Biology Reference
In-Depth Information
5
Proteomics and Sample Pattern Recognition in Systems Biology
Systems biology is aiming at a holistic overview on all regulatory
processes and reactions (phenotypic plasticity) of a biological sys-
tem in response to environmental perturbations. Resolution of
these processes improves with the amount of data available.
Consequently, integration of protein- , metabolite- and transcript
data enhances the resolution. Especially, with untargeted protein
analysis but also by integrating MASS WESTERN analyzes
(Figs.
1
and
2
), huge amounts of data are generated. Therefore
statistical techniques are necessary for comprehensive data
mining and the extraction of biological relevant information.
One of the most important methods for data mining and data
visualization is a pattern recognition strategy based on supervised
and unsupervised multivariate statistics such as Principal
Components Analysis (PCA) (Fig.
2a
). Through pattern recogni-
tion, conclusions about biologically active regulatory processes
and proteins can be drawn. In Fig.
2
MAPA and MASS WESTERN
data of the same sample of
Chlamydomonas reinhardtii
grown
under either mixotrophic or photoautotrophic conditions are
shown. In Fig.
2C
the integration of MAPA and MASS WESTERN
data is shown leading to a clear separation of the two growth
conditions. Data integration and principal components analysis
was performed with COVAIN [
5
]. Each single data set was
uploaded and then integrated into one data matrix by COVAIN.
The data were z-transformed to allow the comparison of MAPA
and MASS WESTERN values and visualized by principal compo-
nents analysis using all the functions of COVAIN. In Fig.
2d
the
loadings of the integrated data are shown demonstrating that
both MASS WESTERN data specifi c for single proteins as well as
tryptic peptide precursors identifi ed by the MAPA algorithm have
an impact on sample classifi cation. All the necessary software and
manuals for a step-by-step description for all of these data pro-
cessing and data mining procedures can be downloaded at
http://www.univie.ac.at/mosys/software.html
.
6 Combining MAPA and MOAC for Quantitative Analysis of Phosphoproteins,
Signaling Cascades, and Novel Protein Kinase Targets
In Fig.
4
a convenient procedure is described for the unbiased
enrichment of phosphoproteins using different variants of a
metal-oxide-affi nity chromatography MOAC [
17
-
19
]. We call
this procedure tandem MOAC because the enrichment of phos-
phoproteins is subsequently combined with the enrichment of
phosphopeptides. The procedure was recently published and
led to the identifi cation of novel targets of the mitogen-activated
