Arch. Pharm. Res.

(2015) 38:1686–1701
DOI 10.1007/s12272-015-0640-5

REVIEW

Role of computer-aided drug design in modern drug discovery

Stephani Joy Y. Macalino1 • Vijayakumar Gosu1 • Sunhye Hong1 •

Sun Choi1

Received: 25 April 2015 / Accepted: 20 July 2015 / Published online: 25 July 2015
Ó The Pharmaceutical Society of Korea 2015

Abstract Drug discovery utilizes chemical biology and Keywords Computer-aided drug design
Virtual
computational drug design approaches for the efficient screening
QSAR
Pharmacophore
identification and optimization of lead compounds.
Chemical biology is mostly involved in the elucidation of
the biological function of a target and the mechanism of Introduction
action of a chemical modulator. On the other hand, com-
puter-aided drug design makes use of the structural Bringing a new drug into the market is a costly process in
knowledge of either the target (structure-based) or known terms of money, manpower, and time. Drug discovery and
ligands with bioactivity (ligand-based) to facilitate the development takes an average of 10–15 years with an
determination of promising candidate drugs. Various vir- approximate cost of US$800 million (DiMasi et al. 2003;
tual screening techniques are now being used by both Song et al. 2009) to US$1.8 billion (Paul et al. 2010).
pharmaceutical companies and academic research groups Innovations in combinatorial chemistry that led to the
to reduce the cost and time required for the discovery of a increase of compound databases covering large chemical
potent drug. Despite the rapid advances in these methods, spaces aided in the expansion of drug discovery and the
continuous improvements are critical for future drug dis- development of high-throughput screening (HTS) (Jhoti
covery tools. Advantages presented by structure-based and et al. 2013; Lavecchia and Di Giovanni 2013). Despite this,
ligand-based drug design suggest that their complementary the number of new molecular entities (NMEs) successfully
use, as well as their integration with experimental routines, launched into the market continued to decrease over the
has a powerful impact on rational drug design. In this last several years (Paul et al. 2010). To this end, employ-
article, we give an overview of the current computational ment of computer-aided drug discovery (CADD) tech-
drug design and their application in integrated rational drug niques by top pharmaceutical companies and other research
development to aid in the progress of drug discovery groups became essential for the preliminary stage of drug
research. discovery to expedite the drug development process in a
more cost-efficient way and to minimize failures in the
final stage. Recent successes involving rational drug design
have been reported elsewhere (Shoichet et al. 2002; Rey-
Stephani Joy Y. Macalino and Vijayakumar Gosu have contributed nolds 2014; Rodrigues and Schneider 2014). While HTS
equally. continues to be a big part of the drug discovery process in
the pharmaceutical industry due to its high success rate,
& Sun Choi
[email protected] lack of primary understanding of the molecular mechanism
behind the activity of the identified hits can hamper the
1
National Leading Research Laboratory of Molecular search of promising candidates (Macarron et al. 2011;
Modeling and Drug Design, College of Pharmacy and
Lionta et al. 2014). The use of rational drug design, as
Graduate School of Pharmaceutical Sciences, and Global Top
5 Research Program, Ewha Womans University, applied in CADD, provides a knowledge-driven approach
Seoul 120-750, Korea that can yield valuable information about the interaction

