spectrometry (AP/MS),^2 which identifies protein complexes, are the most widely
used experimental methods for mapping physical protein-protein interactions
(PPIs) for many organisms. The entire set of proteins, often referred to asproteome,
along with their structure and interactions, is a subject of the study calledproteo-
mics. Other high-throughput experiments provide just a subset or a type ofomics
data: microarray data andgenomicdata^3 enable the study of thegenome(complete
set of genes of a cell); RNA-Sequencing data, ortranscriptomicdata,^4 enable the
study of thetranscriptome(complete set of all RNA molecules, including mRNA,
rRNA, tRNA); andmetabolomicdata generated from mass spectrometry-based
experiments^5 enable the study of themetabolome(complete set of small-molecule
chemicals involved in metabolic processes in organism). Unfortunately, these
experiments often generate noisy and incomplete data sets and with such deficien-
cies can limit our understanding of a biological system. To overcome these exper-
imental limitations, various data integration methods have been devised.
Theomicsdata describe different layers of complex organization in molecular
biology. In the recently emerged multidisciplinary field ofnetwork biology, such
layers are represented by networks (also called graphs), where nodes are metabolites
or macromolecules such as proteins, RNA molecules, genes, etc., and edges repre-
sent various physical, functional, and chemical interactions or relationships between
them^6 (see Fig.1 for an illustration). The whole set of molecular interactions inside a
cell constitutes theiteractomenetwork. This simplified network-based representa-
tion revolutionized our understanding of biological systems as it enabled a shift from
analysis of biological functions at the level of individual genes and proteins to the
Fig. 1 An illustration of the
representation of a cell’s
molecular organization. The
representation includes four
omicslayers each of which
is represented by a
molecular network.
Interactions between nodes
in different layers are
represented bydashed lines
(^2) Gavin et al. (2006) and Krogan et al. ( 2006 ).
(^3) Dahlquist et al. ( 2002 ) and Quackenbush ( 2001 ).
(^4) Marioni et al. ( 2008 ), Mortazavi et al. ( 2008 ), and Wang et al. ( 2009 ).
(^5) Wang et al. ( 2003 ) and Smith et al. ( 2006 ).
(^6) Vidal et al. ( 2011 ).
138 V. Gligorijevic ́and N. Pržulj