plotGainingH: Gaining plots for using h-statistic instead original omic...

In fjcamlab/deco: Decomposing Heterogeneous Cohorts using Omic Data Profiling

Gaining plots for using h-statistic instead original omic data.

Description

This function returns three plots for a particular feature(s), showing the correlation between the h-statistic calculated by DECO and the original omic data.

Usage

plotGainingH(deco, data, ids, print.annot = FALSE, 
             orig.classes = TRUE)

Arguments

deco	a "deco" R object generated by 'decoNSCA' R function.
data	a matrix of normalized omic data with 'f' features (rows) by 'n' samples (columns).
ids	character vector indicating what original IDs from input data should be highlighted and plot in 'Profile' section. If all IDs found have to be plot, input all corresponding IDs here.
print.annot	logical indicating if annotation (gene SYMBOL) should be printed instead original IDs. This option needs previous annotation within 'decoRDA' function or late annotation binding information to @featureTable slot of 'deco' R object.
orig.classes	logical indicating if original sample categories compared (orig.classes = TRUE) or DECO subclasses (orig.classes = FALSE) will be represented.

Details

Diagnostic plot associating omic profile and h-statistic profile of any significant feature found by DECO.

Value

Three plots are returned: Boxplot: distribution of omic data and h-statistic values per DECO subclass. Top-left: parametric correlation between omic data and h-statistic per sample. Top-right: non-parametric correlation (ranking) between omic data and h-statistic per sample.

Author(s)

Francisco Jose Campos Laborie. <[e-mail]>

See Also

decoRDA, decoNSCA

Examples

#### Further information in DECO vignette(s).
## Run after decoRDA and decoNSCA steps on 'ALCLdata'
data(ALCLdata)
ALCL

## Feature to represent
id <- featureTable(deco.results.ma)[1, "ID"]

#### Comparing DECO subclasses against source of samples.
plotGainingH(deco.results.ma, data = assay(ALCL), ids = id, 
             print.annot = FALSE, orig.classes = FALSE)

fjcamlab/deco documentation built on Aug. 18, 2024, 4:28 a.m.