Content uploaded by Anup Som
Author content
All content in this area was uploaded by Anup Som on Jul 08, 2020
Content may be subject to copyright.
In: Recent Trends in ‘Computational Omics' ISBN: 978-1-53617-941-5
Editor: Pramod Katara © 2020 Nova Science Publishers, Inc.
Chapter 9
ROLE OF NETWORK BIOLOGY
IN CANCER RESEARCH
Rinki Singh and Anup Som
Centre of Bioinformatics,
Institute of Interdisplinary Studies,
University of Allahabad, Prayagraj, India
“I think the next century will be the century of complexity.”
— Stephen Hawking, January 2000.
ABSTRACT
Cancer is one of the most prevalent diseases that causes death
worldwide. Due to its extreme heterogeneity and complexity, cancer is
often considered as a model case for complex diseases. Recent advances in
‘omics’ (viz. genomics, proteomics, transcriptomics, and metabolomics)
technologies have generated large volume of biological data with potential
to identify cancer biomarkers and their interaction networks involved in
Corresponding Author’s Email: som.anup@gmail.com.
Complimentary Contributor Copy
Rinki Singh and Anup Som
310
the formation and progression of different types of cancers. Over the past
two decades, high-throughput technologies such as next generation
sequencing (NGS), microarray and mass spectrometry have fundamentally
changed clinical cancer research and revealed novel molecular markers in
different types of cancers. Biological molecules usually exert their
functions through a complex interplay of interactions with other
biomolecules. Thus, the study of complex interaction networks of the
biomolecules is essential for understanding of molecular mechanisms of
the disease development process.
This book chapter concisely reviews cancer biomarkers at different
levels of carcinogenesis and types of cancer biomarkers based on their
implication in the cancer treatment. We comprehensively emphasize the
fundamentals of network topology whose understanding is essential for
using network biology approaches in cancer research. We also highlight
diverse network biology-based approaches that integrate different high
throughput omics data in a single conceptual framework to interpret the
data in a concise manner. We also provide a list of bioinformatics resources
(databases, software and web servers) that are routinely used for analyzing
and interpreting various omics data. Further, this chapter also entails the
challenges associated with network-based biomarker discovery.
Keywords: omics data, cancer biomarker, biological networks, hub gene,
bottleneck gene, network measures, cancer databases
1. INTRODUCTION
Cancer is a complex and heterogenous disease that is characterized by
extensive genomic abnormalities and aberrations in gene expression [1]. The
complexity of cancer is not only associated at physiological, biochemical,
and genetic level, but also at the tissue, organism, and population level. The
physiological complexity of cancers is reflected by the interactions between
tumor and their microenvironment which might promote their growth,
survival, and occurrence of distant metastasis. Cancer tissues often contain
distinct cancer subtypes that have different pathophysiological features
which contribute to cancer tissue complexity [2]. Previous studies reported
that each cancer subtype within a tumor originated through different cancer
specific-developmental mechanism due to distinct genomic alterations [3,
4]. These genomic alterations are of different kinds, such as mutations in
Complimentary Contributor Copy
Role of Network Biology in Cancer Research
311
gene sequences, gene fusions, gene alterations, chromosomal fragment
amplification, chromosomal translocations, and epigenetic modification that
are associated with a dysregulation in genes and noncoding RNAs
expressions. The heterogenous nature of cancer produces significant
challenges in preventing, treating, and gaining a deep understanding of the
pathological mechanism of cancer. Along with the overall improvement in
the quality of medical services and technologies, the prevention and
treatment of cancer have been greatly improved over the past decade;
however, death from cancer is still continuous and ever-increasing. In 2018,
18.1 million new case and 9.6 million estimated cancer deaths were reported
worldwide [5]. Thus, for effective cancer treatment, it is necessary to
identify the genes and their oncogenic pathways that are the driving force
for each cancer and its subtypes at early stage.
Tumor development is a complex process in which many biomolecules,
(i.e., genes, proteins miRNA and metabolites) and molecular processes are
involved. These biomolecules do not work independently but are organized
in co-regulated units at the system level that perform common biological
functions. Alterations in these functional elements lead to certain cancer
phenotypes (e.g., tumor growth, metastasis, drug response, and resistance
against therapy) and, consequently, these cannot be addressed by the
classical one-gene approach. Therefore, rather than focusing on the
individual gene, it is important to identify the interactions among them [6].
Recent technological advances in molecular biology have accelerated the
interest in identifying cancer biomarkers and enabling a system- level
approach of the analyses [7-10].
Network biology has been widely used to represent, compute and model
intracellular interactions to gain insight into molecular architecture of any
complex process [8, 11, 12].
Emergence of the different omics technologies including DNA
microarray, NGS, two-hybrid screening system, and Mass spectrometry
have generated huge amount of biological data that are used for the
construction of different types of network such as gene regulatory network
(GRN), gene co-expression network (GCN), protein-protein interaction
Complimentary Contributor Copy
Rinki Singh and Anup Som
312
network (PPIN), and disease-gene interaction network (DGIN)) that help in
the study of cancer-related biomarker discovery.
2. CANCER BIOMARKERS
Mostly, cancer is detected at the advance stage due to their asymptotic
nature in which cancer is too far to be treated, and most cancer treatments
are effective in only a small number of patients undergoing therapy. Thus,
there is a huge prerequisite to improve the outcomes of cancer patients by
enhancing early detection and treatment strategies [13]. A cancer biomarker
is an objectively measured characteristic that describes a normal or
cancerous state in an organism by analyzing biomolecules, i.e., DNAs,
RNAs, proteins, peptides, and chemical modifications (Figure 1).
Biomarkers are very useful for early detection, monitoring disease
progression, predicting disease recurrence and therapeutic treatment
efficacy. Besides this, cancer biomarker is also associated with dysregulated
pathways that synchronize cell growth, survival, and metastasis which
directs toward the therapeutic application of biomarkers.
Advances in cancer research and high throughput technologies coupled
with bioinformatics tools have given a new hope that biomarkers can be used
for various purposes; (i) to improve cancer screening and detection, (ii) to
improve the drug development process, and (iii) to enhance the effectiveness
of cancer treatment [14]. Ideally, an ideal marker should have the following
features: i) safe and easy to measure, ii) cost efficient, iii) Modifiable with
treatment, and iv) consistent across gender and ethnic groups. A large
number of biomolecules are reported as the biomarkers for various types of
cancer but these biomarkers sometimes do not show sensitivity and
specificity for a particular cancer, thus the estimated success rate of the
biomarkers in the clinical translation is approximately 0.1% [15]. So far,
only handful of the biomarkers are approved by the FDA. These biomarkers
are successively used in cancer detection, diagnosis, prognosis and
measuring the recurrence of the tumor progression.
Complimentary Contributor Copy
Role of Network Biology in Cancer Research
313
Figure 1. Cancer biomarkers at the different level of carcinogenesis.
Figure 2. Types of cancer biomarker based on their implication in the cancer treatment.
Complimentary Contributor Copy
Rinki Singh and Anup Som
314
Cancer biomarkers are classified in the following categories based on
their implications (Figure 2). (a) Diagnostic biomarkers: a biomarker that is
used to detect and identify a given type of cancer in an individual. An
example of a diagnostic biomarker is cologuard, a multigene DNA (KRAS
mutations, aberrant NDRG4, and BMP3 methylation) stool test combined
with fecal immunochemistry designed to screen for colorectal cancer [15].
(b) Prognostic biomarkers: these biomarkers predict the risk of clinical
outcomes such as cancer recurrence or disease progression in the future and
they, therefore, have an important influence on the aggressiveness of the
therapy. The 21-gene recurrence score which is predictive of breast cancer
recurrence and overall survival in node-negative. Tamoxifen-treated breast
cancer is an example of prognostic cancer marker [16]. (c) Predictive
biomarkers: These predict response to specific therapeutic interventions.
