Network Pharmacology: Exploring the Resources and Methodologies

Abstract

Multi-target and combinatorial therapies have been focused for the past several decades. These approaches achieved considerable therapeutic efficacy by modulating the activities of the targets in complex diseases such as HIV-1 infection, cancer and diabetes disease. Most of the diseases cannot be treated efficiently in terms of single gene target, because it involves the cessation of the coordinated function of distinct gene groups. Most of the cellular components work efficiently by interacting with other cellular components and all these interactions together represent interactome. This interconnectivity shows that a defect in a single gene may not be restricted to the gene product itself, but may spread along the network. So, drug development must be based on the network-based perspective of disease mechanisms. Many systematic diseases like neurodegenerative disorders, cancer and cardiovascular cannot be treated efficiently by the single gene target strategy because these diseases involve the complex biological machinery. In clinical trials, many mono-therapies have been found to be less effective. In mono-therapies, the long term treatment, for the systematic diseases make the diseases able to acquired resistance because of the disease nature of the natural evolution of feedback loop and pathway redundancy. Multi-target drugs might be more efficient. Multi-target therapeutics might be less vulnerable because of the inability of the biological system to resist multiple actions. In this study, we will overview the recent advances in the development of methodologies for the identification of drug target interaction and its application in the poly-pharmacology profile of the drug.

Full text

Send Orders for Reprints to reprints@benthamscience.ae
Current Topics in Medicinal Chemistry, 2018, 18, 000-000 1
REVIEW ARTICLE
1568-0266/18 $58.00+.00 © 2018 Bentham Science Publishers
Network Pharmacology: Exploring the Resources and Methodologies
Junaid Muhammad#, Abbas Khan#, Arif Ali, Li Fang, Wang Yanjing, Qin Xu and Dong-Qing Wei*
Department of Bioinformatics and Biostatistics, School of Life Sciences and Biotechnology, Shanghai Jiao Tong
University, Shanghai, China
A R T I C L E H I S T O R Y
Received: December 28, 2017
Revised: March 08, 2018
Accepted: March 14, 2018
DOI:
10.2174/1568026618666180330141351
Abstract: Multi-target and combinatorial therapies have been focused for the past several decades.
These approaches achieved considerable therapeutic efficacy by modulating the activities of the targets
in complex diseases such as HIV-1 infection, cancer and diabetes disease. Most of the diseases cannot
be treated efficiently in terms of single gene target, because it involves the cessation of the coordinated
function of distinct gene groups. Most of the cellular components work efficiently by interacting with
other cellular components and all these interactions together represent interactome. This interconnectiv-
ity shows that a defect in a single gene may not be restricted to the gene product itself, but may spread
along the network. So, drug development must be based on the network-based perspectiv e of disease
mechanisms. Many systematic diseases like neurodegenerative disorders, cancer and cardiovascular
cannot be treated efficiently by the single gene target strategy because these diseases involve the com-
plex biological machinery. In clinical trials, many mono-therapies have b een found to be less effective.
In mono-therapies, the long term treatment, for the systematic diseases make the diseases able to ac-
quired resistance because of the disease nature of the natural evolution of feedback loop and pathway
redundancy. Multi-target drugs might be more efficient. Multi-target therapeutics might be less vulner-
able because of the inability of the biological system to resist multiple actions. In this study, we will
overview the recent advances in the development of methodologies for the identification of drug target
interaction and its application in the poly-pharmacology profile of the drug.
Keywords: Network pharmacology, multi-target virtual screening, drug-target interaction, chemical-protein interactome, tradi-
tional chinese medicine, side effect similarity
1. INTRODUCTION
An empirical drug discovery was the earliest of the drug
designing processes which was totally based on phenotype
but no molecular mechanism of the drug action was under-
stood. The reductionists tried to understand the interacting
mechanism between the drug and target [1]. These efforts
revealed the idea like lock and key model, in which a spe-
cific drug interacts with a specific single target. Targeting
the single targ et by a single molecule is associated with
strong side effects. A continuous decline in the success of
developing drug candidates is being observed over many
years. Reasons were sought out for these failures which were
neither environmental nor scientific but philosophical. It was
a philosophy of “magic Bullet” which explained “one gene,
one drug, one disease” and supported by the molecular bi-
ologist by characterizing an individual disease causing gene
[1, 2]. To validate the fault in the idea of “single drug, single
target and single disease”, many single knockouts were also
carried out which reported no or very little effect [3-5].
These experimental reports were also augmented by the gene
deletion experiment which also revealed the flaws of a single
*Address correspondence to this author at the State Key Laboratory of Mi-
crobial Metabolism, College of Life Sciences and Biotechnology, Shanghai
Jiao Tong University, Shanghai, China, Email: dqwei@sjtu.edu.cn
#Junaid Muhammad and Abbas Khan contributed equally to this work.
target. However, recent advancement in the phenotype and
reported that approximately 19% of the genes are required to
be knocked out to have the effect understood more about
targeting multiple targets by a single drug that could lead to
the desired effect. The processes of all living organisms
make the combined effect on the internally connected net-
works and any abnormality in a single enity would not bring
any disturbance but depends on the components with which
the gene and products interact. The same hypothesis can be
applied to a diseases case, which is not a consequence of a
single entity but a result of many interconnected patho bio-
logical networks [6]. Kun Yang et al. developed a model for
finding multiple target optimal intervention (MTOI) solu-
tions in a disease network [7]. MTOI can identify the poten-
tial target proteins and suggests optimal combinations of the
target intervention that best restores the network into the
normal state. To bring new drug candidates into the market
with effectiveness, the network-based approach can provide
a better understanding with numerous clinical applications.
The knowledge of the intracellular interaction of different
components will help to identify the disease gene and in turn
the disease pathway. This understanding of the disease net-
work will provide a way to address the complexity of the
human disease and will provide new drug candidates to the
clinics [8]. The intra-link of genes and their association with
different diseases can be easily explored through network

2 Current Topics in Medici nal Chemistry, 2018, Vol. 18, No. 00 Muhammad et al.
biology. System biology or network biology explains how a
cellular network leads to phenotypic or disease state. For
dealing with a complex disease, such as cancer, which is not
due to a mutation in a single gene but dysfunction of the
whole network , network medicine or network pharmacology
is a new emerged task which is targeting all the important
nodes in such a complex state. Therefore, targeting all the
networks by a single drug is the basic principle of network
pharmacology [9-11]. So, network pharmacology intents to
comprehend the systematic level of the disease and interac-
tion of the drug with the body through biological networks.
Network pharmacology also plays a key role in predict-
ing drug-drug combinations and in treating a co mplex dis-
ease. To overcome the drug resistance, drug combination
represents a promising strategy. Several computational
methods were developed to predict the synergistic effect of
drugs. Xing Chen et al. developed an algorithm termed as
Network-based Laplacian that regularized Least Square Syn-
ergistic and predicted 13 synergistic drug combinations
against Candida Albicans infection [12]. Out of 13 drug
combinations, 7 combinations were confirmed experimen-
tally. Xing Chen et al. also developed a crosstalk model to
reduce the emergence of drug resistance that targets signal-
ing pathways [13].
2. CONCEPT AND SIGNIFICANCE OF NETWORK
PHARMACOLOGY
The efforts of molecular biology and genomics research
have provided large data which helped in gaining new in-
sights into drug discovery processes. Hopkin, the father of
Network Pharmacology, explained that a single drug can
target multiple nodes in the disease network [14]. Network
pharmacology is based on the integration of multiple disci-
plinary concepts including molecular biology, biochemical
biology and bioinformatics [14-16]. Network pharmacology
has gained more interest due to high success rate in clinical
investigation, less or affordable side effects, enhanced drug
efficacy, regulation of the signaling pathway with multiple
channels, interaction of multiple genes and proteins that
could be easily be targeted causing the disease [17]. In addi-
tion, network pharmacology also helps in finding the disease
node which is an important disease node. Beside these, it
also increases the clinical candidates with potency and re-
duces the attrition rate in the disease network [18]. Around
40% of the current drug discoveries are contributed by net-
work pharmacology rather than a magic bullet philosophy
[17, 18]. Table 1 is depicting the key concepts developed in
the area of network pharmacology.
3. RESEARCH APPROACHES AND AVAILABLE
DATABASE RESOURCES
A newly emerged area in the field of drug discovery is
network pharmacology which uses mainly two approaches,
establishing a network and utilization of public databases.
Prediction of drug target disease network using HTS tech-
nology in combination with bioinformatics is among the
other approaches in this area [19]. In the area of network
pharmacology, the approaches could be divided into compu-
tational and experimental approaches. The computational
approaches mainly include graph theory, statistical methods,
data mining, modeling, and information visualization meth-
ods. The experimental approaches include various high-
throughput omics technologies and biological and pharma-
cological experiments. In network pharmacology, some
common steps include data sources, big data analytics, net-
work construction, interactions prediction and network
analysis.
3.1. Data sources
Experimental verification and public databases are the
two main sources of data collection in network pharmacol-
ogy. By utilizing the existing research and available data, a
target can be identified for the drug followed by an experi-
mental validation. Another approach to collect data is omics
technology [26, 27]. The available databases and resources
are summarized below;
3.1.1. DrugBank
(http://www.drugbank.ca) [28]
The DrugBank database is an abundantly interpreted bio-
informatics and cheminformatics resource. DrugBank com-
bines multi array information regarding the drug candidates.
This information largely comes from pharmaceutical, chemi-
cal and pharmacological sources along with target informa-
tion. Statistics of this database reported 7759 drug entities till
now and 15,199 drug–target interactions.
Table 1. The key concepts of netwo rk pharmacology which are chronologically developed through the era of TCM netwo rk
pharmacology.
Category & Term
Description
Year
References
Hypothesis of the relationsh ip between TCM Syndrome and molecular networks
1999
[20]
Proposed a network-based TCM research framework related to TCM network pharma-
cology
2007
[21]
A network-based case study on Cold/Hot herbal formulae and Hot/Cold Syndromes
2007
[22]
Proposed the “Herb network-Biological network Phenotype network”
2009
[23]
Concepts in TCM network
pharmacology
Proposed the new conc ept of “Network target”
2011
[24, 25]

