Mass Spectrometry in Systems Biology - Systems Biology: Networks, Models, and Applications - page 23

specific, targeted proteins, follow changes of entire cellular pathways,
or dissect cellular organelles for their bona fide proteins (for review see
refs. [69] and [70]).
An example of integrated data from a targeted MS approach to help
answer questions about specific proteins was demonstrated by Hazbun
et al. [71]. This targeted approach to assign function to 100 uncharac-
terized ORFs included MS, classical biochemical techniques (affinity
purification), cell biology protocols (two-hybrid analysis, fluorescence
microscopy), and bioinformatics analysis (structure prediction, Gene
Ontology annotation). Affinity purification followed by mass spectrom-
etry emerged as one of the techniques capable of assigning function
and biological processes to the uncharacterized ORFs [71]. In another
example, integration of MS data from proteomic analysis of mitochon-
dria with publicly available data sets and functional genomics methods
led to construction of a predictive score of protein localization in mito-
chondria [72]. Both these studies formatted the results relative to Gene
Ontology (GO) annotation [73].
Quantitative proteomic analysis using ICAT and MS complemented
with traditional biochemical assays outlined changes in cellular path-
ways [74,75]. Depending on the biological question, one could filter
these results to reveal novel functions of targeted proteins, as with
Myc oncoprotein in Shiio et al. [74], or expand knowledge on cellular
pathways by ascribing interactions of proteins showing changes in
expression in response to an agonist, for example, the response of human
liver cells to interferon [75].
Recently, tissue-specific internal standards have become available by
metabolic labeling of rats [45]. Using this method, changes in proteins
involved in xenobiotic metabolism and protein-folding machinery of
the endoplasmic reticulum between animals treated with cyclohex-
imide (a protein synthesis blocker) and controls were measured.
An alternative to relative quantitation by stable isotope labeling for
shotgun proteomics is to use information available in MudPIT experi-
ments for data clustering based on raw data, statistical modeling [57],
or subtractive proteomics. Protein expression patterns from Plasmodium
falciparum were clustered based on sequence coverage to distinguish
among stage-specific proteins of malaria parasite [76]. Subtractive pro-
teomics (which is the identification of proteins in certain samples or
fractions) [77] identified previously unknown associations of nuclear
membrane proteins with dystrophies [78].
LIPIDOMICS
Global analysis of complex mixtures of cellular lipids is the analytical
goal of the lipidomics. Cellular lipids can be classified in three major
groups: polar, nonpolar, and metabolites. Lipid metabolites are of
significant interest when comparing disease versus control samples in