These biomarkers mainly predict the clinical outcomes from the molecular
characteristics of a patient's tumor, such as positivity/activation of human
epidermal growth factor receptor 2 (HER2) predicts response to trastuzumab
in breast cancer [17]. Similarly, KRAS-activating mutations predict
resistance to EGFR inhibitors such as cetuximab in colorectal cancer [18].
(d) Therapeutic Biomarkers: These are generally the proteins that directly
contribute to the tumor growth that could be used as targets for the drugs
[19]. From last few decades, extensive research have been performed to
identify molecular biomarkers for pre-symptomatic diagnosis, identification
of cancer subtype, assessment of cancer progression, prediction of patient
response to therapy, and detection of recurrences. Development process of
biomarker has been also evolving with the expansion of our omics analysis
capabilities of clinical biospecimens.
3. BIOLOGICAL NETWORKS
In living organisms, biomolecules function by interacting with the other
molecules in the cell and forming many complex biological networks,
including GRN, GCN, PPIN, and metabolic network. The molecules
working together to perform a biological function not necessarily are in
Complimentary Contributor Copy
Role of Network Biology in Cancer Research
315
direct contact, but they might be functionally linked and dysfunction of some
of these interactions causes diseases including cancer [20]. Besides these
cellular interaction networks, several other interactions were reported to
understand the complex mechanism of the tumorigenesis such as a genetic
interaction networks, phenotypic profiling networks, and disease-gene
interactions.
The multitude biological networks are mainly represented as graphs. A
graph G = (V, E) consist of a set of vertices (nodes) V and a set of edges
(links) E, where each edge is assigned to two nodes. These graphs can be
undirected or directed depending on the type of interactions among the
biomolecules. Undirected graph represents graph where connections
between nodes are direction less whereas in the directed graphs, the direction
of the interaction is taken into account. Usually, in PPINs and GCNs,
vertices are represented with arrowheads showing the direction of the
interaction (Figure 3). Interaction between transcription factors and genes is
usually represented as a directed network where the direction goes from the
transcription factors to the genes.
Figure 3. Graphical representation depicted (A) undirected and (B) directed networks.
Complimentary Contributor Copy
Rinki Singh and Anup Som
316
3.1. Protein-protein Interaction Network
Protein-protein interactions (PPIs) are crucial for all biological
processes; thus, their mapping and understanding is essential for the
comprehension of cellular physiology in normal as well as pathological
conditions. PPINs are usually undergo computational analysis and
mathematical modeling to study cellular processes, molecular functions at
the system level [21]. PPINs can be broadly divided in to two categories;
Physical interaction network that represents physical or direct contacts
among the proteins which are determined by various experimental methods
[22] and Functional interaction network which represents functional
interactions of proteins, i.e., when proteins influence the activity of other
proteins through regulation, co-expression, or some other genetic
interactions. These interactions are determined using high throughput
expression techniques like microarray and RNA-sequencing experiments.
Further, the gene expression data from the microarray/RNA-seq are used to
reconstruct co-expression network that may provide a significant overview
of functional interactions among genes.
A systematic investigation of cancer proteins in the protein-protein
interaction network provided important biological information for
uncovering the molecular mechanisms of cancer. Recent works in the area
of biological network have shown how PPIN analysis can elaborate the
difference in cancerous gene products and interactions in cancer patient and
controls due to underlying dissimilarity in network topology of both [23-27].
3.2. Gene Regulatory Network
Gene regulatory network (GRN) describes the interaction and
regulations among transcription factors and downstream genes and is usually
represented with directed graphs and inferred by gene expression data.
Directed edges represent the physical binding of transcription factors to the
genes and regulate their expressions [28]. Chromatin immunoprecipitation
coupled with high-throughput DNA sequencing technologies (ChIP-seq) is
Complimentary Contributor Copy
Role of Network Biology in Cancer Research
317
a powerful technique for the genome-wide profiling of protein-DNA
interactions. In addition to transcription factor activities, overall gene
expressions are also regulated post-transcriptionally by micro RNAs
(miRNAs) that bind to complementary cis-regulatory RNA sequences
usually located in 3'untranslated regions (UTRs) of target mRNAs. miRNAs
are not master regulators, but rather act post-transcriptionally to regulate the
levels of target mRNAs. GRNs are formed by miRNAs interacting with their
targets. In such networks, nodes are miRNAs and target genes, and edges
represent the complementary annealing of the miRNA to the genes. Thus,
GRNs are classified into two types based on the regulatory molecules [20].
It provides a systematic understanding of the molecular mechanisms
underlying biological processes by revealing how the regulatory molecules
interact with genes that carry out cell functions. Identification of GRNs has
proven valuable in a variety of contexts, including identifying druggable
targets, detecting driver genes in different types of cancers, and even
optimizing prognostic and predictive signatures [29-31].
3.3. Gene Co-expression Network
A gene co-expression network identifies which genes have a tendency
to show a coordinated expression pattern across a group of samples. Genes
that function together in a common biological process are expected to show
greater similarities in their expression patterns than random sets of gene
products. High throughput expression technologies such as microarray and
NGS have generated a huge amount of gene expression datasets for many
organisms in different conditions. It provides a matrix that consists of all the
expressed genes in an organism at a particular condition that can be used for
the construction of gene co-expression network. In gene co-expression
networks, nodes represent genes and edge connects pairs of genes that show
correlated co-expression above a set threshold.
Complimentary Contributor Copy
Rinki Singh and Anup Som
318
Gene co-expression network represents gene-to-gene interactions and
makes it possible to investigate the whole-genome architecture under a
certain condition. Gene co-expression network construction and analysis can
be described with the following three steps: (i) Calculate the correlation of
gene expressions between each pair of genes. These relationships describe
the similarity between expression patterns of the gene pair across all the
samples. Different measures of correlation have been used to construct
networks including Pearson’s or Spearman’s correlations. Many of other
methods can also be used to construct co-expression [32] (ii) Construction
of co-expression network by setting a threshold correlation value (iii)
Identification of modules (group of genes) having similar gene expression
pattern across the samples using unsupervised machine learning methods
like k- mean and hierarchical clustering to produce co-expressed genes
rather than only gene pairs. Subsequently functional enrichment analysis of
the modules can be performed to identify the biological and molecular
functions in which these genes are involved.
Gene co-expression networks can be either weighted or unweighted. In
a weighted co-expression network, all genes connected to each other have a
weight value associated with each connection ranging from 0 to 1 which
indicates the strength of a correlation between each gene pairs. In an
unweighted co-expression network, the weight value is binary, i.e., 0 or 1.
Currently, weighted gene co-expression analysis (WGCNA) is preferred
over unweighted because it produces more robust results. WGCNA has been
applied to various diseases, including cancer, for identifying therapeutic
targets and associated genes in tumorigenesis [33-35].
3.4. Metabolic Network
The metabolic network comprehensively describes all possible
biochemical reactions for a particular cell or organism at different time
points.
Complimentary Contributor Copy
Role of Network Biology in Cancer Research
319
In the metabolic networks, nodes are metabolites and edges are either
the reactions that convert one metabolite into another or the enzymes that
catalyze these reactions. In these networks, edges can be directed or
undirected depending on whether a given reaction is reversible or not. A
metabolic network differs from PPINs in terms of technical and functional
properties. PPINs demonstrate that the deletions of the most highly
connected proteins (hubs) may correlate well with lethal phenotypes
whereas, a node (i.e., a metabolite) in metabolic networks cannot be deleted
by genetic techniques, but edges can [36]. Cancer cells undergo significant
metabolic alterations and adaptation that contribute to the invasion,
metastatic and immunosuppression in cancer cells. Metabolic
reprogramming in cancer cell complicates our understanding of the
molecular basis of carcinogenesis. Thus, the analysis of metabolic network
plays a crucial role in identification of the therapeutic drug targets against
cancers [36-39].
