Multivariate Analysis for Metabolomics and Proteomics Data - Proteomic and Metabolomic Approaches to Biomarker Discovery - page 299

FIGURE 5 SELDI-TOF spectra obtained from noncancer and melanoma serum samples, with a overlay of protein mass
spectra. Protein mass spectra obtained from sera of patients with melanoma (red) and noncancer individuals (blue) are
superimposed.

more than 50% in the intralumenal diameter) of all
three major coronary arteries. Patients in the NCA
group had chest pain and a positive exercise test
but normal coronary arteries by angiography.
The similarity in the representative spectra encour-
ages the use of PLS discriminant analysis, which is
a very sensitive classi

of their corresponding NMR chemical shifts in
columns. As mentioned previously, the goal of
the PLS analysis is to build an equation that
predicts Y values (in this case, disease status)
based on X data (in this case, NMR spectra). PCA
works to explain maximum variation between
samples; PLS-DA explains maximum separation
between de

cation method. To perform
this, the spectra were automatically data-reduced
to 245 integral segments, each comprising 0.04
ppm, before analysis, shown in Figure 7 .The
NMR data matrix, X, could then be constructed
with the 66 patients in rows and the 245 segments

ned class samples in the data (X), in
this case the classes are healthy or diseased. Once
a PLS-DA model is calculated and validated, it
can be used for prediction of class membership
for unknown samples.