Network Pharmacology: Exploring the Resources and Methodologies Current Topics in Medicinal Chemistry, 2018, Vol. 18, No. 00 3
3.1.2. TTD: Therapeutic Target Database
(http://bidd.nus.edu.sg/group/ttd/ttd.asp) [29]
Therapeutic Target Database (TTD) provides the infor-
mation about some known and different aspects of a disease.
This information largely includes proteins which are thera-
peutically significant, nucleic acid (DNA & RNA) targets,
disease specific characterization, pathway information and
consistent drugs acting on different targets. Like other data-
bases, TTD also holds 1755 biomarkers for about 365 disor-
ders and 210 scaffolds. These 210 scaffolds are of around
714 drugs. TTD is also enriched with a variety of lead com-
pounds. Targets and drugs included in TTD are of great
clinical importance, under use and trials. These targets and
drugs are found to be very useful in accelerating the process
of modern in silico drug discovery and experiments.
3.1.3. MATADOR
(http://matador.embl.de) [30]
To obtain the information regarding multiple direct and
indirect modes of drug–target interactions and protein–
chemical interactions, Manually Annotated Targets and
Drugs Online Resource (MATADOR) is a frequently ac-
cessed database. Direct and indirect binding of proteins and
chemicals could be accessed by searching a drug or a pro-
tein.
3.1.4. ChEMBL
(https://www.ebi.ac.uk/chembldb) [31]
ChEMBL contains drug candidates with information re-
garding ADMET and binding. Regular literature mining
helped in the collection of this large data. Currently, 5.4 mil-
lion bioactive candidates are added to the database. These
information and candidates are used in the processes of drug
discovery.
3.1.5. STITCH
(http://stitch.embl.de/) [32]
STITCH is a database which incorporates data from in vi-
tro results, literature mining and other resourced databases to
show the information regarding the known and predicted
chemical–protein interactions. Till now 45% increase in the
chemical-protein interaction data in the recent version of
STITCH has been reported.
3.1.6. SuperTarget
(http://bioinf-apache.charite.de/supertarget_v2/) [33]
SuperTarget is a wide-ranging databank for analyzing
drug–target interactions. To date, the database is resourced
with 332 828 drug–target interactions. Rather than the gen-
eral query about the drug-targets and drugs, it also provides
information regarding cytochromes P450s.
3.1.7. TDR Targets
(http://tdrtargets.org/) [34]
The TDR Target Database is a chemo genomics databank
resourced with information regarding neglected tropical dis-
eases. The purpose of TDR is to id entify and prioritize drug-
targets and drugs for neglected disease agents. Functional
genomics data, such as phylogenic reconstruction, differen-
tial expression and essentiality of the disease causing agents
such as bacteria, virus, fungi etc. for genes could be availed
using TDR.
3.1.8. PDTD
(http://www.dddc.ac.cn/pdtd/) [35]
PDTD (Potential Drug Target Database) is a dual-
function resourceful database. PDTD integrates information
from informatics database and a structural database of
known and potential drug targ ets. Known 3D structures tar-
gets are mainly integrated and categorized into 15 and 13
types following therapeutic and biochemical criteria as a
standard of division.
3.1.9. Integrity
(http://integrity.thomson-pharma.com) [36]
This database covers a large number of clinical drug can-
didates corresponding to their drug targets, diseases and the
statistics on clinical phases of the drugs.
3.1.10. FAERS
(http://www.fda.gov/Drugs/GuidanceComplianceRegulator
yInforma-
tion/Surveillance/AdverseDrugEffects/default.htm)
The FDA Adverse Event Reporting System (FAERS) is a
database that contains information obtained from an adverse
event and medication error reports submitted to FDA on side
effect keywords (adverse event keywords) for drugs.
3.1.11. SIDER
(http://sideeffects.embl.de/) [37]
SIDER database collects information regarding the side
effects (i-e frequency) of already approved drug candidates.
Classifications, linking to further information such as drug–
target associations, are also one of the major aims.
3.1.12. JAPIC
(http://www.japic.or.jp/)
Japan Pharmaceutical Information Center (JAPIC) covers
the information regarding pharmaceutical circle in Japan.
Information such as side effects for drugs (pharmaceutical
molecules) is mainly added.
3.1.13. ChemBank
(http://chembank.broadinstitute.org/) [38]
ChemBank is a freely accessed database resourced with
information about small molecules so that insights can be
gained. ChemBank is unique among small-molecule data-
bases in the following three ways: its holds a large space for
raw screening data storage, having rigorous definition of
screening experiments in terms of statistical hypothesis test-
ing and hierarchical metadata-based organization of related
assays into screening projects.
3.1.14. CancerDR
(http://crdd.osdd.net/raghava/cancerdr/) [39]
The CancerDR offers information of 148 anti-cancerous
agents, and their pharmacological profiling across 952 can-

4 Current Topics in Medici nal Chemistry, 2018, Vol. 18, No. 00 Muhammad et al.
cer cell lines. Comprehensive information, such as 1356
unique mutations, gene ontology, pathways, and phylogeny
about the drug targets, is available in this database. The
design of an effective and personalized cancer treatment and
the identification of genes encoding drug targets could be
easily mapped from this database.
3.1.15. BindingDB
(http://www.bindingdb.org/bind) [40]
The BindingDB has information about 1,132,739 ex-
perimentally measured protein-ligand binding affinities.
Among these, 4,894,16 are small molecule such as ligands
while 7020 (receptors) are protein targets. It has become one
of the most extensive public databases of protein–ligand
binding affinities.
3.1.16. ZINC
(http://zinc.docking.org) [41]
ZINC is the largest database for ligand discovery, espe-
cially investigating novel drug candidates for biological tar-
gets. ZINC contains >20 million commercially available
compounds for ligand discovery and virtual screening.
3.1.17. canSAR
(https://cansar.icr.ac.uk) [42]
canSAR is a cancer research database information about
biological data (annotations of biological data, screening of
RNA interference and chemical agents, expression and 3D
structural). The integration of this diverse data set aids in
cancer research and discovery of drug candidates for the
treatment of various cancers.
3.1.18. ASDCD
(http://asdcd.amss.ac.cn/) [43]
DCDB is a database which holds information of antifun-
gal drug research in order to help in drug combination analy-
sis and new antifungal drug development. To date, 210 anti-
fungal drug combinations and 1225 drug–target interactions
involving 105 individual drugs from >12 000 references
have been resourced.
3.1.19. DINIES
(http://www.genome.jp/tools/dinies/) [44]
Drug–target interaction network inference engine is
based on a supervised analysis. DINIES, is a web server to
infer potential drug–target interaction network. DINIES can
accept flexible input data, such as chemical structure, side
effects, amino acid and protein domains. Furthermore, each
data set can be transformed into a kernel similarity, and vari-
ous state-of-the-art machine learning methods are used to
realize the drug–target interactions prediction.
3.1.20. SuperPred
(http://prediction.charite.de/) [45]
Anatomical Therapeutic Chemical (ATC) code and drugs
targets are predicted by the online server SuperPred. For
ATC code prediction, different criteria such as pipeline
search could be used for the integration of 3D, 2D and frag-
ment similarity. Drug target prediction is based on the simi-
larity distribution, wh ich can estimate individual thresholds
and probabilities for a specific target by four input options.
3.1.21. SwissTargetPrediction
(http://www.swisstargetprediction.ch/) [46]
SwissTargetPrediction is a web server to deduce the tar-
gets of bioactive small compounds based on the combination
of 2D and 3D similarity values with the known ligands. Five
different organisms, including Homo sapiens, Mus muscu-
lus, Rattus norvegicus, Bos taurus and Equus caballus can be
inquired using SwissTargetPrediction.
3.2. Big Data Analytics
For large and complex networks, the traditional ap-
proaches may not be sufficient to fully understand the dis-
ease network. Therefore, highly analytical techniques such as
high-performance data mining, predictive analytics, text
mining, forecasting and optimization are required to unveil
the hidden information. In addition, machine learning could
be useful to addressing other needs [47, 48].
3.3. Network Construction and Interactions Prediction
Understanding the network of the disease is the most im-
portant step of the network pharmacology. How to construct
a network disease is another complicated aspect of this
analysis but certain approaches have been made to under-
stand and exploit it for new drug candidates [49]. Some
known approaches are: gene locality [50], phylogenetic re-
construction [49], fusion of genes [51], correlated evolution-
ary rate [52], mirror tree [51], correlated mutations [53], ho-
mologous structural complexes [54] and prediction from
primary structure [55]. Network construction and their inter-
action can be significantly done by using phylogenetic pro-
filing. Node-based network mapping and as well as correla-
tion-based is considered as the promising for future discover-
ies [56].
3.4. Network Analysis
Network is a well-computed mathematical representation
of various connected nodes and edges. A major portion of
the network pharmacology is network analysis which mainly
covers attribute analysis, topological analysis, network struc-
ture and stability, flow (flux) balance analysis and network
models. A network analysis usually measures module,
betweenness, hub, node, edge, shortest path and degree of
hub gene. Fig. (1) shows the topological parameters of a
network.
• Module: A group of nodes that act in concert to perform
a specific function.
• Hub: A node w ith high degree.
• Degree: the number of edges connected to a node.
• Betweenness: the number of shortest paths that go
through a giv en specific node.
• Shortest path: A minimum path between any two nodes
in a network

Network Pharmacology: Exploring the Resources and Methodologies Current Topics in Medicinal Chemistry, 2018, Vol. 18, No. 00 5
3.5. Methods in Network Pharmacology
3.5.1. Identification of Drug Target Interaction
In genomic drug discovery, the identification of drug tar-
get interaction is considered as a key area of interest. The
interactions of small molecules with different pharmaceuti-
cally important protein targets modulate its activity. The
application of various biological assays for the high through-
put screening of large chemical databases enabled the identi-
fication of drugs with different targets [57-59]. Chemical
genomic research aimed to relate the chemical spaces with
genomic spaces, however, the relationship of chemical and
genomic is very limited. For example, the PubChem data-
base has information about millions of compounds but in-
formation about the interactions with their targets is very
limited [60]. The exper imental determination of compound-
protein interactions or potential drug–target interactions is
time-consuming and cost-effective [61, 62]. So, an effective
in silico prediction method needs to be developed.
3.5.2. Prediction of Drug-target Interaction Networks Via
Chemical and Genomic Spaces
In 2008 Yaminishi et al. proposed three methods based
on chemical and genomics spaces [63]. They obtained the
drug target interactions from the SuperTarget, KEGG
BRITE, DrugBank databases and BRENDA [30, 58, 64, 65].
The information about chemical data was obtained from the
KEGG LIGAND database. The structure similarity was
computed by the SIMCOMP methods [66]. The methods
proposed are the nearest profile method, weighted profile
method and bipartite graph learning method. Previously two
research approaches have been used for the identification of
drug-target interactions, the chemical biology and the tradi-
tional drug discovery approach. In traditional drug discov-
ery, new lead compounds are identified for the few targets.
In chemical biology, novel targets are identified for the few
chemical compounds. The methods proposed by Yaminishi
et al. are advantageous for both of the above mentioned ap-
proaches. Among these methods, Bipartite graph learning
method has the advantage to predict the interaction for pre-
viously unseen drug candidate compounds and target candi-
date proteins [63] while other methods, the nearest profile
and weighted profile methods cannot predict the interaction
for the previously unseen drug candidate compounds and
target candidate proteins. Th e nearest profile method predicts
the interaction based on the structure sequence similarity and
hence may give false positive results. Because many target
candidates such as enzymes share sequence similarity but
bind to different chemicals. Some other methods such as
docking simulation can predict the interaction but it needs
three-dimensional structures of the target protein candidates
[67, 68]. Many of the pharmaceutically important drug tar-
gets are GPCRs and ion channels. Predicting the three-
dimensional structure of these proteins is a challenging task,
hence, it limits the molecular docking approach to predict the
drug target interaction. The Bipartite graph learning method
does not need three dimension al structures. Therefore, bipar-
tite graph learning method has an advantage that it is suitable
for screening a huge number of drug candidate compounds
and target proteins at a large scale.
3.5.3. Prediction of Drug–target Interaction Networks
Through Side Effect Similarity
The treatment of human disease with selected drugs re-
sults in regulated recording of side effects. These side effects
are directly attributed to the interaction of drugs with pri-
mary targets and off targets (additional target) and seem to
be one of the most important scenarios [69-71]. The interac-
tion of drugs with off-target derives unexpected and harmful
results. But, sometimes these interactions have a beneficial
effect and lead to a new therapeutic area for drugs [72]. For
example, sildenafil was used to treat angina, but its side ef-
fect in human volunteers prolonged penile erections, which
led to a new therapeutic area for sildenafil [72] . Monica
Fig. (1). The figure is describing a topology of a network. It includes module, Betweenn ess, Hub, node, edge, shortest path and Degree of
hub gene