4. PROPERTIES OF A BIOLOGICAL NETWORK
AND THEIR SIGNIFICANCE
The topology of a biological network plays a crucial role in
understanding of the complex biological systems. Often, biological
networks follow a specific topology that allows scientists to go through a
deeper investigation towards knowledge extraction for identifying the
cellular mechanism. There are various parameters for the network topology
that help to understand the structure of a network and facilitate
understanding of the hidden mechanisms of the biological processes. The
most robust measures include node degree distribution, clustering
coefficient, betweenness centrality, average path length, diameter, and
density which are described below:
Complimentary Contributor Copy
Rinki Singh and Anup Som
320
4.1. Degree and Degree Distribution
One of the most basic properties of a node is its degree (k). The degree
of a node, k, is the number of direct links to this node. In a directed network,
a node has an out-degree that is defined as the number of edges going out of
it and an in-degree defined as the number of edges coming into it. The more
links a node has, more important it becomes in terms of the network stability.
An important measure of a network is the distribution of the number of
connections per node (Degree distribution). Degree distribution P(k) is the
probability that a node has exactly k links in the network [40].
4.2. Scale-free Network, Hubs and Their Attributes
Most biological networks are scale-free which means few nodes have
maximum links (called hubs) and maximum nodes have the minimum
number of links (minimum degrees) (Figure 4). This heterogenous node-
degree distribution is followed by power-law distribution; P(k) ~ kγ, with an
exponent γ that ranges between 2 to 3 [40, 41]. The emergence of hubs is a
consequence of a scale-free property of networks. Hubs have a significant
impact on the network topology in regards to the topological and functional
robustness and have the following attributes:
1. Shortening the path lengths in a network: More hubs in a network
impute more reduced distances between nodes because in a scale-
free network hubs serve as connecting links between the small
degrees nodes. Thus, the distance of two random nodes in a scale-
free network is small. Hence, it is referred to as ‘small’ or ‘ultra-
small’ world network [41].
2. Aging of hub nodes: In a biological network, the hub genes/proteins
were found to be more conserved at the sequence level [42]. Thus,
these high degree nodes are expected to be the ancestral nodes.
Tendency of a newly evolved node (genes/proteins) to connect to an
ancestral node is called preferential attachment.
Complimentary Contributor Copy
Role of Network Biology in Cancer Research
321
Figure 4. Random deletion vs targeted deletion showing the importance of hubs in a
biological network.
3. Robustness and Attack tolerance: Scale-free networks exhibit an
unexpected degree of robustness i.e., because of their
inhomogeneously wired structure, random node disruptions do not
lead to a major loss of connectivity, but the loss of the hubs causes
the breakdown of the network into isolated clusters [43], (Figure 4).
Thus, scale-free networks have the ability to maintain their
functions even the structure of the complex network changed
significantly. Hub nodes have very important biological
significance due to their high connectivity and deletion of a hub
node has catastrophic consequences (centrality-lethality rule) [44].
Consequently, cell is vulnerable to the loss of highly interactive
hubs. A study by Jonsson et al. reported that cancer proteins
participate in central hubs rather than a peripheral one and more
actively participate in protein-protein interactions [45].
4. Degree correlation: It is one of the remarkable features of the scale-
free networks, where the probability that two nodes are attached
depends on their degree. Biological network shows dichotomy in
degree correlation, i.e., assortative (links between nodes with
Complimentary Contributor Copy
Rinki Singh and Anup Som
322
similar degree level are connected) and disassortative (hubs are
connected to the nodes having a small degree). Many properties of
networks explained by this dichotomy of degree correlation include
neighborhood connectivity, sickle shape clustering coefficient
distribution, and modularity. The influence of degree correlation on
robustness and interconnectivity of networks is of fundamental
importance as it implies an evolutionary advantage behind its
existence [46].
5. Spreading phenomenon: Hubs are called super-spreader as they
effectively spread information. In a biological system hub genes are
more prone to mutations and affect the activity of other genes
associated with them and produce disease phenotypes. Therefore,
selective deletion of the hubs might be a good strategy for the
prevention of the complex disease [47].
4.3. Clustering Coefficients and Module
The clustering coefficient, C, of a node is the ratio of the number of
actual connections between the neighbors of node i (ni) to the number of
possible connections.
For a node i with degree , the clustering coefficient is defined as
The parameter of a clustering coefficient measures the degree to which
nodes in a network tend to cluster together. It ranges between the value of 0
to 1 [48]. The overall clustering coefficient of a network with node N is
given by . Usually, biological networks have a higher clustering
coefficient than random networks. The high clustering coefficient means
that the networks contain communities or groups of nodes that are densely
connected and give the modular structure to the network (Figure 5). For
example, a transcriptional module: a set of co-regulated genes involved in a
Complimentary Contributor Copy
Role of Network Biology in Cancer Research
323
common function. Thus, module is a highly connected structure in a
network. The finding of modules in a network is very important because the
biomolecules that occur in modules tend to be enriched in common
biological functions [49]. Therefore, one can identify network modules that
are enriched in genes/proteins known to be associated with cancer
progression. Other proteins in these modules, which are not previously
known to be associated particularly with cancer, may be promising cancer-
associated candidate genes. Currently, the identification of functionally
relevant sub-networks or modules from gene co-expression and protein-
protein interaction networks are gaining popularity in cancer research [50,
51].
4.4. Shortest and Average Path Lengths
Any two nodes in a biological network can be connected through
multiple edges. The length of the path is the number of edges connecting to
it. The shortest path between node i and j is the path with the smallest number
of links and it is usually called as the distance between nodes i and j, and
denoted by dij, or simply d (Figure 5). In an undirected network dij = dji, i.e.,
the distance between node i and j is equal to the distance between node j and
i. In a directed network often dij ≠ dji, as we can have many shortest paths of
the same length between a pair of nodes.
In a network, the average path length (d) is defined as the average of
the all shortest paths between all pairs of nodes. For a directed network of N
nodes, d is;
Most of the biological networks have a very short average path length
leading to the concept of the ‘small world’ where each node is connected to
every other node through the shortest path possible [41].
Complimentary Contributor Copy
Rinki Singh and Anup Som
324
Figure 5. Representing a hub, bottleneck and module structure in a network.
4.5. Diameter
The diameter (dmax= max(dij)) of a network is defined as the maximum
distance between any two nodes i and j. It is a general parameter that
indicates the compactness of the network. For example, the diameter of a
protein-signaling network can be interpreted as the overall easiness of the
proteins to communicate.
4.6. Betweenness Centrality and Bottleneck
Betweenness centrality (BC) is defined as the fraction of shortest path
going through a given node. The betweenness of a node I is given by:
Where, σjk is the total number of shortest paths from node j to k and σjk(i)
is the number of shortest paths from j to k going through i. Thus,
Complimentary Contributor Copy
Role of Network Biology in Cancer Research
325
betweenness centrality shows important nodes that lie on a high proportion
of paths between the other nodes in the network [48]. Without these nodes,
there would be no possible way for two neighbors to communicate with each
other. For example, in PPINs, proteins with high BCs play a role as key
connectors with essential functional and dynamic properties. Nodes that
have high number of ‘shortest paths’ (with high centrality) going through
them have been termed ‘bottleneck nodes’. Bottleneck nodes play crucial
roles in mediating communication within a given network because they
facilitate information flow between modules or relatively densely connected
sub-networks. Such nodes are therefore like chokepoints in a network and
have been described as connecting links between community structures.
Disruption of a bottleneck structure can stop information flow, since there
are few or no alternative routes around the bottleneck (Figure 5). Bottleneck
nodes are highly correlated with essentialities than hub nodes [52].
4.7. Network Density
Network density is defined as the ratio of the existing number of links
in a network to the maximum possible links. It measures the compactness of
a network, thus allows effective flow of information within a shorten path
[41].
5. DIFFERENT LEVELS OF OMICS DATA
Network-based analysis integrates various high throughput omics
profiles (genomics, transcriptomics, proteomics, and metabolomics) that had
vastly improved our knowledge of the molecular basis of cancer progression
and ability to identify novel cancer biomarkers [53]. A review of Yan et al.
described a network-based cancer biomarker discovery approach more
comprehensively using different levels of omics data [54].
