outlyx: Identify Outliers within Methylumi and Minfi packaged objects

In schalkwyk/wateRmelon: Illumina 450 and EPIC methylation array normalization and metrics

Identify Outliers within Methylumi and Minfi packaged objects

Description

Seeks to identify outliers based on multiple (currently 2) outlier detection methods for methylumi and minfi packaged objects.

Usage

outlyx(x, iqr=TRUE, iqrP=2, pc=1,
       mv=TRUE, mvP=0.15, plot=TRUE, ...) 

Arguments

x	A MethyLumiSet, MethylSet, RGChannelSet object or matrix containing raw betas.
iqr	If TRUE, the outliers based on interquartile ranges will be determined
iqrP	The number of interquartile ranges outliers are to be identified from designated principle component.
pc	Desired principal component for outlier identification - only used if other principal components want to be discriminated, only used for IQR outlier detection.
mv	If TRUE, the outliers will detected using pcout
mvP	Arbitrary cut-off point for identifying outliers via pcout
plot	If TRUE, alongside regular output, a plot will be constructed displaying relative 'location' of each sample. Outliers are those that fall within the highlighted regions.
...	Additional arguments passed to pcout

Value

Returns a dataframe of TRUE/FALSE per sample where TRUE is outlying. Dataframe contains 3 columns, the first column (iqr) denotes samples which are outlying according to IQR on Principal component 1, the second column (mv) denotes outliers according to mahalanobis distances. And the third column (outliers) denotes samples that are TRUE in the first two columns.

Note

May perform poorly on normalized data

Author(s)

Tyler Gorrie-Stone - [e-mail]

Examples

 library(wateRmelon)
 data(melon)
 outliers <- outlyx(melon,iqr=TRUE, iqrP=2, pc=1,
                    mv=TRUE, mvP=0.15, plot=TRUE)

schalkwyk/wateRmelon documentation built on April 15, 2024, 12:06 p.m.