6 Current Topics in Medici nal Chemistry, 2018, Vol. 18, No. 00 Muhammad et al.
Campillos et al. mentioned that unrelated drugs that share
similar side effects, must have common off-targets [73]. For
example, the two unrelated drugs, cisapride and astemizole,
bind to the cardiac ion channel hERG, thus inhibit its activity
and both cause cardiac arrhythmias [74]. Monica Campillos
et al. take advantage of the side effect and developed a
method for the side effect similarity and analysis, the likeli-
hood of sharing the target of marketed available drugs [73].
Through in-vitro binding assays, they confirmed experimen-
tally that the side effect similarity of unrelated drugs indeed
shares a common protein target. Through the application of
side effect similarity, Monica Campillos et al. suggest new
targets for marketed drugs of different therapeutic categories
(Table 2). The new targets were found experimentally to
bind with drug candidates with good binding affinity. Side
effects open a new dimensional space for predicting the
polypharmacology of the drug [75]. Feixiong Cheng et al.
developed a database for predicting the side effects named
MetaADEDB [76]. Taking the advantage of the side effects,
Feixiong Cheng et al. found the network pharmacology of
the drugs and found new potential targets (Table 3) [77].
Table 2. Experimentally validated off-targets for the mar-
keted drugs through side effect similarity method.
Drug
Off-target
Ki (µM)
Donepezil
5HTT
9
Fluoxetine
dopamine receptor (DRD3)
2
Rabeprazole
serotonin receptor (HTR1D)
7.6
Rabeprazole
dopamine receptor (DRD3)
1.6
Paroxetine
dopamine receptor (DRD3)
3.8
Zaleplon
HRH1
26
Disopyramide
HRH1
2.7
Clomiphene
HRH1
6.5
Loratodine
BZRP
5
Raloxifene
Seroton in receptor (HTR1D)
0.3
Acitretin
HRH1
15
Doxorubicin
HRH1
10
Ketoconazole
serotonin receptor (HTR1D)
2.8
3.5.4. Prediction of Drug–target Interaction Networks by
Integrating the Pharmacological Space into Chemical and
Genomic Spaces
The in-silico prediction of the drug target interaction
from heterogeneous biological data is important to discover
the drugs and target candidates for the known disease. The
chemical genomics has made it possible to relate the chemi-
cal space with g enomic space, but genome wide detection of
drug target interaction is the key issue in chemical genomic
research [57-59]. Thus, in 2010, Yamanishi et al. proposed a
new method that relates the chemical space with the pharma-
cological space and the integration of drug target network
topology [78]. They showed that the drug–target interaction
is mostly correlated with a pharmacological effect similarity
than with chemical structure similarity. Owing to the pro-
posed method, the unknown drug target interactions are pre-
dict at a large scale from the information of genomic se-
quence, chemical structure and pharmacological effect. The
method consists of two steps: (1) inference of the pharma-
cological information from the structure of a given com-
pound via an algorithm developed by Scheiber et al. [79] (2)
prediction of the interaction between drug and target based
on the pharmacological effect similarity in the supervision of
bipartite graph inference [63, 80]. In fact, the proposed
method here is the extension of the work by Yamanishi et al.
published in 2008 [63]. The performance of the proposed
methods was evaluated for the four different classes of pro-
teins to reconstruct the drug target interaction in terms of
three inputs (i) similarity of the chemical structure (ii) true
pharmacological similarity, and (iii) predicted pharmacol-
ogical similarity. The four different classes include ion
channels, enzymes, nuclear receptors and GPCRs. The statis-
tics of the proposed method is summarized in Table 4. In
Table 4, the input, and chemical structure similarity are
based on the previous method [63] while the input, and true
pharmacological similarity and predicted pharmacological
similarity are based on the proposed method. The previous
study in the same area uses side effect similarity, but the
method is only applicable to the marketed information avail-
able of drugs with side effects [73]. Thus, the method pro-
posed by Yamanishi et al. in 2010 is able to predict the
pharmacological information about not only the marketed
drugs but also any drug candidate.
3.5.5. Prediction of Drug–target Interaction Via Chemical-
protein Interactome (CPI)
Approximately 90 percent of the drug candidates fail dur-
ing the different developing phases before launching into the
market. It makes the research and developing process ex-
tremely expensive and time-consuming. The identification of
novel indication for the already available marketed drug
might lower the research and development costs [72, 81].
The de novo development of a drug takes approximately
10-17 years with regulatory, efficacy and quality risk. The
repurposing of the drugs has the advantage of decreased re-
search and developing cost with launching time due to the
previously collected pharmacokinetic, toxicology and safety
data. The adverse side effects of the drug have been known
as the leading cause of death of hospitalized patients and
have been concerned word-widely [82, 83]. These new indi-
cations and adverse side effects are caused by unwanted
drug-protein interactions [84-91]. The prediction of this in-
teraction is possible by mining the chemical-protein interac-
tome (CPI) [92]. Several other techniques such as drug affin-
ity pull-down and BIACORE biosensors can be used to pre-
dict unwanted or unexpected chemical-protein interactions
[93, 94] but CPI has an advantage of low cost. The first CPI
released by the Lun Yang et al. contains 162 chemicals and
891 binding pockets [92]. The chemicals selected in the CPI
are FDA approved drugs, each of which causes at least one
of the major serious adverse drug reactions (SADRs) includ-
ing deafness, cholestasis, Stevens-Johnson syndrome (SJS)
and rhabdomyolysis. As the human knowledge about SADRs

Network Pharmacology: Exploring the Resources and Methodologies Current Topics in Medicinal Chemistry, 2018, Vol. 18, No. 00 7
Table 3. Polypharmacological profile of the approved drugs.
Drug Name
Primary Target
Predicted Target
IC50 µM
Dobutamine
Beta-1 Adrenergic Receptors
Adrenoceptor alpha 2A (ADRA2A)
10.83
Fenoterol
Beta-2 Adrenergic Receptors
5-hydroxytryptamine receptor 2A (HTR2A)
3.40
Ketotifen
Histamine H1 Receptors
Adrenoceptor alpha 1A (ADRA1A)
10.40
Loxapine
Dopamine Receptor D2, Serotonin Receptor
2c and 7
Muscarinic acetylcholine receptor M2 (CHRM2)
1.12
Tramadol
Opiate Receptors and alpha(2)-adrenergic
receptors
Muscarinic acetylcholine receptor M1 (CHRM1)
11.08
Pimozide
Dopamine D2 receptor
Adrenoceptor alpha 1A (ADRA1A)
0.21
Sertraline
Seroton in Transporter
Histamine H1 Receptors (HRH1)
23.00
Table 4. Statics of the proposed method of Yaminishi et al.
Inputs
Class
Statistics
Chemical structure
similarity
True pharmacological similarity
Predicted phar macol-
ogical similarity
Enzyme
AUC
Sensitivity
Specificity
PPV
0.821
0.239
0.993
0.358
0.892
0.356
0.995
0.527
0.845
0.245
0.993
0.369
Ion channel
AUC
Sensitivity
Specificity
PPV
0.692
0.134
0.996
0.704
0.812
0.137
0.996
0.714
0.731
0.142
0.997
0.742
GPCRs
AUC
Sensitivity
Specificity
PPV
0.811
0.147
0.994
0.519
0.827
0.172
0.996
0.614
0.812
0.164
0.995
0.581
Nuclear
receptors
AUC
Sensitivity
Specificity
PPV
0.814
0.067
0.995
0.560
0.835
0.057
0.994
0.480
0.830
0.077
0.996
0.640
is limited, the target proteins in the CPI were selected from
the literature and protein targetable databases [35, 95-97].
Through the application of CPI, Lun Yang et al. harvested
the genes responsible for the SJS [92]. The CPI has the ad-
vantage of predicting the specific alleles that is more sensi-
tive to the drug attack. HLA-B*57 has been conformed as
the susceptible gene of SADRs causing hypersensitive reac-
tion in response to abacavir [98, 99]. The structure of both
the risk and non-risk allele of HLA-B*57 is availab le [100,
101]. Lun Yang et al. construct the CPI, containing interac-
tion strength for the four structures of risk and non-risk allele
with abacavir, allopurinol. The author found no specificity of
allopurinol to any of the proteins. This result is in accor-
dance with the fact that none of these alleles are the risk al-
leles for allopurinol-induced SADRs (Table 5). It’s clear
from Table 5 that B*5703 is not the susceptible allele be-
cause abacavir cannot fit into the binding site of B*5703.
While the allele B*5701is found to be the risk allele. The
major difference between the two alleles lies in two poly-
morphisms (N114D, Y116S) from B*5703 to B*5701.
Through CPI, it is deduced that B*5701 tends to be the risk
allele compared to B*5703.
Taking the advantage of CPI, Heng Luo et al. introduced
a web server named DRAR-CPI [102]. The server contains
385 human targetable proteins and 254 molecules with
descriptions, indications and ADRs. The server accepts
molecules in mol, ml2, mol2, pdb, sdf and SMILES. Dock
programs implemented in the server are used to predict the
binding energy of the submitted molecule and targets. The
author developed an algorithm based on connectivity analyt-
ics [103] which calculate the positiv e or negative association
scores between the submitted drug and the server molecules.
The two-directional Z-transformation (2DIZ) is applied to

8 Current Topics in Medici nal Chemistry, 2018, Vol. 18, No. 00 Muhammad et al.
association scores [104]. The target having association score
less than 1 is treated as a favorable target while greater than
1 is treated as unfavorable. It can also predict the off-targets
for the submitted molecule so that users can predict potential
indications or ADRs based on the association scores of their
molecule across our library molecules. The reliability of the
server was checked by comparing the predicted drug-drug
associations and drug-drug association through gene-
expression profiles. The matching rate was found to be 74%.
Heng Luo et al. found a new indication and ADRs for the
Rosiglitazone, a drug used as an anti-diabetic through the
application of DRAR-CPI server [102, 105]. Several studies
have been published, using DRAR-CPI server, regarding the
discovery of the new indication and ADRs for the different
drugs [106-114].
Owing to the complex network-based theory [115-118],
Feixiong Cheng et al. proposed the three methods named
target-based similarity inference (TBSI), network-based in-
ference (NBI) and drug-based similarity inference (DBSI)
[119]. The performance of these methods is checked with
four benchmark data sets. Four major drug targets were in-
cluded in these data sets named as ion channels, enzymes,
GPCRs, nuclear receptors and GPCRs. After several statisti-
cal analyses, the NBI method was found to be the best.
Based on the NBI method, the drug target interaction of the
FDA approved and experimental drug was determined. New
targets were successfully predicted for the five approved
drugs following the NBI method [119]. The NBI method was
further improved by Feixiong Cheng et al. by weighting the
edge and nodes of the CPI to achieve the better accuracy of
drug target interaction [120]. In the Edge Weighted Net-
work-based Inference (EWNBI), each edge of the CPI is
weighted according to the strength of the inhibitory activity
or binding affinity of chemical and protein node [120]. In
Node Weighted Network-based Inference (NWNBI), a new
expression of initial resource distribution of nodes is used
which takes into account the influence of resources associ-
ated with the receiver nodes in the CPI bipartite network
proposed by Jia et al. [121]. This method is based on the
general knowledge that the hub node with more resources is
more difficult to be influenced. These two improved meth-
ods slightly outperformed the original NBI.
Because of the lack of connections between the newly
synthesized chemical or failed drugs, in phases II and III,
and the existing DTI network, the aforementioned methods
cannot predict the new potential targets for the known drugs
unless the known target present in the existing DTI network.
To overcome this pitfall, in 2016, Zengrui Wu et al. pro-
posed chemoinformatics tool and an integrated network
named substructure–drug–target network-based inference
(SDTNBI) [122]. To bridge the gap between the newly syn-
thesized structure and known drugs, SDTNBI uses a sub-
structure which is shared by the chemical structures. The
chemical substructure has a significant role in the computa-
tional evaluation of drug pharmacokinetics and DTI predic-
tion suggested by the previous studies [123-126]. Thus,
SDTNBI can prioritize potential targets for old drugs, clini-
cally failed drugs, and new synthesized chemicals at a large
scale. However, several pitfalls exist in the SDTNBI, sas
they cannot predict the potential DTIs for the subject targets
that are absent from the existing DTIs because of lack of
connection among those targets and the known network.
Moreover, it cannot predict the accurate DTIs for the new
chemical molecule that shares no substructure or few sub-
structures.
Zengrui Wu et al. have made an improvement in the
original SDTNBI by introducing the three parameters [127]:
(i) the initial resource allocation of different nodes (i.e. sub-
structure nodes and target nodes), (ii) the weighted values of
different edges (i.e. drug-substructure associations and drug-
target interactions), and (iii) the influence of hub nodes, re-
spectively. The improved SDTNBI was named as a balanced
substructure-drug-target network-based inference
(bSDTNBI). Zengrui Wu et al. found the molecular mecha-
nism of action (MoA) of tricyclic anti-depressant agent pro-
methazine and clomipramine via bSDTNBI [127]. Previous
studies suggested that both promethazine and clomipramine
induce cell apoptosis in different cancer cells but the anti-
cancer MoA of these drugs remains unclear [128-131].
Through bSDTNBI, promethazine and clomipramine were
found to target the serotonin receptors (HTR1A and HTR1D)
with high score. These receptors might be involved in differ-
ent cancer types. Moreover, through bSDTNBI, several anti-
diabetics drugs such as pioglitazone, rosiglitazone and dapa-
gliflozin were repurposed for cancer treatment by targeting
the nuclear receptors such as CA1, PPA RG, RARB, and
RXRA [127]. Collectively, bSDTNBI would provide a pow-
erful tool for the identification of chemical MoA in drug
discovery and development.
3.5.6. Prediction of Drug-target Interaction Through a
Network-based Random Walk with Restart on the Hetero-
geneous Network
The drugs with similar structures often interact with simi-
lar proteins. Chen et al. developed a model of Network-
Table 5. Chemical-protein interactome among abacavir, allopurinol and four HLA-B*57 structures.
Abacavir
Allopurinol
PDB ID
Allele
Dock-Score
Z-Score
Risk Allele (yes/no)
Dock-Score
Z-Score
Risk Allele (yes/no)
2BVO
B*5703
-33.73
0.460
no
-26.95
1.77
no
2BVQ
B*5703
-34.07
0.416
no
-28.56
0.141
no
2BVP
B*5703
-32.52
0.618
no
-26.88
1.84
no
2RFX
B*5701
-48.71
-1.49
yes
-28.69
0.00565
no