Complimentary Contributor Copy
Table 1. Various “omics” technologies and associated databases
Data type
Description
*Platforms & databases
Genomics
A variety of genomic alterations, contribute to tumor formation and progression. Cancer genome
sequencing is an ongoing process that enables identifications of these genomic alterations causing
cancer progression. A systems biology-based study of cancer genome sequencing outputs is
useful for understanding tumorigenesis as well as for the identification of valuable biomarkers
and drug targets.
Platform: Microarray, DNA-seq
(WGS, WES)
Databases: TCGA, ICGC, SRA,
GEO, dpGap, EGA, cBioportal,
UCSC Genome browser gateway, and
CCLE
Transcriptomics
At the transcriptomic level, GCN and GRNs are most widely used for biomarker discoveries.
These networks are reconstructed using gene expression data. Over the last three decades,
methodological breakthroughs have repeatedly revolutionized transcriptomic data analysis and
more knowledge about the transcriptome lead to earlier and more accurate diagnosis of cancer
subtypes as well as the development of targeted therapeutics.
Platform: Microarray/ RNA-seq
Databases: TCGA, ICGC, SRA,
GEO, dpGap, EGA, cBioportal,
UCSC Genome browser gateway,
CCLE, and ArrayExpress
Proteomics
Proteomics analyses allow quantitative and qualitative assessments of protein expression changes
related to cellular responses, thus providing a very efficient tool for monitoring targeted therapy.
Proteomics has been increasingly applied to cancer research, especially for biomarker discovery
based on PPINs. Various tools have been developed for PPINs analysis.
Platform: Protein microarray/Mass
spectrometry
Databases: Pride archive, protein
atlas, and GEO
Metabolomics
Metabolomics is the study of the identification and quantification of small molecules in the
biological system. In recent years, extensive efforts have been devoted to reveal the mechanisms
underlying metabolic alterations in cancer. It supported the hypothesis that metabolic changes
are related to the shifted balance among growth, apoptosis, and differentiation. Therefore, it can
provide novel potential biomarkers for the detection and treatment of cancer and response to
therapeutic intervention.
Platform: NMR/Mass spectrometry
Databases: HMDB, BIGG, BMRB,
MMCD, and others
Complimentary Contributor Copy
Data type
Description
*Platforms & databases
Epigenomics
Epigenetic changes such as modification of histones, chromatin, or DNA methylation status,
may affect the gene expression patterns without alteration of DNA sequences. Cancer cells are
prone to epigenetic changes. Aberrant chromatin-remodeling mechanisms such as DNA
methylation and histone acetylation are epigenetic modifications that have been observed in
tumor tissues leading to the inhibition of tumor suppressor gene expression. Understanding
these mechanisms may provide biomarkers of response to cancer chemotherapy.
Platform: Chip seq, chip microarray
Databases: NCBI epigenomics,
Encode, TCGA, SRA, EpiFactors,
MethCNA, and dbEM
*GEO, Gene expression omnibus; SRA, The sequence read archives; UCSC, The university of California santa cruz Genome browser; CCLE, Cancer cell
line encyclopedia; HMDB, Human metabolome database; BIGG, Biochemical genetic and genomic knowledge; BMRB, Biological magnetic
resonance database; MMCD, Madison metabolomic consortium database; EGA, European-genome-phenome archives.
Complimentary Contributor Copy
Rinki Singh and Anup Som
328
A brief description of different levels of omics data used in the field of
cancer biomarker discovery and associated platforms and databases are
summarized in Table 1.
The emergence of the omics has given us a pipeline around which we
can develop the infrastructure required for biomarker discovery. Biological
networks are constructed by integrating data from different resources like
interactions data from different biological databases, or directly inferred
from high throughput experimental data. Here, we give a comprehensive
overview of the steps used for the identification of cancer biomarkers from
different levels of omics data using computational network approach (Figure
6).
Figure 6. Pipeline for biomarker identification using biological network-based
approach.
Complimentary Contributor Copy
Role of Network Biology in Cancer Research
329
6. BIOINFORMATICS RESOURCES (DATABASES,
SOFTWARE AND WEB SERVERS) FOR CANCER RESEARCH
Over the past few decades omics data of tumor samples became publicly
accessible to analyze. Various tools and software are developed that provide
great opportunities to explore important cancer associated molecules,
therapeutic targets, diagnostic, and prognostic biomarkers. In this section,
we describe network-based databases, tools, and software that are used in
cancer research.
6.1. Cancer Specific Gene/Protein/Mirna Interaction Databases
With technological advances and international efforts such as
International cancer genome consortium (ICGC) [55] and The Cancer
Gnome Atlas (TCGA) [56], there is an exponential growth of cancer-
associated data from different resources, such as genome-wide association
studies, gene expression experiments, gene-gene or protein-protein
interaction data, enzymatic assays, epigenomics, immunomics, and
cytogenetics, stored in relevant repositories.
These data are complex and heterogeneous, ranging from unprocessed
and unstructured data to well-annotated and structured data. Therefore, the
storage, mining, retrieval, and analysis of these data in an efficient and
meaningful manner pose a major challenge to the biomedical researchers.
These obstacles are being overcome by developing a number of databases.
These databases can be categorized depending on the source of information
they have, such as the database of cancer driving genes, genetic variation
databases, transcriptomic databases, oncomiRs, and interaction network
databases. The list of such databases is provided in the review of
Pavlopoulou et al. [57].
Besides these databases, various other cancer-related databases have
been developed that stored the data from other sources like epigenetic
modification, phosphorylation, cell line information, drug resistance,
Complimentary Contributor Copy
Rinki Singh and Anup Som
330
cytogenetics, and anticancer agents. There are several databases that focused
on certain types or subtypes of cancers such as Cervical Cancer Gene
Database (CCDB), Dragon Database of Genes associated with Prostate
Cancer (DDPC), CuratedOvarianData, and HLungDB [58-62]. As this
chapter focuses on the network-based approaches for cancer research, we
provide a list of cancer related genes/proteins/miRNAs interaction network
databases.
Table 2. List of useful cancer gene/protein/miRNA
interaction databases
Database
Description/Reference
NCG
A web-based repository of systems-level properties of cancer genes and
associated miRNA. It has information on 2,372 candidate genes from manually
curated publications. http://ncg.kcl.ac.uk/, [63].
CDGnet
It combines information from biological networks related to the cancer types and
specific alterations, FDA-approved targeted cancer therapies and indications,
additional gene-drug information, and data on whether given genes are
oncogenes. http://epiviz.cbcb.umd.edu/shiny/CDGnet/, [64].
CancerNet
A cancer specific database which contains cancer associated PPIs, miRNA-
target, miRNA-miRNA interaction networks across 33 human cancer types.
http://bis.zju.edu.cn/CancerNet, [65].
Oncoppi
A publicly available user-friendly interface that bridges cancer genomics,
clinical and pharmacological data with a network of experimentally determined
direct interactions between cancer-associated proteins. It also has flexible query
options that allows exploration, visualization, and export of the PPIs network of
cancer- associated genes http://oncoppi. emory. edu. /[66].
TCSBN
It is a database of tissue and cancer specific biological networks. It contains co-
expression networks of 46 human tissues and 17 cancers using data from GTEx
and TCGA databases. It also contains tissue specific integrated networks for
liver, muscle and adipose tissues by integrating metabolic, regulatory and PPINs.
http://inetmodels.com/, [67].
CGIdb
Cancer genetic interaction database (CGIdb) contain synthetic viability (SV)
interactions and synthetic lethality (SL) interactions data for the specific types
of cancers. These data were generated by mining of copy number alteration and
whole-exome mutation profiles from The Cancer Genome Atlas (TCGA) and
other studies. http://www.medsybio.org/CGIdb, [68].
