PCA:Multidrug

Run PCA

result <- pca(X, ncomp = 3, center = TRUE, scale. = TRUE)
result # will output the eigenvalues, proportional of explain variance in the
       # case of non missing data

Samples representation

Let us have a first glimpse at the individual plots for the first 2 principal components

plotIndiv(result, comp = c(1, 2), ind.names = TRUE)

Not very useful for the moment. How about adding some colours?

## Let's see how many cell lines types we have we have
as.numeric(as.factor(multidrug$cell.line$Class))

## Now lets add colours to each type of cell line
col.drug = as.numeric(as.factor(multidrug$cell.line$Class))
col.drug[col.drug == 1] <- 'red'
col.drug[col.drug == 2] <- 'blue'
col.drug[col.drug == 3] <- 'yellow'
col.drug[col.drug == 4] <- 'violet'
col.drug[col.drug == 5] <- 'green'
col.drug[col.drug == 6] <- 'black'
col.drug[col.drug == 7] <- 'pink'
col.drug[col.drug == 8] <- 'brown'
col.drug[col.drug == 9] <- 'darkgreen'

## plot of the samples for the first two components
plotIndiv(result, comp = 1:2, ind.names = FALSE, col = col.drug, pch = 16)

## Add a legend
legend(0.15, -4.5, unique(multidrug$cell.line$Class), col = unique(col.drug),
pch = 16, title = "Cancer Cell line", ncol = 3, cex = 0.6)

Same individual plots but in 3D

plot3dIndiv(result, comp = 1:3, ind.names = FALSE, col = col.drug,
                    cex = 0.2, axes.box = "both")

## This command can be used to capture the image on linux
rgl.postscript('cs_multi_plotIndiv_final.ps')

NEXT: Variable representation

Pages: 1 2 3 4