Network Pharmacology: Exploring the Resources and Methodologies Current Topics in Medicinal Chemistry, 2018, Vol. 18, No. 00 9
based Random Walk with Restart on the Heterogeneous
network (NRWRH ) that effectively predicts the drug target
interaction based on the above assumption and by the inte-
gration of drug-drug similarity network, protein-protein
similarity network and know drug-target interaction net-
works into a heterogeneous network [132]. For the integra-
tion of data and drug–target interactions prediction, NRWRH
makes the full use of the network tool that is different from
the traditional random walk with restart. In case of NRWRH,
the random walk is applied to the heterogeneous network
which consists of different sub-networks such as drug
chemical structure similarity network, and target protein se-
quence similarity network and drug–target interactions net-
work. This method has an advantage of predicting the novel
target for the subject drug which has no known target. The
potential target can be predicted based on the known targets
of drugs, which are similar to given subject drug.
3.5.7. Prediction of Drug –target Interaction Through A
Rotation Forest-based Predictor
Based on the hypothesis that drug target interactions are
mainly determined by the primary structure of the target pro-
tein sequence and substructure fingerprints of drug mole-
cules, Lei Wang et al. proposed a novel method for the drug
target interactions [133]. In the proposed method, the inter-
actions of the drug with the target are predicted under the
theory that each drug-target interaction pair can be repre-
sented by the structural properties of the drugs and evolu-
tionary information derived from proteins. The biological
evolutionary information of the protein sequences is encoded
as Position-Specific Scoring Matrix (PSSM) descriptor. The
drug molecules are encoded as fingerprint feature vectors
which represent the existence of certain functional groups.
First, the protein sequence is converted into the PSSM ma-
trix and then the auto covariance algorithm was used to ex-
tract features from PSSM containing biological evolution
information to combine it with molecular substructure fin-
gerprints information to form a feature vector. At last, the
drug target interaction is predicted by a rotation forest (RF)
classifier. The prediction accuracy of the proposed method
was found to be 71.1%, 84.1%, 89.1% and 91.3% for four
datasets’ nuclear receptors, GPCRs, ion channels and en-
zymes, respectively. Later, in 2016, Yu-An Huang et al. pro-
posed a model based on the same assumption of Lei Wang et
al. [133, 134]. Here the protein sequences were encoded by
the Pseudo Substitution Matrix Representation (Pseudo-
SMR) descriptor due to which the influence of biological
evolutionary information retained. The drug molecules were
represented by the structural activity relationship (SAR). The
extremely randomized trees (ETs) classifier was used instead
of RF classifier to build the model for the four datasets’ nu-
clear receptors, GPCRs, ion channels and enzymes. The pre-
diction accuracy of the Lei Wang et al. model was 81.67%,
82.99%, 87.87% and 89.85% for the four nuclear receptors.
4. NETWORK PHARMACOLOGY AND TRADI-
TIONAL CHINESE HERBAL MEDICINES:
Traditional Chinese medicine (TCM) holds a long history
of development around thousands of years which has ac-
quired clinical significance. It has been identified to have
unique and successful clinical applications. The administra-
tion of TCM herbal formulae is a remarkable feature of the
treatment based on Syndrome (ZHENG in Chinese) differen-
tiation, as well as holistic thinking in TCM theory. In tradi-
tional Chinese medicine network pharmacology, understand-
ing the “Network” is the prerequisite. Here the network is
also a well-computed and mathematical representation of
various connected nodes and edges. The concept of TCM
network pharmacology revealed many methodologies with
itself. Different models were proposed and utilized as given
below in the Table 1 [19]. A series of TCM network phar-
macology methods (Table 6) was created, including the net-
work-based prediction of disease genes [17, 135], drug tar-
gets [136] and drug functions [137, 138], the construction of
disease-specific networks [135, 136, 139], the construction
of herb networks [140], and a drug-gene-disease co-module
quantitative analysis [138, 141, 142]. These methods pat-
ented key procedures [143-145] and data-bases [146] which
provide a bone to support all the network pharmacology re-
search. These approaches have continuously led to (i) the
identification of active ingredien ts and synergistic ingredient
pairs in TCM herbal formulae (Fig. 2) [25, 135, 139] and (ii)
exploration of the network characteristics of the classic the-
ory of TCM herbal formulae, such as Cold or Hot herb prop-
erties [22], and the combinatorial rules of ‘Jun-Chen-Zuo-
Shi’ [141]. Furthermore, these network characteristics can be
exploited to predict clinical biomarkers of TCM h erbal for-
mulae and rationally design multi-component therapeutics.
These methods are explored and discussed by Shao and
ZHANG in 2013 [147]. Shao Li et al. developed different
databases given in Table 7 for TCM network pharmacology
which include HerbBioMap which is a molecular data source
for herbs and TCM phenotypes developed in 2010 [148] as
well as dbNEI which is another database for neuro-
endocrine-immune interactions and drug-NEI-disease net-
work developed in 2006 and 2008 [42, 149].
5. POLYPHARMACOLOGICAL PROfiLES AND
MULTI-TARGET VIRTUAL SCREENING OF THE
DRUGS
In the past history, pharmaceutical companies relied on
specific families of druggable protein targets [162]. The
chemists had made attempt to develop compounds with the
desired action [163, 164]. Owing to the development of net-
work biology, systems biology, and polypharmacology, a
new concept of network pharmacology has gained interest of
many researchers [14, 15, 165, 166]. It has been considered
to be the next standard framework for the drug development.
It aims to investigate the relation of drug to disease and un-
ravel the synergistic effect of multicomponent of the drug on
the multi-targets. It has been suggested that relatively week
interactions with multi-targets may prove more satisfactory
than the strong interaction with a single target [147]. Several
studies regarding polypharmacology are listed below.
Montelukast is effective in the treatment of Asthma and
has been considered a cysteinyl leukotriene 1 receptor an-
tagonist [167]. It is sold in several countries with brand name
Singulair by Merck. Recently, Langlois et al. reported that
Montelukast regulates eosinophil protease activity through a
mechanism of action that is leukotriene-independent [168].
So far, there has been no study regarding its binding with
dipeptidyl peptidase-IV. Feixiong Cheng et al. predicted

10 Current Topics in Medicinal Chemistry, 2018, Vol. 18, No. 00 Muhammad et al.
Table 6. The table is showing the available methodologies in TCM network pharmacology.
Methods
Description
Year
Ref
CIPHER
Network-based prediction for disease genes
2008
[17]
drugCIPHER
Network-based prediction for drug (herbal ingredient) targets and functions
2010
[145, 150]
comCIPHER
Drug–gene-disease co-module analysis
2012
[150]
CIPHERHIT
Modularity-based disease gene prediction
2011
[135]
DMIM
Herb network construction and co-module analysis for herbal formulae
2010
[141]
NADA
Network-based assessment for drug (herbal ingredient) action
2010
[139]
NIMS
Network-based identification of multi- component synergy and dru g (herbal
ingredient) combinations
2011
[25, 140]
Drug combination mode l
A formal model for analyzing drug combination effects
2010
[151]
LMMA
Disease-specific biomolecular network construction
2006
[143]
CSPN
Disease-specific pathway network construction
2010
[137]
ClustEx
Disease-specific responsive gene module identification
2010
[152]
Fig. (2). A general schematic diagram of TCM network ph armacology in the discovery of an herbal formula.
montelukast as a new dipeptidyl peptidase-IV inhibitor
[120]. The author validated it by experiment with an IC50
value 9.79 mM. So, montelukast might have new indication
in anti-diabetic treatment. This new indication is supported
by the study of Faul et al. [169]. The author found that oral
administration of montelukast changed the level of Insulin.
Moreover, the Tanimoto similarity of montelukast and a
classical inhibitor of dipeptidyl peptidase-IV was 0.38 based
on MACCS keys [170], indicating that NBI could success-
fully predict novel structural skeleton molecules for a given
target.
Diclofenac, an acetic acid nonsteroidal anti-inflammatory
drug is widely used to treat pain, dysmenorrhea and ocular
inflammation. The anti-inflammatory eff ects of diclofenac
were thought to be linked with the inhibition of leukocyte
migration and cyclooxygenase (COX-1 and COX-2), leading
to the peripher al inhibition of prostaglandin synthesis [171].
For the first time, Feixiong Cheng et al. predicted new po-
tential targets for the diclofenac through NBI method [119].
The author reported ERα and ERβ as a potential target for
diclofenac. Through experimental assays, the IC50 values
were found to be 7.59 and 2.32 mM, respectively. Few pre-
vious studies support the fact that nonsteroidal anti-
inflammatory drugs can target the nuclear receptors. Leh-
mann JM et al. demonstrated indomethacin, a nonsteroidal
anti-inflammatory drug that activates the Peroxisome Prolif-

Network Pharmacology: Exploring the Resources and Methodologies Current Topics in Medicinal Chemistry, 2018, Vol. 18, No. 00 11
erator-activated Receptors α and γ at micro-molecular con-
centration [172][. Zhou et al. reported that Sulindac sulfide
binds to the retinoid X receptor α with an IC50 value of 80
µM and induces retinoid X receptor α-depended apoptosis
[173]. These results demonstrated that diclofenac has an
anti-inflammatory effect which might be through novel
mechanism of action by targeting ERα and ERβ.
In humans, the biosynthesis of cholesterol plays an im-
portant role in hypercholesterolemia [174]. The rate limiting
enzyme in this process is 3-hydroxy-3-methylglutaryl-
coenzyme A reductase which has found to be the potent tar-
get of simvastatin and derivatives [174]. Polypharmacology
profile of the simvastatin includes another potent target, ERβ
[119]. It binds to the ERβ with an IC50 value 3.12 mM. Evi-
dence from the previous studies supports the binding of sim-
vastatin to ERβ. Benjamin Wolozin et al. found that simvas-
tatin causes a strong reduction in the in cidence of dementia
and Parkinson's disease [175]. The same author also found
the association of simvastatin with the decreased prevalence
of Alzheimer’s diseases but the effect on Alzheimer’s dis-
ease associated with 3-hydroxy-3-methylglutaryl-coenzyme
A reductase is not significant [176]. Therefore, the decrease
in the incidence of Alzheimer’s disease, dementia and Park-
inson’s disease could be explained through a new b iological
pathway of inhibition to ERβ by simvastatin.
6. LIMITATIONS
Network pharmacology has revolutionized the process of
drug discovery but still this area has some limitations and
has to go a long way. One of the limitations of network
pharmacology is that it relies on a single network. Address-
ing this issue, network pharmacology needs to integrate het-
erogeneous networks (i-e: drug chemical structure similarity
network, protein sequence similarity network, known drug–
target interaction networks, drug side-effect network, meta-
bolic network related to specific disease and target–protein
interaction network). This will further improve and spread
the successfu l implementation of network pharmacology in
the area of drug discovery. To capture the association be-
tween target proteins and drugs, network-based drug discov-
ery needs global network information. This would accelerate
the process in the area of personalized medicines which will
be more useful in a complex disease like cancer. Implemen-
tation of methodology for personalized medicine develop-
ment based on the network is another important requirement
of today’s world. This could be achieved through the integra-
tion of cancer hallmark-based network, tumor clone-based
network and sequencing technologies. Frequent development
and implementation of network-based models could be con-
structed to solve many important problems as follows: (i) the
prediction of personalized drug targets; (ii) prediction of
drug resistance (iii) personalized drug effect prediction (iv)
personalized molecular signature identification for therapeu-
tic evaluation after cancer tr eatment; (v) personalized cancer
risk prediction for healthy individuals. Successful network-
based models for these important problems would have a
critical impact on timely diagnosis, personalized treatment,
prognosis and personalized prevention of cancer [177]. Pa-
hikkala et al. [178] pointed out the following four important
facts, which should be taken into consideration for the model
development and evaluation because they can strongly influ-
ence the prediction performance: (i) problem formulation by
more realistic regression formulation rather than standard
binary classification; (ii) model prediction based on quantita-
tive bioactivity data rather than on/off interaction data; (iii)
model validation based on simple or nested cross validation;
(iv) model performance report based on different experimen-
Table 7. The table is showing Herbs Related Databases Biomolecular network resources (only shows ppi databases) and Pheno-
type Network Resources related to TC M network pharmacology.
Name
Web
Ref
TCM-ID
http://tcm.cz3.nus.edu.sg/group/tcm-id/
[153]
TCM Database@Taiwan
http://tcm.cmu.edu.tw
[154]
TCMGeneDIT
http://tcm.lifescience.ntu.edu.tw
[155]
CHMIS-C
http://sw16.im.med.umich.edu/chmis-c
[156]
TCMID
http://www.megabionet.org/tcmid/
[157]
HPRD
http://hprd.org/
[158]
MINT
http://mint.bio.uniroma2.it/mint/
[159]
STRING
http://string-db.org
[142]
DIP
http://dip.doe-mbi.ucla.edu
[160]
BioGRID
http://thebiogrid.org/
[144]
OMIM
http://www.omim.org
[161]
UMLS
http://www.nlm.nih.gov/research/umls/
[138]
HPO
http://www.human-phenotype-ontology .org
[136]