Complimentary Contributor Copy
Role of Network Biology in Cancer Research
331
6.2. Other Protein-Protein and Protein-Chemical Interaction
Databases Used for Cancer Research
Various interaction data generated by high throughput experiments or
predicted computationally, were submitted in various databases that are
publicly available.
These databases have various applications in cancer research such as:
(i) identification of new candidate genes involve in cancer progression
(ii) annotation of newly identified cancer genes
(iii) identification of interacting partner(s) of the candidate genes
(iv) identification of drug like small molecules interacting with some
particular genes
(v) prediction of regulatory molecules like transcription factors and
miRNAs of the cancer genes.
Here, we present a list consisting important interaction databases that
can be used in the cancer research (Table 3).
6.3. Tools and Software for Network Construction, Visualization
and Analysis
To gain insights into the biological networks, it is important to analyze
the networks. Currently, various tools and software are available for
biological network construction and visualization. Here, we give a brief
description of freely available software and web-based tools (Table 4).
These software and web-based tools facilitate network analysis by
calculating complex network parameters like average clustering
coefficients, shortest paths, and node degrees, as well as centrality measures
like stress centrality, betweenness centrality, and closeness centrality.
Complimentary Contributor Copy
Table 3. List of important protein-protein and protein-chemical interaction databases that can be used for
cancer research
Database
Description/ Reference
MINT
Molecular interaction database (MINT) stores information of the experimentally identified protein-protein interactions from the
scientific literature by expert curators. http://mint. bio. uniroma2. it/minthttp://mint. bio. uniroma2. it/mint, [69].
DIP
The database of interacting proteins (DIP) consists of experimentally determined protein-protein interaction data. It provides an
integrated set of tools for retrieving information about protein- protein interaction networks. http://dip. doe-mbi. ucla. edu, [70].
InAct
It is an open source database which provides toolkit for storage, visualization and analysis of the protein-protein interaction network.
http://www. ebi. ac. uk/intact, [71].
HPRD
Human protein reference database (HPRD), contains information about human protein–protein interactions, post-translational
modifications, enzyme–substrate relationships and its disease association. Data are manually curated by experts. https://hprd. org/,
[72].
BioGrid
The Biological General Repository for Integration Datasets (BioGrid) is a public database that contains information of protein-protein
and genetic interactions, chemical associations and post transcriptional modification. It provides various tools/software to make the
use of its interaction data [73].
PINA
Protein interaction network analysis (PINA) is an integrated platform for protein interaction network construction, visualization, and
analysis. It retrieves various protein-protein interaction data from other databases such as InAct, BioGRID, MINT, DIP, HPRD, and
MIPS/Mpact and builds a complete, non-redundant protein interaction dataset for several model organisms. Moreover, it provides a
variety of built-in tools to filter and analyze the network for gaining insights from the network. https://omics. bjcancer. org/pina/,
[74].
HitPredict
Integrates experimentally identified protein-protein interaction data from various databases including InAct, BioGrid, DIP, MINT,
and others. http://www. hitpredict. org/, [75].
Complimentary Contributor Copy
Database
Description/ Reference
PICKLE
It is a meta-database that integrates the data curated by several primary databases in one repository. http://www. pickle. gr/, [76].
STRING
It is a comprehensive protein-protein interaction database that contains >2000 million interactions from 5090 organisms. It retrieves
interactions from several evidences (text mining, experiments, databases, co-expression, neighborhood, gene fusion and co-
occurrence) and filters interactions data based on the confidence scores. https://string-db. org/, [77].
APID
Agile protein interaction data analyzer (APID) allows the exploration and analysis of currently known information about protein-
protein interactions for more than 400 organisms. These interactions are experimentally validated with protein-protein physical
interactions from primary molecular interactions databases and also from experimentally resolved 3D structures (PDB) where more
than two distinct proteins have been identified, http://cicblade.dep.usal.es:8080/APID/init.action, [78].
Complimentary Contributor Copy
Table 4. List of resources (software/web-based tools) for network construction, visualization and analysis
Resources
Description/Reference
Software
Cytoscape
An open source platform providing biological network analysis and visualization. With more than 200 registered plugins contributed by
the community, it is very versatile in network applications, such as network importing, network integrating, inference customization,
literature mining, topological clustering, functional enrichment analysis, network comparison, and programmatic access.
NetworkAnalyzer, CentiScape and Cytohubba are Cytoscape plugins providing simple interface to compute various topological
parameter for undirected and/or directed networks. https://cytoscape. org/, [79].
Medusa
A powerful tool for visualization and analysis of the molecular interaction network. It is highly interactive and supports weighted and/or
unweighted multi-edged directed and/or undirected networks. It contains a variety of layouts (Grid, random, circular, hierarchical, spring
embedding, distance geometry, parallel coordinates, and Fruchterman-Reingold force directed layout) and clustering methods such as
the affinity propagation, k-Means, and spectral clustering for comprehensive visualization and advanced data analysis. It is offered both
as a standalone application and an applet. https://sites. google. com/site/medusa3visualization/, [80].
Gephi
An open source software for network construction and visualization. Different modules in Gephi can import, visualize, spatialize, filter,
manipulate and export all types of networks. For network visualization, it uses a 3D render engine to display large networks in real-
time. It runs on Windows, Linux, and Mac operating system. https://gephi. org/, [81].
BioLayout
Express3D
It is released under GNU public license for visualization and analysis of biological networks. It runs on Windows, Linux, Mac, and other
operating systems. It is highly interactive and provides 3D visualization of molecular interaction networks. In the newly released version,
the Markov clustering algorithm (MCL) has become an integral part of BioLayout Express3D for clustering analysis and offers various
analytical approaches to microarray data analysis. https://kajeka. com/graphia-professional/, [82].
MCODE
It is a cytoscape plugin that detects densely connected region in a network. Specifically oriented to find out modular structure in the
networks [83].
Web-based tools
CellMaps
An open-source web-based tool that brings advanced visualization and network analysis to users. It supports local files in SIF, XLSX
or text formats to import network and can be exported in SVG or SIF formats. http://cellmaps. babelomics. org/, [84].
Complimentary Contributor Copy
Resources
Description/Reference
RegNetwork
It integrates a comprehensive set of experimentally observed or predicted transcriptional and post-transcriptional regulatory
relationships among transcription factors (TFs), microRNAs (miRNAs) and target genes. http://www. regnetworkweb. org/, [85].
GeNeCK
An efficient and easy-to-use web application for gene regulatory network construction from expression data. It also gives the
information of hub genes in the networks. http://lce. biohpc. swmed. edu/geneck, [86].
VisANT
An integrative visual analysis tool for biological networks and pathways. http://www. visantnet. org/visantnet. html, [87].
3OMICS
A web-based tool for analysis, integration, visualization of human transcriptomics, proteomics, and metabolomics data.
https://3omics. cmdm tw/, [88].
CoExpNetViz
A tool for comparative co-expression network construction and visualization. It is freely available both as web server and as a
Cytoscape plugin. Both versions of the CoExpNetViz tool support LINUX and Windows operating system. http://bioinformatics.
psb. ugent. be/webtools/coexpr/, [89].
Network
Analyst
An online visual analytic platform that offers transcriptome profiling, network analysis, and meta-analysis for gene expression data.
NetworkAnalyst also contains advanced network analysis features that can be used to identify potentially functionally relevant
network modules. https://www. networkanalyst. ca/, [90].
Complimentary Contributor Copy
Rinki Singh and Anup Som
336
7. CHALLENGES OF NETWORK-BASED
BIOMARKER DISCOVERY
The past decade has seen rapid developments in network models for
cancer biomarker discovery. Different network-based methods have
provided a new paradigm and hold a great promise for the future study of
cancer. However, we still have several future challenges in the field of
biomarker discovery based on the network approach.