123
Role of computer-aided drug design in modern drug discovery 1687

pattern between protein and ligand (complex), as well as of CADD approaches continued to improve and evolve
the binding affinity. Furthermore, the availability of separately. However, combining different structure-based
supercomputers, parallel processing, and advanced soft- and ligand-based design strategies in the drug discovery
wares have greatly facilitated the rate of lead identification effort have been established to be more effective than any
in pharmaceutical research. single approach since both methods are able to complement
The current scenario of the drug discovery process their strengths and weaknesses (Prathipati et al. 2007;
involves several disciplines such as chemical and structural Grinter and Zou 2014; Lionta et al. 2014). Here, we sum-
biology, computational chemistry, organic synthesis, and marize the recent advances in computational drug design
pharmacology. Accordingly, it is comprised of several that could provide insights to modern drug discovery.
stages: (a) Target identification involves the discovery and
isolation of individual targets to investigate their functions
and association with a specific disease (Anderson 2003). Chemical biology in target identification
(b) Target validation is the stage where the drug target is and validation
linked to the disease of interest, as well as their capacity to
regulate biological functions in the body after binding to a Target identification can be performed either at the start
partner molecule. Numerous studies are performed to (target-based or reverse chemical genetics) or at the end
ascertain that the target macromolecule is linked to the (phenotype-based or forward chemical genetics) of a bio-
diseased state (Chen and Chen 2008). (c) Lead identifica- logical screening. Direct biochemical, genetic interaction,
tion entails the discovery of a synthetic chemical that and computational inference techniques are used to verify
shows a degree of potency and specificity against a bio- the protein target involved in the biological pathway being
logical target and is assumed to have the makings of a drug studied. Direct biochemical approach pinpoints the target
that can cure the intended disease (Kalyaanamoorthy and of a small molecule via biochemical affinity purification
Chen 2011). (d) Lead optimization covers the improvement (Burdine and Kodadek 2004). Enzyme assays and other
of potency and other significant properties through iterative biochemical tests can provide beneficial information for
cycles of evaluation of the lead compound(s) and their lead optimization when target structure information is
analogs. Thus, both in vitro and in vivo experiments are available. In addition, numerous DNA and RNA analysis
conducted to prioritize and select candidates with optimum methods permit target identification using genetic or
potential for development as a safe and efficient drug. genomic methods to measure or modify the presence and
Moreover, structure–activity relationships (SARs) are function of protein targets in a controlled model (Schenone
developed to determine pertinent pharmacokinetic and et al. 2013). Computational inference methods are also able
pharmacodynamic properties that can be applied to analogs to predict the target of identified inhibitors using structure-
that will be synthesized for evaluation (Andricopulo et al. based and profiling (ligand-based) methods. In particular,
2009). (e) Preclinical stage involves drug synthesis and these methods are often used for drug repositioning to
formulation research, in vivo animal studies for potency explain the off-target interactions in drug discovery
and toxicity, and characterization of mechanistic toxicity research. The methods mentioned above are essential in
(Cavagnaro 2002; Silverman 2004). (f) Clinical trials collecting information necessary to perform CADD and
include three phases that investigate safety, adverse side- have been extensively reviewed in a recent paper by
effects, dosage, efficacy, and pharmacokinetic and phar- Schenone (Schenone et al. 2013).
macological properties of the candidate drug on human
volunteers (Silverman 2004).
In general, modeling approaches are categorized into Structure-based drug design (SBDD)
structure-based and ligand-based methods (Fig. 1). The
structure-based approach consists of using the 3D structure In SBDD, the knowledge acquired from the binding site of
of the target (enzyme/receptor) for the generation or a 3D macromolecule structure is used to design and eval-
screening of potential ligands (modulators), followed by uate ligands based on their predicted interactions with the
synthesis, biological testing, and optimization. In contrast, protein binding site (Lavecchia and Di Giovanni 2013;
ligand-based approach consists of subjecting a collection of Grinter and Zou 2014). Thus, identification of a valid drug
molecules with diverse structures and known potency to target and the acquisition of its structural information are
computational modeling methods to develop theoretical the first vital steps in SBDD. Researches from structural
predictive models. These models are then used for struc- and computational biology aided in the generation of pro-
tural optimization to enhance potency and for identification tein structures with the use of X-ray crystallography,
of new chemical entities through virtual screening of a nuclear magnetic resonance (NMR), cryo-electron micro-
large chemical database. Over the decades, these two types scopy (EM), homology modeling, and molecular dynamic

123
1688 S. J. Y. Macalino et al.

Fig. 1 Representative workflow for computer-aided drug design

(MD) simulations (Anderson 2003; Kalyaanamoorthy and Binding site prediction or identification
Chen 2011). SBDD can be divided into two categories
(Kalyaanamoorthy and Chen 2011; Lionta et al. 2014): the The ideal binding site is a concave region containing sev-
de novo approach and the virtual screening approach. De eral chemical functionalities that interact with a ligand to
novo drug design exploits information from the 3D achieve the desired result (activation, modulation, or
receptor to find small fragments that match well with the inhibition) (Anderson 2003; Kalyaanamoorthy and Chen
binding site. These fragments should be linked based on 2011). Proteins co-crystallized with their substrates or
connection rules to ensure synthetic accessibility, provid- known inhibitors, as well as mutation studies identifying
ing a structurally novel ligand that can be synthesized for key residues for interaction, provide beneficial knowledge
further screening (Kutchukian and Shakhnovich 2010; in SBDD. However, when no information about the bind-
Rodrigues and Schneider 2014). On the other hand, virtual ing site is available, additional analyses are needed in order
screening (VS) uses available small molecule libraries to perform structure-based rational drug discovery. Cur-
(Table 1) to identify compounds with specific bioactivity rently, several in silico approaches have been reported in a
to act as replacements for existing ligands of target bio- number of papers and are available for the recognition of
molecules or to discover compounds for unexplored known binding regions in proteins (Laurie and Jackson 2006;
targets with available structural information (Andricopulo Zhang et al. 2011). The binding site of a small molecule
et al. 2009; Lavecchia and Di Giovanni 2013). compound can be predicted using available tools such as