12 Current Topics in Medicinal Chemistry, 2018, Vol. 18, No. 00 Muhammad et al.
tal setting to find out whether training and test sets share
common drugs and targets, only drugs or targets or neither.
CONCLUSION
Network pharmacology approach has unique features of
being probable and regular. This approach is totally different
from “trial and error” and “Magic Built approach”. Network
pharmacology made the drug discovery process predictable
due to the computational supremacies. So, th is approach has
the capacity to manage big data. Unlike, the reductionist
method, network pharmacology can make the systematic
study of different drugs formulae achievable. Therefore, al-
though network pharmacology is stillat its initial stage , such
a novel approach will initiate new directions and lead to a
probable revolution in the modernization of network drugs,
and also provide new insights into the current drug discovery
field.
CONSENT FOR PUBLICATION
Not applicable.
CONFLICT OF INTEREST
The authors declare no conflict of interest, financial or
otherwise.
ACKNOWLEDGEMENTS
Declared none.
REFERENCES
[1] Kaufmann, S.H., Paul Ehrlich: founder of chemotherapy. Nature
reviews. Drug discovery, 2008, 7, (5), 373.
[2] Sams-Dodd, F., Target-based drug discovery: is something wrong?
Drug discovery today, 2005, 10, (2), 139-147.
[3] Zambrowicz, B.P.; Sands, A.T., Modeling drug action in the mouse
with knockouts and RNA interference. Drug Discovery Today:
TARGETS, 2004, 3, (5), 198-207.
[4] Winzeler, E.A.; Shoemaker, D.D.; Astromoff, A.; Liang, H.; An-
derson, K.; Andre, B.; Bangham, R.; Benito, R.; Boeke, J.D.;
Bussey, H., Functional characterization of the S. cerevisiae genome
by gene deletion and parallel analysis. Science (New York, N.Y.),
1999, 285, (5429), 901-906.
[5] Giaever, G.; Chu, A.M.; Ni, L.; Connelly, C.; Riles, L.; Veronneau,
S.; Dow, S.; Lucau-Danila, A.; Anderson, K.; Andre, B., Func-
tional profiling of the Saccharomyces cerevisiae genome. Nature,
2002, 418, (6896), 387-391.
[6] Goldstein, D.B., Common genetic variation and human traits. New
England Journal of Medicine, 2009, 360, (17), 1 696.
[7] Yang, K.; Bai, H.; Ouyang, Q.; Lai, L.; Tang, C., Finding multiple
target optimal intervention in disease-related molecular network.
Mol Syst Biol, 2008, 4, 228.
[8] Barabási, A.-L.; Mass Medical Soc, 2007.
[9] Pe'er, D.; Hacoh en, N., Principles and strateg ies for develop ing
network models in cancer. Cell, 2011, 144, (6), 864-873.
[10] Schadt, E.E., Molecular networks as sensors and drivers of com-
mon human diseases. Nature, 2009, 461, (7261), 218-223.
[11] Liu, Z.-P.; Wang, Y.; Zhang, X.-S.; Chen, L.-n., Network-based
analysis of complex diseases. IET Systems Biology, 2012, 6, (1),
22-33.
[12] Chen, X.; Ren, B.; Chen, M.; Wang, Q.; Zhang, L.; Yan, G.,
NLLSS: predicting synergistic drug combinations based on semi-
supervised learning. PLoS computational biology, 2016, 12, (7),
e1004975.
[13] Sun, X.; Bao, J.; You, Z.; Chen, X.; Cui, J., Modeling of signaling
crosstalk-mediated drug resistance and its implications on drug
combination. Oncotarget, 2016, 7, (39), 63995-64006.
[14] Hopkins, A.L., Network pharmacology: the next paradigm in drug
discovery. Nature chemical biology, 2008, 4, (11), 682-690.
[15] Hopkins, A.L., Network pharmacology. Nature biotechnology,
2007, 25, (10), 1110-1110.
[16] Roth, B.L.; Sheffler, D.J.; Kroeze, W.K., Magic shotguns versus
magic bullets: selectively non-selective drugs for mood disorders
and schizophrenia. Natu re reviews Drug discover y, 2004, 3, (4 ),
353-359.
[17] Wu, X.; Jiang, R.; Zhang, M.Q.; Li, S., Network!based global
inference of human disease genes. Molecular systems biology,
2008, 4, (1), 189.
[18] Jia-hu, P., New paradigm for drug discovery based on network
pharmacology [J]. Chinese Journal of New Drugs and Clinical
Remedies, 2009, 10, 002.
[19] Zhang, G.-b.; Li, Q.-y.; Chen, Q.-l.; Su, S.-b., Network pharmacol-
ogy: a new approach for chinese herbal medicine research. evi-
dence-based Complementary and Alternative Medicine, 2013,
2013.
[20] Li, S. In Proceedings of the 1st Academic Annual Meeting of the
China Association for Science and Technology, 1999.
[21] Li, S., Framework and practice of n etwork-based studies for Chi-
nese herbal formula. Journal of Chinese Integrative Medicine,
2007, 5, (5), 489-493.
[22] Li, S.; Zhang, Z.; Wu, L.; Zhang, X.; Li, Y.; Wang, Y., Under-
standing ZHENG in traditional Chinese medicine in the context of
neuro-endoc rine-immune network. IET Systems Biology, 2007, 1,
(1), 51-60.
[23] Li, S., Network systems underlying traditional Chinese medicine
syndrome and herb formula. Current Bioinformatics, 2009, 4, (3),
188-196.
[24] Shao, L., Network target: a starting point for traditional Chinese
medicine network pharmacology [J]. China Journal of Chinese
Materia Medica, 2011, 15, (36), 2017-2020.
[25] Li, S.; Zhang, B.; Zhang, N., Network target for screening syner-
gistic drug combinations with application to traditional Chinese
medicine. BMC systems biolog y, 2011, 5, (1), S10.
[26] Budovsky, A.; Fraifeld, V.E., Medicinal plants growing in the
Judea region: network approach for searching potential therapeutic
targets. Network Biology, 2012, 2, (3), 84.
[27] Zhou, W.; Cheng, X.; Zhang, Y., Network pharmacology- a new
philosophy for understanding of drug action and discove ry of new
drugs. Zhongguo Yaolixue yu Dulixue Zazhi- Chinese Journal of
Pharmacology and Toxicology, 2012, 26, (1), 4-9.
[28] Law, V.; Knox, C.; Djoumbou, Y.; Jewison, T.; Guo, A.C.; Liu, Y.;
Maciejewski, A.; Arndt, D.; Wilson , M.; Neveu, V., DrugBank 4.0:
shedding new light on drug metabolism. Nucleic acids research,
2013, 42, (D1), D1091-D10 97.
[29] Qin, C.; Zhang, C.; Zhu, F.; Xu, F.; Chen, S.Y.; Zhang, P.; Li,
Y.H.; Yang, S.Y.; Wei, Y.Q.; Tao, L., Therapeutic target database
update 2014: a resource for targeted therapeutics. Nucleic acids re-
search, 2013, 42, (D1), D1118-D1123.
[30] Günther, S.; Kuhn, M.; Dunkel, M.; Campillos, M.; Senger, C.;
Petsalak i, E.; Ahmed, J.; Urdiales, E.G.; Gewiess, A.; Jensen, L.J. ,
SuperTarget and Matador: resources for exploring drug-target rela-
tionships. Nucleic acids research, 2007, 36, (suppl_1), D919-D922.
[31] Gaulton, A.; Bellis, L.J.; Bento, A.P.; Chambers, J.; Davies, M.;
Hersey, A.; Light, Y.; McGlinchey, S.; Michalovich, D.; Al-
Lazikani, B., ChEMBL: a large-scale bioactivity database for drug
discovery. Nucleic acids research, 2011, 40, (D1), D1100-D1107.
[32] Kuhn, M.; Szklarczyk, D.; Pletscher-Frankild, S.; Blicher, T.H.;
Von Mering, C.; Jensen, L.J.; Bork, P., STITCH 4: integration of
protein–chemical interactions with user data. Nucleic acids re-
search, 2013, gkt1207.
[33] Hecker, N.; Ahmed, J.; von Eichborn, J.; Dunkel, M.; Macha, K.;
Eckert, A.; Gilson, M.K.; Bourne, P.E.; Preissner, R., SuperTarget
goes quantitative: update on drug–target interactions. Nucleic acids
research, 2011, gkr912.
[34] Magariños, M.P.; Carmona, S.J.; Crowther, G.J.; Ralph, S.A.;
Roos, D.S.; Shanmugam, D.; Van Voorhis, W.C.; Agüero, F., TDR
Targets: a chemogenomics resource for neglected diseases. Nucleic
acids research, 2011, 40, (D1), D1118-D1 127.
[35] Gao, Z.; Li, H.; Zhang, H.; Liu, X.; Kang, L.; Luo, X.; Zhu, W.;
Chen, K.; Wang, X.; Jiang, H., PDTD: a web-accessible protein da-
tabase for drug target identification. BMC bioinformatics, 2008, 9,
(1), 104.

