makePCA: 3D Plot with Principal Component Analysis (PCA)

In machalen/QualityGraphs: Make quality graphs for microarray data based on objects obtained witharoma affymetrix normalization.

Description

This function performs a principal component analysis with the prcomp function and uses the prcomp class object to plot the three first principal components.

Usage

makePCA(est_noctrls, picname, conditions = NULL, colors = NULL, dist = 2, resDir = NULL)

Arguments

est_noctrls	Numeric matrix where columns are samples and rows genes. est_noctrls corresponds to a matrix without controls.
picname	Character with the output file name.
conditions	Optional. Vector with sample conditions to colour sample names in the plot. conditions vector has to have the same order as the order of samples in the columns of the input expression matrix. IMPORTANT Sample columns in the expression matrix and sample conditions have to be in order mixedsort(labels).
colors	Optional. Vector with colors corresponding to each sample in mixedsort(samplename) order. If this parameter is not defined, default colors will be used.
dist	Distance where the name of the sample is located with respect to the corresponing point in the plot. Optimal distance is different deppending on the sample number and PC distances. Default is 2.
resDir	Output results directory. Default is ResultsDir.

Details

The first three principal components obtained with the prcomp function capture part of the variability explaining the sample distribution. Each PC is represented in an axis of the plot, x.axis=PC1, y.axis=PC2 and z.axis=PC3. This plot is useful to check the overall similarity among samples. Column names in the expression matrix are used as sample names.

Value

Output is a .png file with the results in resDir.

machalen/QualityGraphs documentation built on Oct. 22, 2019, 8:29 p.m.