123
 Table 1 Various software packages for computational drug design
Function Program/server Free/commercial Description Institute/company Websites

Binding CASTp Dundas et al. (2006) Free Uses weighted Delaunay triangulation University of Illinois at Chicago http://sts.bioe.uic.edu/castp/
site and the alpha complex for shape
prediction measurements
Cavitator Gao and Skolnick Free Pocket prediction using a grid based Center for the Study of Systems http://cssb.biology.gatech.edu/
(2013) geometric analysis Biology, Georgia Institute of Cavitator
Technology
ConCavity Capra et al. (2009) Free Based on combining evolutionary Princeton University http://compbio.cs.princeton.edu/
sequence and 3D structures concavity/
eFindSite Brylinski and Feinstein Free Common ligand binding site prediction Louisiana State University http://brylinski.cct.lsu.edu/efindsite
(2013) using set of evolutionary related
proteins
SiteComp Lin et al. (2012) Free Binding site comparison based on Mount Sinai School of Medicine http://scbx.mssm.edu/sitecomp/
molecular interaction fields sitecomp-web/Input.html
Docking AutoDock/AutoDock Vina Morris Free Flexible side chains (genetic algorithm) Scripps Research Institute http://autodock.scripps.edu/
et al. (2009), Trott and Olson
(2010)
Role of computer-aided drug design in modern drug discovery

BP-Dock Bolia et al. (2014) NA Flexible docking Arizona State University NA

cDocker Commercial Rigid body rotations followed by Biovia (formerly Accelrys) http://accelrys.com/
simulated annealing
Docking server Free/commercial Flexible side chains (Genetic algorithm) Virtua Drug ltd http://www.dockingserver.com/web
FLIPDock Zhao and Sanner Free for Flexible ligand and protein Scripps Research Institute http://flipdock.scripps.edu/
(2007) academic
usage
GOLD Hartshorn et al. (2007) Commercial Protein flexibility Astex Therapeutics, Ltd. and http://www.ccdc.cam.ac.uk/
Cambridge Crystallographic Data Solutions/GoldSuite/Pages/
Centre (CCDC) GOLD.aspx
Glide Commercial Ligand exhaustive search and flexible Schrödinger http://www.schrodinger.com/Glide
side chains
idock Hongjian et al. (2012), Li Free Flexible ligand docking Chinese University of Hong Kong http://istar.cse.cuhk.edu.hk/idock/
et al. (2014)
SwissDock Free Rigid ligand with less than 10 rotational Swiss Institute of Bioinformatics http://www.swissdock.ch/
bonds
VLifeDock Commercial GRID based, Genetic algorithm, GRIP vLifeÒ http://www.vlifesciences.com/
based docking products/Functional_products/
VLifeDock.php
1689

123
 Table 1 continued
1690

Function Program/server Free/commercial Description Institute/company Websites

123
Modeling Amber Free/commercial Simulation package Amber Development http://ambermd.org/
packages Team
BioSolveIT Commercial Molecular modeling package BioSolveIT http://www.biosolveit.de/
Desmond Free/commercial Molecular simulation package David. E. Shaw Research https://www.deshawresearch.com/resources_desmond.html
Discovery studio Commercial CHARMM molecular mechanics Biovia (formerly http://accelrys.com/products/discovery-studio/simulations.html
simulation program Accelrys)
Gromacs Pronk Free Simulation package for Gromacs Development www.gromacs.org
et al. (2013) macromolecules Team
Molecular Commercial Molecular modeling and simulation Chemical Computing https://www.chemcomp.com/MOE-Molecular_Modeling_and_
operating Group Simulations.htm
environment
NAMD Phillips Free High performance simulations for University of Illinois at http://www.ks.uiuc.edu/Research/namd/
et al. (2005) large biomolecular system Urbana-Champaign
SYBYL-X Commercial Molecular modeling and simulation Tripos International http://tripos.com/index.php?family=modules,SimplePage,,,&page=
package SYBYL-X&s=a100e3718f5d92a7690270f9f204739c
Yasara dynamics Commercial Molecular graphics, modeling and YASARA http://www.yasara.org/md.htm
simulation package
Databases DrugBank Free Detailed drug and target information DrugBank Development http://www.drugbank.ca/
Team
GLIDA Free GPCR ligand database Kyoto University http://pharminfo.pharm.kyoto-u.ac.jp/services/glida/index.php
PubChem Free Small molecule database NCBI https://pubchem.ncbi.nlm.nih.gov/
ZINC Free Database of commercially available University of California, http://zinc.docking.org/
compounds San Francisco
Most of the information about the listed programs can be found at http://www.click2drug.org/
S. J. Y. Macalino et al.
Role of computer-aided drug design in modern drug discovery 1691

