clusters and a model to be fitted during the EM phase of clustering for Gaussian Mixture Models. The function provides summaries and helps to visualise gene clusters based on generated data using score_genes function. Weighed gene expression is clustered based on a specific disease score which can be either the association or specificity for a disease, i.e., if the gene has known links to disease phenotypes or how specific it is when describing a pathology. The function also provides scatter plots and dimension reduction plots to analyse the clusters and features in the experimental data.

Usage

cluster_links(
  data,
  max_range = 20,
  type = "association",
  clusters = NULL,
  modelNames = NULL
)

Arguments

data	data frame containing processed expression file from score_genes with LFCscore; Class - data frame
max_range	number of clusters to consider during model selection; default 20 clusters. Class - integer
type	type of score to consider which can be either "association" or "specificity"; default "association". Class - string
clusters	number of clusters to test not based on the best BIC output, user also needs to supply modelNames; class - integer
modelNames	can only be supplied when clusters are also specified, this option will model based on the user parameters; class- string

Value

A data frame object that contains a summary of clusters; class - data frame

Examples

 ## Not run: 
path_to_test_data<- system.file("extdata", "test_data.tabular", package="OmicInt")
# basic usage of cluster_links
df<-utils::read.table(path_to_test_data)
df<-cluster_links(df)
head(df)

## End(Not run)

OmicInt documentation built on Oct. 28, 2021, 5:09 p.m.