PCA produced
the shown pattern of distribution without using information about
the classes. In the other words, the PC’s in PCA are calculated only
from absorbance data matrix. So this PC’s might not necessarily be
the components relevant for discrimination or regression.