PASS (Brady and Stouten 2000), Q-SiteFinder (Laurie and families and ligand series being considered (Wang et al.
Jackson 2005), CASTp (Dundas et al. 2006), LIGSITEcsc 2003; Warren et al. 2006; Cheng et al. 2009; Cross et al.
(Huang and Schroeder 2006), SiteMap (Halgren 2009), 2009; Xu et al. 2015). Combining the results from two or
FPocket (Le Guilloux et al. 2009), ConCavity (Capra et al. more scoring functions and forming a consensus has
2009), MED-SuMo (Doppelt-Azeroual et al. 2010), proved to be more beneficial in scoring and ranking com-
MDPocket (Schmidtke et al. 2011), FTMAP (Schmidtke pounds. Consensus scoring has superior accuracy, thereby
et al. 2011), POOL (Somarowthu and Ondrechen 2012), improving the probability of discovering true potential hits
and many others. Alternatively, specific approaches that (Cheng et al. 2009; Houston and Walkinshaw 2013).
can identify peptide binding sites include PepSite (Trabuco
et al. 2012), PeptiMap (Lavi et al. 2013), and PEP- Flexibility and solvent effects
SiteFinder (Saladin et al. 2014). Lastly, difficulties in
detecting allosteric sites can be alleviated by recently Difficulties experienced in scoring and ranking compounds
developed open-access web servers such as SPACER mainly arise from restraints and approximations applied to
(Goncearenco et al. 2013), MCPath (Kaya et al. 2013), conformational flexibility and solvation effects. The
Allosite (Huang et al. 2013), and PARS (Panjkovich and dynamic nature of proteins and ligands in a physiological
Daura 2014). A systematic assessment of a number of setting allows them to adopt a considerable number of
available web servers and stand-alone protein–ligand different conformations. However, incorporating molecular
binding site prediction programs was published previously flexibility in virtual screening of large chemical libraries is
(Chen et al. 2011). The report detailed that while these not feasible due to high computational costs (Anderson
methods can be valuable in finding putative binding sites, 2003; Feixas et al. 2014). To solve this, different strategies
the predictive quality of the algorithms used may be that can partially account for protein flexibility in docking
dependent on several factors, including template similarity calculations are being used in various software packages
and the size of the pocket (Chen et al. 2011). (Table 1): (a) ‘Soft scoring’ or the softening of van der
Waals potentials reduces steric energy penalties to permit
Docking and scoring some degree of overlap between ligand and target atoms
(Jiang and Kim 1991; Lavecchia and Di Giovanni 2013).
Molecular docking is one of the most well-known SBDD (b) Ensemble-based approach exploits multiple receptor
methods that predicts possible binding modes of a com- structures that are experimentally determined (X-ray crys-
pound in a particular target binding site and estimates tallography or NMR) or computationally produced using
affinity based on its conformation and complementarity molecular dynamics [relaxed complex scheme (RCS)
with the features found in the binding pocket (Cheng et al. (Amaro et al. 2008; Amaro and Li 2010)]. Instead of using
2012; Hung and Chen 2014). Accurately scoring the a single rigid structure, carefully chosen representative
binding and ranking of docked compounds is a crucial step structures accounting for conformational changes in pro-
in virtual screening. Frequently used scoring functions, as tein (backbone and/or side chain) is capable of yielding a
reported in several papers (Huang et al. 2010; Cheng et al. higher enrichment of potential compounds (Cheng et al.
2012; Lavecchia and Di Giovanni 2013; Liu and Wang 2008; Totrov and Abagyan 2008; Amaro and Li 2010;
2015), are divided into three categories: (a) Force field- Korb et al. 2012; Tuffery and Derreumaux 2012). A few
based functions calculate binding affinity derived from of the algorithms that use this approach include BP-Dock
physical atomic contacts in the target-ligand complex. Both (Bolia et al. 2014), MedusaDock (Ding and Dokholyan
solvation and entropy terms can also be evaluated. 2013), and RosettaBackrub (Lauck et al. 2010). (c) Induced
(b) Empirical-based functions employ simpler energy fit approach is based on the principle of induced fit effects
terms, such as that for hydrogen bonding and hydrophobic that take place upon binding of a ligand to a receptor. As
interactions, fitted to experimental binding affinity data. introduced by Koshland (Koshland 1958), the mechanism
(c) Knowledge-based functions derive binding energy from of induced fit happens after the initial formation of recep-
statistical analyses taken from a training set containing tor-ligand complex where the ligand ‘induces’ conforma-
protein–ligand atom pair frequencies. Numerous studies tional change in the protein, prompting a tighter binding
have compared different docking programs and their score (Sotriffer 2011; Feixas et al. 2014). (d) Molecular simu-
functions to evaluate their ability to correctly score and lation based on force fields depicting intra- and inter-
rank ligands that bind to a protein target. Even though molecular interactions between receptor and ligand is also
current scoring functions have greatly advanced, more possible. The downside is that protein flexibility is often
development is necessary to increase accuracy in hit constrained to the ligand binding site for computational
ranking. Performance of scoring functions varies depend- efficiency, leading to the occurrence of energy barriers in
ing on the purpose (scoring or ranking) and on the protein MD calculations. As a result, unrealistic conformations are