Network Pharmacology: Exploring the Resources and Methodologies Current Topics in Medicinal Chemistry, 2018, Vol. 18, No. 00 13
[36] Emig, D.; Ivliev, A.; Pustovalo va, O.; Lancashire, L.; Bureeva, S.;
Nikolsky, Y.; Bessarabova, M., Drug target prediction and reposi-
tioning using an integrated network-based approach. PLoS One,
2013, 8, (4), e60618.
[37] Kuhn, M.; Campillos, M.; Letunic, I.; Jensen, L.J.; Bork, P., A side
effect resource to capture phenotypic effects of drugs. Mole cular
systems biology, 2010, 6, (1), 343.
[38] Seiler, K.P.; George, G.A.; Happ, M.P.; Bodycombe, N.E.; Carrin-
ski, H.A.; Norton, S.; Brudz, S.; Sullivan, J.P.; Muhlich, J.; Ser-
rano, M., ChemBank: a small-molecule screening and cheminfor-
matics resource database. Nucleic acids resea rch, 2007, 36,
(suppl_1), D351-D359.
[39] Kumar, R.; Chaudhary, K.; Gupta, S.; Singh, H.; Kumar, S.;
Gautam, A.; Kapoor, P.; Raghava, G.P., CancerDR: cancer drug re-
sistance database. Scientific repo rts, 2013, 3, 1445.
[40] Liu, T.; Lin, Y.; Wen, X .; Jorissen, R.N.; Gilson, M.K., Bind-
ingDB: a web-accessible database of experimentally determined
protein–ligand binding affin ities. Nucleic acids research, 2006, 35,
(suppl_1), D198-D201.
[41] Irwin, J.J.; Sterling, T.; Mysinger, M.M.; Bolstad, E.S.; Coleman,
R.G., ZINC: a free tool to discover chemistry for biology. Journal
of chemical information and modeling, 2012, 52, (7), 1757-1768.
[42] Halling-Brown, M.D.; Bulusu, K.C.; Patel, M.; Tym, J.E.; Al-
Lazikani, B., canSAR: an integrated cancer public translational re-
search and drug discovery resource. Nucleic acids research, 2011,
40, (D1), D947-D956.
[43] Chen, X.; Ren, B.; Chen, M.; Liu, M.-X.; Ren, W.; Wang, Q.-X.;
Zhang, L.-X.; Yan, G.-Y., ASDCD: antifungal synergistic drug
combination database. PloS one, 2014, 9, (1), e86499.
[44] Yamanishi, Y.; Kotera, M.; Moriya, Y.; Sawada, R.; Kanehisa, M.;
Goto, S., DINIES: drug–target interaction netwo rk inference engine
based on supervised analy sis. Nucleic acids research, 2014, 42,
(W1), W39-W45.
[45] Nickel, J.; Gohlke, B.-O.; Erehman, J.; Banerjee, P.; Rong, W.W.;
Goede, A.; Dunkel, M.; Preissner, R., SuperPred: update on drug
classification and target prediction. Nucleic acids research, 2014,
42, (W1), W26-W31.
[46] Gfeller, D.; Grosdidier, A.; Wirth, M.; Daina, A.; Michielin, O.;
Zoete, V., SwissTargetPrediction: a web server for target prediction
of bioactive small molecules. Nucleic acids research, 2014, 42,
(W1), W32-W38.
[47] Russom, P., Big data analytics. TDWI best practices report, fo urth
quarter, 2011, 19, 40.
[48] Zhang, W., Pattern classification and recognition of invertebrate
functional groups using self-organizing neural networks. Environ-
mental monitoring and assessment, 2007, 130, (1), 415-422.
[49] GAASTERLAND, T.; RAGAN, M.A., Microbial genescapes:
phyletic and functional patterns of ORF distribution among pro-
karyotes. Microb ial & compa rative genomics, 1998, 3, (4), 199-
217.
[50] Dandekar, T.; Snel, B.; Huynen, M.; Bork, P., Conservation of
gene order: a fingerprint of proteins that physically interact. Trends
in biochemical sciences, 1998 , 23, (9), 324-328.
[51] Marcotte, E.M.; Pellegrini, M.; Ng, H.-L.; Rice, D.W.; Yeates,
T.O.; Eisenberg, D., Detecting protein function and protein -protein
interactions from genome sequences. Science (New York, N.Y.),
1999, 285, (5428), 751-753.
[52] Fryxell, K.J., The co evolution of gen e family tre es. Trends in Ge-
netics, 1996, 12, (9), 364-369.
[53] Göbel, U.; Sander, C.; Schneider, R.; Valencia, A., Correlated
mutations and residue contacts in proteins. Proteins: Structure,
Function, and Bioinformatics, 1994, 18, (4), 309-317 .
[54] Aloy, P.; Russell, R.B., InterPreTS: protein interaction prediction
through tertiary structure. Bioinformatics (Oxford, England), 2003,
19, (1), 161-162.
[55] Bock, J.R.; Gough, D.A., Predicting protein–protein interactions
from primary structure. Bioinformatics (Oxford, England), 2001,
17, (5), 455-460.
[56] Zhang, W., Network Biology: an exciting frontier science. Network
Biology, 2011, 1, (1), 79-80.
[57] Dobson, C.M., Chemical space and biology. Nature, 2004, 432,
(7019), 824-828.
[58] Kanehisa, M.; Goto, S.; Hattori, M.; Aoki-Kinoshita, K.F.; Itoh,
M.; Kawashima, S.; Katayama, T.; Araki, M.; Hirakawa, M., From
genomics to chemical genomics: new developments in KEGG. Nu-
cleic acids research, 2006, 34, (suppl 1), D354-D357.
[59] Stockwell, B.R., Chemical genetics: ligand-based discovery of
gene function. Nature Reviews Genetics, 2000, 1, (2), 116-125.
[60] Wheeler, D.L.; Barrett, T.; Benson, D.A.; Bryant, S.H.; Canese, K.;
Chetvernin, V.; Church, D.M.; DiCuccio, M.; Edgar, R.; Federhen,
S., Database resources of the national center for biotechnology in-
formation. Nucleic acids research, 2007, 35, (supp l 1), D5-D12.
[61] Haggarty, S.J.; Koeller, K.M.; Wong, J.C.; Butcher, R.A.; Schrei-
ber, S.L., Multidimensional chemical genetic analysis of diversity-
oriented synthesis-derived deacetylase inhibitors using cell-based
assays. Chemistry & biology, 2003, 10, (5), 383-396.
[62] Kuruvilla, F.G.; Shamji, A.F.; Sternson, S.M.; Hergenrother, P.J.;
Schreiber, S.L., Dissecting glucose signalling with diversity-
oriented synthesis and small-molecule microarrays. Nature, 2002,
416, (6881), 653-657.
[63] Yamanishi, Y.; Araki, M.; Gutteridge, A.; Honda, W.; Kanehisa,
M., Prediction of drug–target interaction networks from the integra-
tion of chemical and genomic spaces. Bioinformatics (Oxford, Eng-
land), 2008, 24, (13 ), i232-i240 .
[64] Schomburg, I.; Chang, A.; Ebeling, C.; Gremse, M.; Heldt, C.;
Huhn, G.; Schomburg, D., BRENDA, the enzyme database: up-
dates and major new developments. Nucleic acids research, 2004,
32, (suppl 1 ), D431-D433.
[65] Wishart, D.S.; Knox, C.; Guo, A.C.; Cheng, D.; Shrivastava, S.;
Tzur, D.; Gautam, B.; Hassanali, M., DrugBank: a knowledgebase
for drugs, drug actions and drug targets. Nucleic acids research,
2007, 36, (suppl_1), D901-D906.
[66] Hattori, M.; Okuno, Y.; Goto, S.; Kanehisa, M., Development of a
chemical structure comparison method for integrated analysis of
chemical and genomic information in the metabolic pathways.
Journal of the American Chemical Society, 2003, 125, (39), 11853-
11865.
[67] Cheng, A.C.; Coleman, R.G.; Smyth, K.T.; Cao, Q.; Soulard, P.;
Caffrey, D.R.; Salzberg, A.C.; Huang, E.S., Structure-based maxi-
mal affinity model predicts small-molecule druggability. Nature
biotechnology, 2007, 25, (1), 71-75.
[68] Rarey, M.; Kramer, B.; Lengauer, T.; Klebe, G., A fast flexible
docking method using an incremental construction algorithm.
Journal of molecular biology, 1996, 261, (3), 470-489.
[69] Liebler, D.C.; Guengerich, F.P., Elucidating mechanisms of drug-
induced toxicity. Nature reviews Drug discovery, 2005, 4, (5), 4 10-
420.
[70] Blagg, J., Structure–activity relationships for in vitro and in vivo
toxicity. Annual Reports in Medicinal Chemistry, 2006, 41, 353-
368.
[71] Whitebread, S.; Hamon, J.; Bojanic, D.; Urban, L., Keynote re-
view: in vitro safety pharmacology profiling: an essential tool for
successful drug development. Drug discovery today, 2005, 10,
(21), 1421-1433.
[72] Ashburn, T.T.; Thor, K.B., Drug repositioning: identifying and
developing new uses for existing drugs. Nature reviews Drug dis-
covery, 2004, 3, (8), 673-683.
[73] Campillos, M.; Kuhn, M.; Gavin, A.-C.; Jensen, L.J.; Bork, P.,
Drug target identification using side-effect similarity. Science (New
York, N.Y.), 2008, 321, (5886), 263-266.
[74] Finlayson, K.; Witchel, H.J.; McCulloch, J.; Sh arkey, J., Acquired
QT interval prolongation and HERG: implications for drug discov-
ery and development. European journal of pharmacology, 2004,
500, (1), 129-142.
[75] Tatonetti, N.P.; Liu, T.; Al tman, R.B., Predicting drug side-effects
by chemical systems biology. Genome biology, 2009, 10, (9), 238.
[76] Cheng, F.; Li, W.; Wang, X.; Zhou, Y.; Wu, Z.; Shen, J.; Tang, Y.,
Adverse drug events: database construction and in silico prediction.
Journal of chemical information and modeling, 2013, 53, (4), 744-
752.
[77] Cheng, F.; Li, W.; Wu, Z.; Wang, X.; Zhang, C.; Li, J.; Liu, G.;
Tang, Y., Prediction of polypharmacological profiles of drugs by
the in tegration of chemical, side effect, an d therapeutic space.
Journal of chemical information and modeling, 2013, 53, (4), 753-
762.
[78] Yamanishi, Y.; Kotera, M.; Kanehisa, M.; Goto, S., Drug-target
interaction prediction from chemical, genomic and pharmacologi-
cal data in an integrated framework. Bioinformatics (Oxford, Eng-
land), 2010, 26, (12 ), i246-i254 .
[79] Scheiber, J.; Jenkins, J.L.; Sukuru, S.C.K.; Bender, A.; Mikhailov,
D.; Milik, M.; Azzaoui, K.; Whitebread, S.; Hamon, J.; Urban, L.,