1. Due to the heterogeneous properties of cancer, there are differential
responses from individual biomarkers, making the identification of
clinically useful and precise biomarkers for cancer diagnosis and
predicting clinical outcomes quite difficult. The current methods
lack effective validation, especially in large and multiple datasets,
which is also a significant problem for identifying efficient and
clinically useful biomarkers.
2. Since a particular type of biological network provides only one
viewpoint of the cell, integration with other data types is necessary
to get a more elaborative picture of cellular events. Methods for
integration of various omics level data are still relatively weak, and
further, most of the methods can only integrate two or three different
omics data.
3. The biological interaction networks are often static, as they do not
show the dynamic changes in the cellular state over time.
4. The continued refinement of the algorithms and tools based on
networks is critically needed and will have a significant impact on
the development of personalized biomarkers.
5. The development of robust and standardized methods for the
assessment of molecular biomarkers, especially the sub-network
biomarkers, will be essential in the future.
Complimentary Contributor Copy
Role of Network Biology in Cancer Research
337
CONCLUSION
Biological networks provide a conceptual and intuitive framework to
investigate, model, characterize, and understand complex interactions of the
biomolecules of any living system. Different network-based methods have
provided a new paradigm and hold a great promise for unfolding the
complex molecular mechanisms behind cancer diseases. It has been
described that how bimolecular interaction networks can elucidate the
molecular basis of cancer progression, which in turn can appraise methods
for prevention, diagnosis and treatment. Analysis of the topological
properties of the networks will reveal crucial genes associated with cancer
progression. Current limitations of network-based biomarker discovery can
be overcome by employing multilayer network-based approach combining
with multi-scale omics data.
CONFLICT OF INTEREST
The author (s) declared no potential conflicts of interest with respect to
the research, authorship, and /or publication of this article.
ACKNOWLEDGMENTS
The authors would like to thank University Grant Commission (UGC)
and Department of Biotechnology (DBT) India for the support. RS thanks
Priyanka Kumari, Arindam Ghosh, Anamika, Aparna Chaturvedi, Amresh
Sharma, and Birendra Singh Yadav for giving valuable suggestions
regarding the improvement of this chapter. AS thank Pramod Katara for
reviewing the chapter.
Complimentary Contributor Copy
Rinki Singh and Anup Som
338
REFERENCES
[1] Hassanpour, SH, Dehghani M. (2017). Review of cancer from
perspective of molecular. Journal of Cancer Research and Practice
4(4):127-129.
[2] Le Van T, van Leeuwen M, Carolina Fierro A. et al. (2016).
Simultaneous discovery of cancer subtypes and subtype features by
molecular data integration. Bioinformatics. 32(17):i445–i454.
doi:10.1093/bioinformatics/btw434.
[3] He L, Lv Y, Song Y, Zhang B. (2019). The prognosis comparison of
different molecular subtypes of breast tumors after radiotherapy and
the intrinsic reasons for their distinct radiosensitivity. Cancer Manag
Res. 11:5765–5775. doi:10.2147/CMAR.S213663.
[4] Barretina J, Taylor BS, Banerji S. et al. (2010). Subtype-specific
genomic alterations define new targets for soft-tissue sarcoma therapy.
Nat Genet. 42(8):715–721. doi:10.1038/ng.619.
[5] Bray F, Ferlay J, Soerjomataram I, Siegel RL, Torre LA, Jemal A.
(2018). Global cancer statistics 2018: GLOBOCAN estimates of
incidence and mortality worldwide for 36 cancers in 185 countries. CA
Cancer J Clin. 68(6):394–424. doi:10.3322/caac.21492.
[6] Koutsogiannouli E, Papavassiliou AG, Papanikolaou NA. (2013).
Complexity in cancer biology: is systems biology the answer?. Cancer
Med. 2(2):164–177. doi:10.1002/cam4.62.
[7] Abu-Asab MS, Chaouchi M, Alesci S. et al. (2011). Biomarkers in the
age of omics: time for a systems biology approach. OMICS.
15(3):105–112. doi:10.1089/omi.2010.0023.
[8] Verma, Y, Yadav A, Katara P. (2020). Mining of cancer core-genes
and their protein interactome using expression profiling based PPI
network approach. Gene Reports. 18:100583.
[9] Strunz S, Wolkenhauer O, de la Fuente A. (2016). Network-Assisted
Disease Classification and Biomarker Discovery. Methods Mol Biol.
1386:353–374. doi:10.1007/978-1-4939-3283-2_16.
[10] Singh R, Som A. (2018). Profiling of ovarian cancer reveals common
features shared by sub-types of ovarian cancer: A network biology
Complimentary Contributor Copy
Role of Network Biology in Cancer Research
339
approach towards finding therapeutic drug targets. International
conference on bioinformatics and systems biology (BSB). IEEE, 2018.
[11] Zhu X, Gerstein M, Snyder M. (2007). Getting connected: analysis and
principles of biological networks. Genes Dev. 21(9):1010–1024.
doi:10.1101/gad.1528707.
[12] Ghosh A, Som A. (2020). RNA-Seq analysis reveals pluripotency-
associated genes and their interaction networks in human embryonic
stem cells. Computational Biology and Chemistry. 85:107239.
[13] Golemis EA, Scheet P, Beck TN. et al. (2018). Molecular mechanisms
of the preventable causes of cancer in the United States. Genes Dev.
32(13-14):868–902. doi:10.1101/gad.314849.118.
[14] Ilyin SE, Belkowski SM, Plata-Salamán CR. (2004). Biomarker
discovery and validation: technologies and integrative approaches.
Trends Biotechnol. 22(8):411–416. doi:10.1016/j.tibtech.
2004.06.005.
[15] Goossens N, Nakagawa S, Sun X, Hoshida Y. (2015). Cancer
biomarker discovery and validation. Transl Cancer Res. 4(3):256–
269. doi:10.3978/j.issn.2218-676X.2015.06.04.
[16] Mamounas EP, Russell CA, Lau A, Turner MP, Albain KS. (2018).
Clinical relevance of the 21-gene Recurrence Score® assay in
treatment decisions for patients with node-positive breast cancer in the
genomic era. NPJ Breast Cancer. 4:27. doi:10.1038/s41523-018-
0082-6.
[17] Damodaran S, Olson EM. (2012). Targeting the human epidermal
growth factor receptor 2 pathway in breast cancer. Hosp Pract (1995).
40(4):7–15. doi:10.3810/hp.2012.10.997.
[18] Shankaran V, Obel J, Benson AB (2010). 3rd. Predicting response to
EGFR inhibitors in metastatic colorectal cancer: current practice and
future directions. Oncologist. 15(2):157–167. doi:10.1634/
theoncologist.2009-0221.
[19] Shen, Z. (2013). Cancer biomarkers and targeted therapies. Cell Biosci
3, 6https://doi.org/10.1186/2045-3701-3-6.
[20] Vidal M, Cusick ME, Barabási AL. (2011) Interactome networks and
human disease. Cell. 144(6):986–998. doi:10.1016/j.cell.2011.02.016.
Complimentary Contributor Copy
Rinki Singh and Anup Som
340
[21] Hakes L, Pinney JW, Robertson DL, Lovell SC. (2008) Protein-protein
interaction networks and biology--what's the connection?. Nat
Biotechnol. 26(1):69–72. doi:10.1038/nbt0108-69.
[22] Rao VS, Srinivas K, Sujini GN, Kumar GN. (2014). Protein-protein
interaction detection: methods and analysis. Int J Proteomics.
2014:147648. doi:10.1155/2014/147648.
[23] Sun J, Zhao Z. (2010). A comparative study of cancer proteins in the
human protein-protein interaction network. BMC Genomics. 11 Suppl
3(Suppl 3):S5. doi:10.1186/1471-2164-11-S3-S5.
[24] Kar G, Gursoy A, Keskin O. (2009). Human cancer protein-protein
interaction network: a structural perspective. PLoS Comput Biol.
5(12):e1000601. doi:10.1371/journal.pcbi.1000601.