123
1692 S. J. Y. Macalino et al.

procured and computationally demanding simulations will variations in a set of compounds (Lavecchia and Di Gio-
be necessary to obtain an accurate complex structure (May vanni 2013; Melo et al. 2014). A statistical model is gen-
et al. 2008). The application of side-chain flexibility, which erated from this correlation to develop and mathematically
is critical in receptor flexibility, permits a certain measure predict the biological property of novel compounds (Melo
of freedom in the binding region of the target (Naj- et al. 2014). Several restrictions are required to generate a
manovich et al. 2000; Bostrom et al. 2006; Waszkowycz reliable QSAR model: (a) the bioactivity data should be of
et al. 2011). Even so, backbone displacement should also sufficient number (minimum of 20 compounds with
be considered since it can strengthen the accuracy of side activity) and acquired from a common experimental pro-
chain orientations (Bolia et al. 2014). Existing methods that tocol such that the potency values are comparable,
incorporate partial protein flexibility include Glide, (b) proper selection of compounds for the training and test
RosettaLigand (Davis and Baker 2009), FLIPDock (Zhao sets, (c) molecular descriptors for the ligands should have
and Sanner 2007), and FITTED (Corbeil et al. 2007, 2008; no autocorrelation to avoid over-fitting, (d) the model
Corbeil and Moitessier 2009; Englebienne and Moitessier should be validated using internal and/or external valida-
2009a, b; Pottel et al. 2014; Therrien et al. 2014). tion to determine its applicability and predictivity (Cher-
Solvation also plays a crucial role in ligand binding kasov et al. 2014; Melo et al. 2014). Table 2 lists some of
stability by mediating the interaction between receptor and the known programs or web servers commonly used for
ligand. Entropic and enthalpic contributions of the solvent QSAR studies.
via displacement of structural water have also been noted Comparative molecular field analysis (CoMFA) (Cramer
to enhance binding affinity (Abel et al. 2008; Yang et al. et al. 1988), which was established more than three decades
2013). Inclusion of explicit solvent assists in accurately ago, is still one of the most widely used 3D-QSAR method.
predicting the binding mode of ligands in docking studies, More recent 3D-QSAR strategies include Topomer
especially in complexes where water is required for CoMFA (Cramer 2003), spectral structure activity rela-
biomolecular recognition (Lie et al. 2011; Liu et al. 2013). tionship (S-SAR) (Putz and Lacrama 2007), adaptation of
Still, preliminary research regarding the number and the fields for molecular comparison (AFMoC) (Gohlke and
location of relevant structural and bridging water mole- Klebe 2002), and comparative residue interaction analysis
cules is necessary to avoid misleading results (Roberts and (CoRIA) (Datar et al. 2006). Despite its notable successes
Mancera 2008; Thilagavathi and Mancera 2010). Recently in the drug discovery field, 3D-QSAR still has numerous
developed methods that take into account solvation con- shortcomings that can be solved by more advanced multi-
tributions include WaterMap (Abel et al. 2008; Yang et al. dimensional QSAR strategies in the form of 4D, 5D, and
2013) and AcquaAlta (Rossato et al. 2011). 6D-QSAR. 4D-QSAR was developed to address ligand
conformation and orientation in the target binding site,
while 5D-QSAR incorporates issues like receptor flexibil-
Ligand-based drug design (LBDD) ity and induced fit effects. Finally, 6D-QSAR takes note of
the solvation effects to include its critical role in receptor-
In cases where 3D structure of the target protein is lacking, ligand interaction (Damale et al. 2014). Advances in
information taken from a set of ligands active against a computational power and software performance has also
relevant target (receptor or enzyme) can be used to identify been applied in improving QSAR model development and
significant structural and physicochemical properties validation through Discovery Bus (Cartmell et al. 2005)
(molecular descriptors) responsible for the observed bio- and AutoQSAR (Davis and Wood 2013). In both methods,
logical activity. Here, there is an assumption that struc- hundreds of highly predictive statistical models can be
turally similar compounds display similar biological objectively discovered, updated, and validated by contin-
response and interaction with the target (Prathipati et al. uously integrating new machine learning agents and
2007). The compound set should encompass a wide range descriptors into the system.
of concentration (at least 4 orders of magnitude) (Melo
et al. 2014) to generate a reliable ligand-based screening Pharmacophore modeling
model. Common ligand-based design techniques are
quantitative structure–activity relationships (QSARs) and Pharmacophore screening aims to identify compounds
pharmacophore-based methods. containing different scaffolds, but with a similar 3D
arrangement of key interacting functional groups (Vuori-
Quantitative structure–activity relationship (QSAR) nen and Schuster 2015). Binding site information can be
incorporated into the pharmacophore model by exploiting
QSAR studies are based on the premise that changes in the bioactive conformation of candidate compounds.
bioactivity are associated with structural and molecular Pharmacophore modeling is also often performed in the

