Description Usage Arguments Details Value Author(s) Examples
Principal Component Analysis (PCA) is a very powerful technique that has wide applicability in data science, bioinformatics, and further afield. It was initially developed to analyse large volumes of data in order to tease out the differences/relationships between the logical entities being analysed. It extracts the fundamental structure of the data without the need to build any model to represent it. This 'summary' of the data is arrived at through a process of reduction that can transform the large number of variables into a lesser number that are uncorrelated (i.e. the ‘principal components'), whilst at the same time being capable of easy interpretation on the original data. PCAtools provides functions for data exploration via PCA, and allows the user to generate publicationready figures. PCA is performed via BiocSingular  users can also identify optimal number of principal component via different metrics, such as elbow method and Horn's parallel analysis, which has relevance for data reduction in singlecell RNAseq (scRNAseq) and high dimensional mass cytometry data.
1 2 3 4 5 6 7 8 9 10 
mat 
A datamatrix or dataframe containing numerical data only. Variables are expected to be in the rows and samples in the columns by default. 
metadata 
A datamatrix or dataframe containing metadata. This will be stored in the resulting pca object. Strictly enforced that rownames(metadata) == colnames(mat). 
center 
Center the data before performing PCA? Same as prcomp() 'center' parameter. 
scale 
Scale the data? Same as prcomp() 'scale' parameter. 
rank 
An integer scalar specifying the number of PCs to retain. OPTIONAL for an exact SVD, whereby it defaults to all PCs. Otherwise REQUIRED for approximate SVD methods. 
removeVar 
Remove this % of variables based on low variance. 
transposed 
Is 
BSPARAM 
A BiocSingularParam object specifying the algorithm to use for the SVD. Defaults to an exact SVD. 
Principal Component Analysis (PCA) is a very powerful technique that has wide applicability in data science, bioinformatics, and further afield. It was initially developed to analyse large volumes of data in order to tease out the differences/relationships between the logical entities being analysed. It extracts the fundamental structure of the data without the need to build any model to represent it. This 'summary' of the data is arrived at through a process of reduction that can transform the large number of variables into a lesser number that are uncorrelated (i.e. the ‘principal components'), whilst at the same time being capable of easy interpretation on the original data. PCAtools provides functions for data exploration via PCA, and allows the user to generate publicationready figures. PCA is performed via BiocSingular  users can also identify optimal number of principal component via different metrics, such as elbow method and Horn's parallel analysis, which has relevance for data reduction in singlecell RNAseq (scRNAseq) and high dimensional mass cytometry data.
A pca
object, containing:
rotated
, a data frame of the rotated data, i.e., the centred and scaled (
if either or both are requested) input data multiplied by the variable loadings
('loadings'). This is the same as the 'x' variable returned by prcomp().
loadings
, a data frame of variable loadings ('rotation' variable returned
by prcomp()).
variance
, a numeric vector of the explained variation for each principal component.
sdev
, the standard deviations of the principal components.
metadata
, the original metadata
xvars
, a character vector of rownames from the input data.
yvars
, a character vector of colnames from the input data.
components
, a character vector of principal component / eigenvector names.
Kevin Blighe <kevin@clinicalbioinformatics.co.uk>
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54  options(scipen=10)
options(digits=6)
col < 20
row < 20000
mat1 < matrix(
rexp(col*row, rate = 0.1),
ncol = col)
rownames(mat1) < paste0('gene', 1:nrow(mat1))
colnames(mat1) < paste0('sample', 1:ncol(mat1))
mat2 < matrix(
rexp(col*row, rate = 0.1),
ncol = col)
rownames(mat2) < paste0('gene', 1:nrow(mat2))
colnames(mat2) < paste0('sample', (ncol(mat1)+1):(ncol(mat1)+ncol(mat2)))
mat < cbind(mat1, mat2)
metadata < data.frame(row.names = colnames(mat))
metadata$Group < rep(NA, ncol(mat))
metadata$Group[seq(1,40,2)] < 'A'
metadata$Group[seq(2,40,2)] < 'B'
metadata$CRP < sample.int(100, size=ncol(mat), replace=TRUE)
metadata$ESR < sample.int(100, size=ncol(mat), replace=TRUE)
p < pca(mat, metadata = metadata, removeVar = 0.1)
getComponents(p)
getVars(p)
getLoadings(p)
screeplot(p)
screeplot(p, hline = 80)
biplot(p)
biplot(p, colby = 'Group', shape = 'Group')
biplot(p, colby = 'Group', colkey = c(A = 'forestgreen', B = 'gold'),
legendPosition = 'right')
biplot(p, colby = 'Group', colkey = c(A='forestgreen', B='gold'),
shape = 'Group', shapekey = c(A=10, B=21), legendPosition = 'bottom')
pairsplot(p, triangle = TRUE)
plotloadings(p, drawConnectors=TRUE)
eigencorplot(p, components = getComponents(p, 1:10),
metavars = c('ESR', 'CRP'))

Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.