[25] Suman S, Mishra A. (2018). An interaction network driven approach
for identifying biomarkers for progressing cervical intraepithelial
neoplasia. Sci Rep. 8(1):12927. doi:10.1038/s41598-018-31187-x.
[26] Luo T, Wu S, Shen X, Li L. (2013). Network cluster analysis of
protein-protein interaction network identified biomarker for early
onset colorectal cancer. Mol Biol Rep. 40(12):6561–6568.
doi:10.1007/s11033-013-2694-0.
[27] Li BQ, Huang T, Liu L, Cai YD, Chou KC. (2012). Identification of
colorectal cancer related genes with mRMR and shortest path in
protein-protein interaction network. PLoS One. 7(4):e33393.
doi:10.1371/journal.pone.0033393.
[28] Emmert-Streib F, Dehmer M, Haibe-Kains B. (2014). Gene regulatory
networks and their applications: understanding biological and medical
problems in terms of networks. Front Cell Dev Biol. 2:38.
doi:10.3389/fcell.2014.00038.
[29] Chudasama D, Bo V, Hall M. et al. (2018). Identification of cancer
biomarkers of prognostic value using specific gene regulatory
networks (GRN): a novel role of RAD51AP1 for ovarian and lung
cancers. Carcinogenesis. 39(3):407–417. doi:10.1093/carcin/bgx122.
[30] Emmert-Streib F, Dehmer M, Haibe-Kains B. (2014). Gene regulatory
networks and their applications: understanding biological and medical
Complimentary Contributor Copy
Role of Network Biology in Cancer Research
341
problems in terms of networks. Front Cell Dev Biol. 2:38.
doi:10.3389/fcell.2014.00038.
[31] Zhang W, Zang J, Jing X. et al. (2014). Identification of candidate
miRNA biomarkers from miRNA regulatory network with application
to prostate cancer. J Transl Med. 12:66. doi:10.1186/1479-5876-12-
66.
[32] Kumari S, Nie J, Chen HS. et al. (2012). Evaluation of gene association
methods for coexpression network construction and biological
knowledge discovery. PLoS One. 7(11):e50411. doi:10.1371/
journal.pone.0050411.
[33] Singh R, Som A. (2020). Identification of common candidate genes
and pathways for progression of ovarian, cervical and endometrial
cancers. Meta Gene 23:100634.
[34] van Dam S, Võsa U, van der Graaf A, Franke L, de Magalhães JP.
(2018). Gene co-expression analysis for functional classification and
gene-disease predictions. Brief Bioinform. 19(4):575–592.
doi:10.1093/bib/bbw139.
[35] Xia WX, Yu Q, Li GH. et al. (2019). Identification of four hub genes
associated with adrenocortical carcinoma progression by WGCNA.
PeerJ. 7:e6555. doi:10.7717/peerj.6555.
[36] Mahadevan R, Palsson BO. (2005). Properties of metabolic networks:
structure versus function [published correction appears in Biophys J.
2006 Jan 15;90(2):708]. Biophys J. 88(1):L07–L9.
doi:10.1529/biophysj.104.055723.
[37] Hu J, Locasale JW, Bielas JH. et al. (2013). Heterogeneity of tumor-
induced gene expression changes in the human metabolic network. Nat
Biotechnol. 31(6):522–529. doi:10.1038/nbt.2530.
[38] Arakaki AK, Mezencev R, Bowen NJ, Huang Y, McDonald JF,
Skolnick J. (2008). Identification of metabolites with anticancer
properties by computational metabolomics. Mol Cancer. 7:57.
doi:10.1186/1476-4598-7-57.
[39] Haanstra JR, Gerding A, Dolga AM. et al. (2017). Targeting pathogen
metabolism without collateral damage to the host. Sci Rep. 7:40406.
doi:10.1038/srep40406.
Complimentary Contributor Copy
Rinki Singh and Anup Som
342
[40] Albert R, Barabasi AL. (2000). Topology of evolving networks: local
events and universality. Phys Rev Lett. 85(24):5234–5237.
doi:10.1103/PhysRevLett.85.5234.
[41] Barabási AL, Oltvai ZN. (2004). Network biology: understanding the
cell's functional organization. Nature reviews genetics 5.2:101-113.
[42] Fraser HB, Hirsh AE, Steinmetz LM, Scharfe C, Feldman MW.
(2002). Evolutionary rate in the protein interaction network. Science.
296(5568):750–752. doi:10.1126/science.1068696.
[43] Albert R, Jeong H, Barabasi AL. (2000). Error and attack tolerance of
complex networks. Nature. 406(6794):378–382. doi:10.1038/
35019019.
[44] Jeong H, Mason SP, Barabási AL, Oltvai ZN. (2001). Lethality and
centrality in protein networks. Nature. 411(6833):41–42.
doi:10.1038/35075138.
[45] Jonsson PF, Bates PA. (2006). Global topological features of cancer
proteins in the human interactome. Bioinformatics. 22(18):2291–2297.
doi:10.1093/bioinformatics/btl390.
[46] Hao D, Li C. (2011). The dichotomy in degree correlation of biological
networks. PLoS One. 6(12):e28322. doi:10.1371/journal.
pone.0028322.
[47] Liu YY, Slotine JJ, Barabási AL. (2011). Controllability of complex
networks. Nature. 473(7346):167–173. doi:10.1038/nature10011.
[48] Albert R. (2005). Scale-free networks in cell biology. J Cell Sci118(Pt
21):4947–4957. doi:10.1242/jcs.02714.
[49] Charitou T, Bryan K, Lynn DJ. (2016). Using biological networks to
integrate, visualize and analyze genomics data. Genet Sel Evol. 48:27.
doi:10.1186/s12711-016-0205-1.
[50] Li J, Lenferink AE, Deng Y. et al. (2010). Identification of high-
quality cancer prognostic markers and metastasis network modules.
Nat Commun. 1(4):34. doi:10.1038/ncomms1033.
[51] Silverbush D, Cristea S, Yanovich-Arad G, Geiger T, Beerenwinkel
N, Sharan R. (2019). Simultaneous Integration of Multi-omics Data
Improves the Identification of Cancer Driver Modules. Cell Syst.
8(5):456–466.e5. doi:10.1016/j.cels.2019.04.005.
Complimentary Contributor Copy
Role of Network Biology in Cancer Research
343
[52] Yu H, Kim PM, Sprecher E, Trifonov V, Gerstein M. (2007). The
importance of bottlenecks in protein networks: correlation with gene
essentiality and expression dynamics. PLoS Comput Biol. 3(4):e59.
doi:10.1371/journal.pcbi.0030059.
[53] Lu M, Zhan X. (2018). The crucial role of multiomic approach in
cancer research and clinically relevant outcomes. EPMA J. 9(1):77–
102. doi:10.1007/s13167-018-0128-8.
[54] Yan W, Xue W, Chen J, Hu G. (2016). Biological Networks for Cancer
Candidate Biomarkers Discovery. Cancer Inform. 15(Suppl 3):1–7.
doi:10.4137/CIN.S39458.
[55] International Cancer Genome Consortium, Hudson TJ, Anderson W.
et al. (2010). International network of cancer genome projects. Nature.
464(7291):993–998. doi:10.1038/nature08987.
[56] Cancer Genome Atlas Research Network, Weinstein JN, Collisson
EA. et al. (2013). The Cancer Genome Atlas Pan-Cancer analysis
project. Nat Genet. 45(10):1113–1120. doi:10.1038/ng.2764.
[57] Pavlopoulou A, Spandidos DA, Michalopoulos I. (2015). Human
cancer databases (review). Oncol Rep. 33(1):3–18. doi:10.3892/
or.2014.3579.
[58] Agarwal SM, Raghav D, Singh H, Raghava GP. (2011). CCDB: a
curated database of genes involved in cervix cancer. Nucleic Acids
Res. 39(Database issue):D975–D979. doi:10.1093/nar/gkq1024.