14 Current Topics in Medicinal Chemistry, 2018, Vol. 18, No. 00 Muhammad et al.
Mapping adverse drug rea ctions in chemical space. Journal of me-
dicinal chemistry, 2009, 52, (9), 3103-3107.
[80] Yamanishi, Y. In Advances in Neural Information Processing Sys-
tems, 2009, pp 1841-1848.
[81] Tobinick, E.L., The value of drug repositioning in the current
pharmaceutical market. Drug news & perspectives, 2009, 22, (2),
119-125.
[82] Bandekar, M.; Anwikar, S.; Kshirsagar, N., Quality check of spon-
taneous adverse drug reaction reporting forms of different coun-
tries. Pharmacoepidemiology and drug safety, 2010, 19, (11),
1181-1185.
[83] Yang, L.; Xu, L.; He, L., A CitationRank algorithm inheriting
Google technology designed to highlight genes responsible for se-
rious adverse drug reaction. Bioinformatics (Oxford, England),
2009, 25, (17), 2244-2250.
[84] Rognan, D., Structure!based approache s to target fish ing and liga nd
profiling. Molecula r info rmatics, 2010, 29, (3), 176-187.
[85] Berger, S.I.; Iyengar, R., Role of systems pharmacology in under-
standing drug adverse events. Wiley Interdisciplinary Reviews: Sys-
tems Biology and Medicine, 2011, 3, (2), 129-135.
[86] De Franchi, E.; Schalon, C.; Messa, M.; Onofri, F.; Benfenati, F.;
Rognan, D., Binding of protein kinase inhibitors to synapsin I in-
ferred from pair-wise binding site similarity measurements. PloS
one, 2010, 5, (8), e12214.
[87] Keiser, M.J.; Setola, V.; Irwin, J.J.; Laggner, C.; Abbas, A.I.;
Hufeisen, S.J.; Jensen, N.H.; Kuijer, M.B.; Matos, R.C.; Tran, T.B.,
Predictin g new molecular targets for known drugs. Nature, 2009,
462, (7270), 175-181.
[88] Li, Y .Y.; An, J .; Jones, S.J., A large-scale computational approach
to drug repositioning. Genome Informatics, 2006, 17, (2), 239-247.
[89] Lee, S.; Park, K.; Kim, D., Building a drug–target network and its
applications. Expert opinion on drug discovery, 2009, 4, (11),
1177-1189.
[90] Ma, X.H.; Shi, Z.; Tan , C.; Jiang, Y.; Go, M.L.; Low, B.C.; Chen,
Y.Z., In-Silico approaches to multi-target drug discovery. Pharma-
ceutical research, 2010, 27, (5), 739-749.
[91] Keiser, M.J.; Irwin, J.J.; Shoichet, B.K., The chemical basis of
pharmacology. Biochemistry, 2010, 49, (48), 10267-10276.
[92] Yang, L.; Chen, J.; He, L., Harvesting candidate genes responsible
for serious adverse drug reactions from a chemical-protein interac-
tome. PLoS co mputational biology, 2009, 5, (7), e1000441.
[93] Rich, R.L.; Day, Y.S.; Morton, T.A.; Myszka, D.G., High-
resolution and high-throug hput protocols for measuring
drug/human serum albumin interactions using BIACORE. Analyti-
cal biochemistry, 2001, 296, (2), 197-207.
[94] Von Rechenberg, M.; Blake, B.K.; Ho, Y.S.J.; Zhen, Y.;
Chepanoske, C.L.; Richardson, B.E.; Xu, N.; Kery, V., Ampicil-
lin/penicillin!binding protein interactions as a model drug!target
system to optimize affinity pull!down and mass spectrometric
strategies for target and pathway identification. Proteomics, 2005,
5, (7), 1764-1773.
[95] Ji, Z.L.; Han, L.Y.; Yap, C.W.; Sun, L.Z.; Chen, X.; Chen, Y.Z.,
Drug Adverse Reaction Target Database (DART). Drug safety,
2003, 26, (10), 685-690.
[96] Zhang, J.-X.; Huang, W.-J.; Zeng, J.-H.; Huang, W.-H.; Wang, Y.;
Zhao, R.; Han, B.-C.; Liu, Q.-F.; Chen, Y.-Z.; Ji, Z.-L., DITOP:
drug-induced toxicity related protein database. Bioinformatics (Ox-
ford, England), 2007, 23, (13), 1710-1712.
[97] Chen, X.; Ji, Z.L.; Chen, Y.Z., TTD: therapeutic target database.
Nucleic acids research, 2002, 30, (1), 412-41 5.
[98] Mallal, S.; Phillip s, E.; Carosi, G.; Molina, J.-M.; Workman, C.;
Tomažič, J.; Jägel-Guedes, E.; Rugina, S.; Kozyrev, O.; Cid, J.F.,
HLA-B* 5701 screening for hypersensitivity to abacavir. New Eng-
land Journal of Medicine, 2008, 358, (6), 568-579.
[99] Martin, A.M.; Nolan, D.; Gaudieri, S.; A lmeida, C.A.; Nolan, R.;
James, I.; Carvalho, F.; Phillips, E.; Christiansen, F.T.; Purcell,
A.W., Predisposition to abacavir hypersensitivity conferred by
HLA-B* 5701 and a haplotypic Hsp70-Hom variant. Proceedings
of the National Academy of Sciences, 2004, 101, (12), 4180-4185.
[100] Chessman, D.; Kostenko, L.; Lethborg, T.; Purcell, A.W.; William-
son, N.A.; Chen, Z.; Kjer-Nielsen, L.; Mifsud, N.A.; Tait, B.D.;
Holdsworth, R., Human leukocyte antigen class I-restricted activa-
tion of CD8+ T cells provides the immunog enetic basis of a sys-
temic drug hypersensitivity. Immunity, 2008, 28, (6), 822-832.
[101] Stewart-Jones, G.B.; Gillespie, G.; Overton, I.M.; Kaul, R.; Roche,
P.; McMichael, A.J .; Rowland-Jones, S.; Jones, E.Y., Structures of
three HIV-1 HLA-B* 5703-peptide complexes and identification of
related HLAs potentially associated with long-term nonprogres-
sion. The Journal of Immunology, 2005, 175, (4), 2459-2468.
[102] Luo, H.; Chen, J.; Shi, L.; Mikailov, M.; Zhu, H.; Wang, K.; He,
L.; Yang, L., D RAR-CPI: a server for identifying drug reposition-
ing potential and adverse drug reactions via the chemical–protein
interactome. Nucleic acids research, 2011, 39, (suppl 2), W492-
W498.
[103] Lamb, J.; Crawford, E.D.; Peck, D.; Mod ell, J.W.; Blat , I.C. ;
Wrobel, M.J.; Lerner, J.; Brunet, J.-P.; Subramanian, A.; Ross,
K.N., The Connectivity Map: using gene-exp ression signatures to
connect small molecules, genes, and disease. Science (New York,
N.Y.), 2006, 313, (5795), 1929-1935.
[104] Yang, L.; Luo, H.; Chen, J.; Xing, Q.; He, L., SePreSA: a server for
the prediction of populations susceptible to serious adverse drug
reactions implementing the methodology of a chemical–protein in-
teractome. Nucleic acids research, 20 09, 37, (suppl 2), W406-
W412.
[105] Young, P.W.; Buckle, D.R.; Cantello, B.C.; Chapman, H.; Clap-
ham, J.C.; Coyle, P.J.; Haigh, D.; Hindley, R.M.; Holder, J.C.;
Kallender, H., Identification of high-affinity binding sites for the
insulin sensitizer rosiglitazone (BRL-49653) in rodent and human
adipocytes using a radioiodinated ligand for peroxisomal prolifera-
tor-activated receptor γ. Journal of Pharmacology and Experimen-
tal Therapeutics, 1998, 284, (2), 751-759.
[106] Chen, H.; Wang, X.; Jin, H.; Liu, R.; Hou, T., Discovery of the
molecular mechanisms of the novel chalcone-based Magnaporthe
oryzae inhibitor C1 using transcriptomic profiling and co-
expression network analysis. SpringerPlu s, 2016, 5, (1), 1851.
[107] Shu, M.; Zai, X.; Zhang, B.; Wang, R.; Lin, Z., Hypothyroidism
side effect in patients treated with sunitinib or sorafenib: clinical
and structural analyses. PloS one, 2016, 11, (1), e0 147048.
[108] Ye, X.-y.; Ling, Q.-z.; Chen, S.-j., Identification o f a potential
target of capsaicin by computational target fishing. Evidence-Based
Complementary and Alternative Medicine, 2015, 2015.
[109] Iyer, P.; Bolla, J.; Kumar, V.; Gill, M.S.; Sobhia, M.E., In silico
identification of targets for a novel scaffold, 2-thiazolylimino-5-
benzylidin-thiazolidin-4-one. Molecular diversity, 2015, 19, (4),
855-870.
[110] Qiu, J.-X.; Zhou, Z.-W.; He, Z.-X.; Zhao, R.J.; Zhang, X.; Yang,
L.; Zhou, S.-F.; Mao, Z.-F., Plumbag in elicits differential proteo-
mic responses mainly involving cell cycle, apoptosis, autophagy,
and epithelial-to-mesenchymal transition pathways in human pros-
tate cancer PC-3 and DU145 cells. Drug design, development and
therapy, 2015, 9, 349.
[111] Su, J.; Chang, C.; Xiang, Q.; Zhou, Z.-W.; Luo, R.; Yang, L.; He,
Z.-X.; Yang, H.; Li, J.; Bei, Y., Xyloketal B, a marine compound,
acts on a network of molecular proteins and regulates the activity
and expression of rat cytochrome P450 3a: a bioinformatic and
animal study. Drug design, development and therapy, 2014, 8,
2555.
[112] Chen, S.-J., A potential target of Tanshinone IIA for acute promye-
locytic leukemia revealed by inverse docking and drug repurpos-
ing. Asian Pacific journal of cancer prevention : APJCP, 2014, 15,
(10), 4301-4305.
[113] Qi, L.; Ding, Y., Potential antitumor mechanisms of phenothiazine
drugs. Science China Life Sciences, 2013, 56, (11), 1020-1027.
[114] Faraone, S.V.; Zhang!James, Y., Can sodium/hydrogen exchange
inhibitors be repositioned for treating attention deficit hyperactivity
disorder? An in silico approach. American Journal of Medical Ge-
netics Part B: Neuropsychiatric Genetics, 2013, 162, (7), 711-717.
[115] Zhou, T.; Su, R.-Q.; Liu, R.-R.; Jiang, L.-L.; Wang, B.-H.; Zhang,
Y.-C., Accurate and diverse recommendations via eliminating re-
dundant correlations. New Journal of Physics, 2009, 11, (12),
123008.
[116] Zhou, T.; Kuscsik, Z.; Liu, J.-G.; Medo, M.; Wakeling, J.R.;
Zhang, Y.-C., Solving the apparent diversity-accuracy dilemma of
recommender systems. Proceedings of the National Academy of
Sciences, 2010, 107, (10), 4511-4515.
[117] Sarwar, B.; Karypis, G.; Konstan , J.; Riedl, J. In Proceedings of the
10th international conference on World Wide Web; ACM, 2001, pp
285-295.
[118] Herlocker, J.L.; Konstan, J.A.; Terveen, L.G.; Riedl, J.T., Evaluat-
ing collaborative filtering recommender systems. ACM Transac-
tions on Information Systems (TOIS), 2004, 22, (1), 5-53.