[59] Maqungo M, Kaur M, Kwofie SK. et al. (2011). DDPC: Dragon
Database of Genes associated with Prostate Cancer. Nucleic Acids Res.
39(Database issue):D980–D985. doi:10.1093/nar/gkq849.
[60] Ganzfried BF, Riester M, Haibe-Kains B. et al. (2013).
curatedOvarianData: clinically annotated data for the ovarian cancer
transcriptome. Database (Oxford). 2013:bat013. doi:10.1093/data
base/bat013.
[61] Wang L, Xiong Y, Sun Y. et al. (2010). HLungDB: an integrated
database of human lung cancer research. Nucleic Acids Res.
38(Database issue):D665–D669. doi:10.1093/nar/gkp945.
Complimentary Contributor Copy
Rinki Singh and Anup Som
344
[62] Som A, Kumari P, Ghosh A. (2019). Advancing India’s bioinformatics
education and research: an assessment and outlook. J Proteins
Proteom. 10:257–267.
[63] Repana D, Nulsen J, Dressler L. et al. (2019). The Network of Cancer
Genes (NCG): a comprehensive catalogue of known and candidate
cancer genes from cancer sequencing screens. Genome Biol. 20(1):1.
doi:10.1186/s13059-018-1612-0.
[64] Kancherla J, Rao S, Bhuvaneshwar K. et al. (2020). Evidence-Based
Network Approach to Recommending Targeted Cancer Therapies.
JCO Clin Cancer Inform. 4:71–88. doi:10.1200/CCI.19.00097.
[65] Meng X, Wang J, Yuan C. et al. (2015). CancerNet: a database for
decoding multilevel molecular interactions across diverse cancer
types. Oncogenesis. 4(12):e177. doi:10.1038/oncsis.2015.40.
[66] Li Z, Ivanov AA, Su R. et al. (2017). The OncoPPi network of cancer-
focused protein-protein interactions to inform biological insights and
therapeutic strategies. Nat Commun. 8:14356. doi:10.1038/
ncomms14356.
[67] Lee S, Zhang C, Arif M. et al. (2018). TCSBN: a database of tissue
and cancer specific biological networks. Nucleic Acids Res.
46(D1):D595–D600. doi:10.1093/nar/gkx994.
[68] Han Y, Wang C, Dong Q. et al. (2019). Genetic Interaction-Based
Biomarkers Identification for Drug Resistance and Sensitivity in
Cancer Cells. Mol Ther Nucleic Acids. 17:688–700. doi:10.1016/
j.omtn.2019.07.003.
[69] Chatr-aryamontri A, Ceol A, Palazzi LM. et al. (2007). MINT: the
Molecular INTeraction database. Nucleic Acids Res. 35(Database
issue):D572–D574. doi:10.1093/nar/gkl950.
[70] Xenarios I, Rice DW, Salwinski L, Baron MK, Marcotte EM,
Eisenberg D. (2000). DIP: the database of interacting proteins. Nucleic
Acids Res. 28(1):289–291. doi:10.1093/nar/28.1.289.
[71] Hermjakob H, Montecchi-Palazzi L, Lewington C. et al. (2004).
IntAct: an open source molecular interaction database. Nucleic Acids
Res. 32(Database issue):D452–D455. doi:10.1093/nar/gkh052.
Complimentary Contributor Copy
Role of Network Biology in Cancer Research
345
[72] Mishra GR, Suresh M, Kumaran K. et al. (2006). Human protein
reference database--2006 update. Nucleic Acids Res. 34(Database
issue):D411–D414. doi:10.1093/nar/gkj141.
[73] Stark C, Breitkreutz BJ, Reguly T, Boucher L, Breitkreutz A, Tyers
M. (2006). BioGRID: a general repository for interaction datasets.
Nucleic Acids Res. 34(Database issue):D535–D539. doi:10.1093/
nar/gkj109.
[74] Cowley MJ, Pinese M, Kassahn KS. et al. (2012). PINA v2.0: mining
interactome modules. Nucleic Acids Res. 40(Database issue):D862–
D865. doi:10.1093/nar/gkr967.
[75] Patil A, Nakai K, Nakamura H. (2011). HitPredict: a database of
quality assessed protein-protein interactions in nine species. Nucleic
Acids Res. 39(Database issue):D744–D749. doi:10.1093/nar/gkq897.
[76] Gioutlakis A, Klapa MI, Moschonas NK. (2017). PICKLE 2.0: A
human protein-protein interaction meta-database employing data
integration via genetic information ontology. PLoS One.
12(10):e0186039. doi:10.1371/journal.pone.0186039.
[77] Szklarczyk D, Morris JH, Cook H. et al. (2017). The STRING database
in 2017: quality-controlled protein-protein association networks, made
broadly accessible. Nucleic Acids Res. 45(D1):D362–D368.
doi:10.1093/nar/gkw937.
[78] Prieto C, De Las Rivas J. (2006). APID: Agile Protein Interaction
DataAnalyzer. Nucleic Acids Res. 34(Web Server issue):W298–
W302. doi:10.1093/nar/gkl128.
[79] Shannon P, Markiel A, Ozier O. et al. (2003). Cytoscape: a software
environment for integrated models of biomolecular interaction
networks. Genome Res. 13(11):2498–2504. doi:10.1101/gr.1239303.
[80] Hooper SD, Bork P. (2005). Medusa: a simple tool for interaction
graph analysis. Bioinformatics. 21(24):4432–4433. doi:10.1093/
bioinformatics/bti696.
[81] Bastian M, Heymann S, Jacomy M. (2009). Gephi: an open source
software for exploring and manipulating networks. International AAAI
Conference on Weblogs and Social Media.
Complimentary Contributor Copy
Rinki Singh and Anup Som
346
[82] Theocharidis A, van Dongen S, Enright AJ, Freeman TC. (2009).
Network visualization and analysis of gene expression data using
BioLayout Express (3D). Nat Protoc. 4(10):1535–1550.
doi:10.1038/nprot.2009.177.
[83] Bader GD, Hogue CW. (2003). An automated method for finding
molecular complexes in large protein interaction networks. BMC
Bioinformatics. 4:2. doi:10.1186/1471-2105-4-2.
[84] Salavert F, García-Alonso L, Sánchez R. et al. (2016). Web-based
network analysis and visualization using CellMaps. Bioinformatics.
32(19):3041–3043. doi:10.1093/bioinformatics/btw332.
[85] Liu ZP, Wu C, Miao H, Wu H. (2015). RegNetwork: an integrated
database of transcriptional and post-transcriptional regulatory
networks in human and mouse. Database (Oxford). 2015:bav095.
doi:10.1093/database/bav095.
[86] Zhang M, Li Q, Yu D. et al. (2019). GeNeCK: a web server for gene
network construction and visualization. BMC Bioinformatics.
20(1):12. doi:10.1186/s12859-018-2560-0.
[87] Hu Z, Mellor J, Wu J, DeLisi C. (2004). VisANT: an online
visualization and analysis tool for biological interaction data. BMC
Bioinformatics. 5:17. doi:10.1186/1471-2105-5-17.
[88] Kuo TC, Tian TF, Tseng YJ. (2013). 3Omics: a web-based systems
biology tool for analysis, integration and visualization of human
transcriptomic, proteomic and metabolomic data. BMC Syst Biol. 7:64.
doi:10.1186/1752-0509-7-64.
[89] Tzfadia O, Diels T, De Meyer S, Vandepoele K, Aharoni A, Van de
Peer Y. (2016). CoExpNetViz: Comparative Co-Expression Networks
Construction and Visualization Tool. Front Plant Sci. 6:1194.
doi:10.3389/fpls.2015.01194.
[90] Xia J, Gill EE, Hancock RE. (2015). NetworkAnalyst for statistical,
visual and network-based meta-analysis of gene expression data. Nat
Protoc. 10(6):823–844. doi:10.1038/nprot.2015.052.
Complimentary Contributor Copy