Network Pharmacology: Exploring the Resources and Methodologies Current Topics in Medicinal Chemistry, 2018, Vol. 18, No. 00 15
[119] Cheng, F.; Liu, C.; Jiang, J.; Lu, W.; Li, W.; Liu, G.; Zhou, W.;
Huang, J.; Tang, Y., Prediction of drug-target interactions and drug
repositioning via network-based inference. PLoS computational bi-
ology, 2012, 8, (5), e1002503.
[120] Cheng, F.; Zhou, Y.; Li, W.; Liu, G .; Tang, Y., Prediction of
chemical-protein interactions network with weighted network-
based inference method. PloS one, 2012, 7, (7), e41064.
[121] Jia, C.-X.; Liu, R.-R.; Sun, D.; Wang, B.-H., A new weighting
method in network-based recommendation. Physica A: Statistical
Mechanics a nd its Application s, 2008, 387, (23), 5887-5891.
[122] Wu, Z.; Cheng, F.; Li, J.; Li, W.; Liu, G.; Tang, Y., SDTNBI: an
integrated network and chemoinformatics tool for systematic pre-
diction of drug–target interactions and drug repositioning. Briefin gs
in bioinformatics, 2016, bbw012.
[123] Cheng, F.; Zhou, Y.; Li, J.; Li, W.; Liu, G.; Tang, Y., Prediction of
chemical–protein interactions: multitarget-QSAR versus computa-
tional chemogenomic me thods. Molecular bioSystems, 2012, 8, (9),
2373-2384.
[124] Shen, J.; Cheng, F.; Xu, Y.; Li, W.; Tang, Y., Estimation of ADME
properties with substructure pattern recognition. Journal of chemi-
cal information and modeling, 2010, 50, (6), 1034-1041.
[125] Yamanishi, Y.; Pauwels, E.; Saigo, H.; Stoven, V., Extracting sets
of chemical substructures and protein domains governing drug-
target interactions. Journal of chemical information and modeling,
2011, 51, (5), 1183-1194.
[126] Klekota, J.; Roth, F.P., Chemical substructures that enrich for bio-
logical activity. Bioinformatics (Oxford, England), 2008, 24, (21),
2518-2525.
[127] Wu, Z.; Lu, W.; Wu, D.; Luo, A.; Bian, H.; Li, J.; Li, W.; Liu, G. ;
Huang, J.; Cheng, F., In silico prediction of chemical mechanism of
action via an improved network!based inference method. British
journal of pharmacology, 2016, 173, (23), 337 2-3385.
[128] Jahchan, N.S.; Dudley, J.T.; Mazur, P.K.; Flores, N.; Yang, D.;
Palmerton, A.; Zmoos, A.-F.; Vaka, D.; Tran, K.Q.; Zhou, M., A
drug repositioning approach identifies tricyclic antidepressants as
inhibitors of small cell lung cancer and other neuroendocrine tu-
mors. Cancer discovery, 2013, 3, (12), 1364-1377.
[129] Shchors, K.; Massaras, A.; Hanahan, D., Dual targeting of the
autophagic regulatory circuitry in gliomas with repurposed drugs
elicits cell-lethal autophagy and therapeutic benefit. Cancer cell,
2015, 28, (4), 456-471.
[130] Cheng, F.; Hong, H.; Yang, S.; Wei, Y., Individualized network-
based drug repositioning infrastructure for precision oncology in
the pano mics era. Briefings in bioinformatics, 2016, bbw051.
[131] Li, J.; Lei, K.; Wu, Z.; Li , W.; Liu, G.; Liu, J.; Cheng, F.; Tang, Y.,
Network-based identification of microRNAs as potential pharma-
cogenomic biomarkers for anticancer drugs. Oncotarget, 2016, 7,
(29), 45584.
[132] Chen, X.; Liu, M.X.; Yan, G.Y., Drug-target interaction prediction
by random walk on the heterogeneous network. Molecu lar bioS ys-
tems, 2012, 8, (7), 19 70-1978.
[133] Wang, L.; You, Z.H.; Chen, X.; Yan, X.; Liu, G.; Zhang, W.,
RFDT: A Rotation Forest-based Predictor for Predicting Drug-
Target Interactions using Drug Structure and Protein Sequence In-
formation. Current protein & peptide science, 2016.
[134] Huang, Y.A.; You, Z.H.; Chen, X., A systematic prediction of
drug-target interactions using molecular fingerprints and protein
sequences. Current protein & peptide science, 2016.
[135] Yao, X.; Hao, H.; Li, Y.; Li, S., Modularity-based credible predic-
tion of disease genes an d detection of disease subtypes o n the phe-
notype-gene heterogeneous network. BMC systems biology, 2011,
5, (1), 79.
[136] Robinson, P.N.; Mundlos, S., The human phenotype ontology.
Clinical genetics, 2010, 77, (6), 525-534.
[137] Gu, J.; Chen, Y.; Li, S.; Li, Y., Identification of responsive gene
modules by network-based gene clustering and extending: applica-
tion to inflammation and angiogenesis. BMC systems biology,
2010, 4, (1), 47.
[138] Bodenreider, O., The unified medical language system (UMLS):
integrating biomedical terminology. Nucleic acids research, 2004,
32, (suppl 1 ), D267-D270.
[139] Li, L.; Zhang, N.; Li, S., Ranking effects of candidate drugs on
biological process by integrating network analysis and Gene Ontol-
ogy. Chinese Science Bulletin, 2010, 55, (26), 2974-2980.
[140] Zhang, N.; Li, S.; Zhang, B.; Google Patents, 2012.
[141] Li, S.; Zhang, B.; Jiang, D.; Wei, Y.; Zhang, N., Herb network
construction and co-module analysis for uncovering the combina-
tion rule of traditional Chinese herbal formulae. BMC bioinformat-
ics, 2010, 11, (11), S6.
[142] Franceschini, A.; Szklarczyk, D.; Frankild, S.; Kuhn, M.; Simono-
vic, M.; Roth, A.; Lin, J.; Minguez, P.; Bork, P.; Von Mering, C.,
STRING v 9. 1: protein-protein interaction networks, with increased
coverage and integration. Nucleic acids research, 2012, 41, (D1),
D808-D815.
[143] Li, S.; Wu, L.; Zhang, Z., Constructing biological networks through
combined literature mining and microarray analysis: a LMMA ap-
proach. Bioinformatics (Oxford, England ), 2006, 22, (17), 2143-
2150.
[144] Stark, C.; Breitk reutz, B.-J.; Chatr-Aryamontri, A.; Boucher, L.;
Oughtred, R.; Livstone, M.S.; Nixon, J.; Van Auken, K.; Wang, X.;
Shi, X., The BioGRID interaction database: 2011 update. Nucleic
acids research, 2010, 39, (suppl_1), D698 -D704.
[145] Zhao, S.; Li, S., Network-based relating pharmacological and ge-
nomic spaces for drug target identification. PloS one, 2010 , 5, (7),
e11764.
[146] Zhuang, Y.; Li, S.; Li, Y ., dbNEI: a specific database for neuro-
endocrine-immune interactions. Neuroendocrinology Letters, 2006,
27, (1-2), 53-59.
[147] Shao, L.; ZHANG, B., Traditional Chinese medicine network
pharmacology: theory, methodology and application. Chinese jour-
nal of natural medicines, 2013, 11, (2), 110-120.
[148] Zhu, F.; Shi, Z.; Qin, C.; Tao, L.; Liu, X.; Xu, F.; Zhang, L.; Song,
Y.; Liu, X.; Zhang, J., Therapeutic target database update 2012: a
resource for facilitating target-oriented drug discovery. Nucleic ac-
ids research, 2011, gkr797.
[149] Wishart, D.S.; Knox, C.; Guo, A.; Shrivastava, S.; Hassanali, M.;
Stothard , P.; Woolsey, J., DrugBank: A comprehensive resource
for in silico drug discovery and explorat. 2006.
[150] Li, S.; Zhao, S., 2010.
[151] Yan, H.; Zhang, B.; Li, S.; Zhao, Q., A formal model for analyzing
drug combination effects and its application in TNF-α-induced
NFκB pathway. BMC systems biology, 2010, 4, (1), 50.
[152] Gu, J.; Li, S., Towards integrative annotation of the cell-type spe-
cific gene functional and signaling map in vascular endothelial
cells. Molecula r bioSystems, 2012, 8, (8), 2041-2049.
[153] Ji, Z.; Zhou, H.; Wang, J.; Han, L.; Zheng, C.; Chen, Y., Tradi-
tional Chinese medicine information database. Journal of ethno-
pharmacology, 2006, 103, (3), 501.
[154] Chen, C.Y.-C., TCM Database@ Taiwan: the world's largest tradi-
tional Chinese medicine database for drug screening in silico. PloS
one, 2011, 6, (1), e15939.
[155] Fang, Y.-C.; Huang, H.-C.; Chen, H.-H.; Juan, H.-F.,
TCMGeneDIT: a d atabase for associated traditional Chinese medi-
cine, gene and disease information using text mining. BMC com-
plementary and alternative medicine, 2008, 8, (1), 58.
[156] Fang, X.; Shao, L.; Zhang, H.; Wang, S., CHMIS-C: a comprehen-
sive herbal medicine information system for cancer. Journal of me-
dicinal chemistry, 2005, 48, (5), 1481-1488.
[157] Xue, R.; Fang, Z.; Zhang, M.; Yi, Z.; Wen, C.; Shi, T., TCMID:
traditional Chinese medicine integrative database for herb molecu-
lar mechanism analysis. Nucleic acids research, 2012, gks1100.
[158] Goel, R.; Harsha, H.; Pandey, A.; Prasad, T.K., Human Protein
Reference Database and Human Proteinpedia as resources for
phosphoproteome analysis. Molecular b ioSystems, 2012, 8, (2),
453-463.
[159] Licata, L.; Briganti, L.; Peluso, D.; Perfetto, L.; Iannuccelli, M.;
Galeota, E.; Sacco, F.; Palma, A.; Nardozza, A.P.; Santonico, E.,
MINT, the molecular interaction database: 2012 update. Nucleic
acids research, 2011, 40, (D1), D857-D861.
[160] Xenarios, I.; Salwinski, L.; Duan, X.J.; Higney, P.; Kim, S.-M.;
Eisenberg, D., DIP, the Database of Interacting Proteins: a research
tool for studying cellular networks of protein interactions. Nucleic
acids research, 2002, 30, (1), 303-305.
[161] Hamosh, A.; Scott, A.F.; Amberger, J.S.; Bocchini, C.A.;
McKusick, V.A. , Online Mendelian Inheritance in Man (OMIM), a
knowledgebase of human genes and genetic disorders. Nucleic ac-
ids research, 2005, 33, (suppl 1 ), D514-D517.
[162] Russ, A.P.; Lampel, S., The druggable genome: an update. Drug
discovery today, 2005, 10, (23), 160 7-1610.

16 Current Topics in Medicinal Chemistry, 2018, Vol. 18, No. 00 Muhammad et al.
[163] Imming, P.; Sinning, C.; Meyer, A., Drugs, their targets and the
nature and number of drug targets. Nature reviews Drug discovery,
2006, 5, (10), 821-834.
[164] Overington, J.P.; Al-Lazikani, B.; Hopkins, A.L., How many drug
targets are there? Nature reviews Drug discovery, 2006, 5, (12),
993-996.
[165] Attur, M.; Dave, M.; Tsunoyama, K.; Akamatsu, M.; Kobori, M.;
Miki, J .; Abramson, S.; Katoh, M.; Amin, A., " A System Biology"
Approach to Bioinformatics and Functional Genomics in Complex
Human Diseases: Arthritis. Current issues in molecular biology,
2002, 4, 129-146.
[166] Liu, A.; Du, G., Network pharmacology: new guidelines for drug
discovery. Yao xue xue bao= Acta pharmaceutica Sinic a, 201 0, 45,
(12), 1472-1477.
[167] Barnes, N.; Thomas, M.; Price, D .; Tate, H ., The national montelu-
kast survey. Journal of allergy and clinical immunology, 2005,
115, (1), 47-54.
[168] Langlois, A.; Ferland, C.; Tremblay, G.M.; Laviolette, M., Monte-
lukast regulates eosinophil protease activity through a leukotriene-
independent mechanism. Journal of allergy and clinical immunol-
ogy, 2006, 118, (1), 113-119.
[169] Faul, J.L.; Wilson, S.R.; Chu, J.W.; Canfield, J.; Kuschner, W.G.,
The effect o f an inhaled corticosteroid on glucose control in type 2
diabetes. Clinical medicine & research, 2009, 7, (1-2), 14-20.
[170] Willett, P.; Barnard, J.M.; Downs, G.M., Chemical similarity
searching. Journal of chemical information and computer sciences,
1998, 38, (6), 983-996.
[171] Knox, C.; Law, V.; Jewison, T.; Liu, P.; Ly, S.; Frolkis, A.; Pon,
A.; Banco, K.; Mak, C.; Neveu, V., DrugBank 3.0: a comprehen-
sive resource for ‘omics’ research on drugs. Nucleic acids re-
search, 2010, 39, (suppl_1), D1035-D1041.
[172] Lehmann, J.M.; Lenhard, J.M.; Oliver, B.B.; Ringold, G.M.;
Kliewer, S.A., Peroxisome proliferator-activated receptors α and γ
are activated by indomethacin and other non-steroidal anti-
inflammatory drugs. Journal of Biological Chemistry , 1997, 272,
(6), 3406-3410.
[173] Zhou, H.; Liu, W.; Su, Y.; We i, Z.; L iu, J.; Kolluri, S.K.; Wu, H.;
Cao, Y.; Chen, J.; Wu, Y., NSAID sulindac and its analog bind
RXRα and inhibit RXRα-dependent AKT signaling. Cancer cell,
2010, 17, (6), 560-573.
[174] Liu, L.; Zhang, R.; Zhao, J.J.; Rogers, J.D.; Hsieh, J.Y.-K.; Fang,
W.; Matuszewski, B.K.; Dobrinska, M.R., Determination of sim-
vastatin-derived HMG-CoA reductase inhibitors in biomatrices us-
ing an automated enzyme inhibition assay with radioactivity detec-
tion. Journal of pharmaceutical and biomed ical analysis, 2003, 32,
(1), 107-123.
[175] Wolozin, B.; Wang, S.W.; Li, N.-C.; Lee, A.; Lee, T.A.; Kazis,
L.E., Simvastatin is associated with a red uced incidence of demen-
tia and Parkinson's disease. BMC medicine, 2007, 5, (1), 20.
[176] Wolozin, B.; Kellman, W.; Ruosseau, P.; Celesia, G.G.; Siegel, G.,
Decreased prevalence of Alzheimer disease associated with 3-
hydroxy-3-methyglutaryl coenzyme A reductase inhibitors. Ar-
chives of neurology, 2000, 57, (10), 1439-1443.
[177] Wang, E.; Zaman, N.; Mcgee, S.; Milanese, J.-S.; Masoudi-Nejad,
A.; O’Connor-McCourt, M. In Seminars in cancer biology; El-
sevier, 2015; Vol. 30, pp 4-12.
[178] Pahikkala, T.; Airola, A.; Pietilä, S.; Shakyawar, S.; Szwajda, A.;
Tang, J.; Aittokallio, T., Toward more realistic drug–target interac-
tion predictions. Briefings in bioinformatics, 2014, bbu010.
DISCLAIMER: The above article has been published in Epub (ahead of print) on the basis of the materials provided by the author. The
Editorial Department reserves the righ t to make minor modifications for further improvement of the manuscript.