123
Role of computer-aided drug design in modern drug discovery 1693

Table 2 Programs for molecular descriptor calculation or QSAR model analysis

Function Program/server Free/commercial Description Institute/company Websites

QSAR McQSAR Free Uses genetic function approximation Åbo Akademi University http://users.abo.fi/
(Vainio and paradigm mivainio/mcqsar/
Johnson 2005) index.php
SYBYL-X Commercial Includes modeling and QSAR packages Tripos International http://www.certara.
(CoMFA, HQSAR and Topomer com/products/
CoMFA) molmod/sybyl-x/
qsar/
MOLFEAT Free Computes molecular fingerprints and National University of http://jing.cz3.nus.edu.
descriptors derived from published Singapore sg/cgi-bin/molfeat/
QSAR models molfeat.cgi
Open3DQSAR Free Generates chemometric analysis of University of Turin http://open3dqsar.
molecular interaction fields (MIFs) sourceforge.net/
E-Dragon Free Computes molecular descriptors for Helmholtz Zentrum http://www.vcclab.
structure–activity or structure–property Muenchen Institute of org/lab/edragon/
relationship studies Structural Biology
Most of the information about the listed programs can be found at http://www.click2drug.org/

molecular alignment stage of QSAR modeling studies targets that have limited or no receptor and ligand infor-
(Melo et al. 2014). Several commonly used programs for mation available (Klabunde et al. 2009).
automatic pharmacophore generation include Discovery
Studio, PHASE (Dixon et al. 2006a, b), LigandScout
(Wolber and Langer 2005), and MOE. These softwares and Compound selection
other algorithms have already been extensively reviewed
(Vuorinen and Schuster 2015). Accordingly, generating a Generally, compound selection (or ‘cherry picking’) is
model with well-balanced sensitivity and specificity is done before proceeding to in vitro and in vivo assessment
important to reduce false negative and false positive of lead potency. As mentioned in the Docking and Scoring
results, respectively. Spatial constraints can be employed in section, one of the serious limitations of the current scoring
areas occupied by inactive compounds and refined to avoid functions include incorrectly ranking compounds, which
making the model too restrictive. Furthermore, features not lead to difficulty in identifying true hits. Apart from re-
consistently observed in active compounds should be made scoring poses and creating a consensus list of hits (Lionta
optional or removed from the model. After model refine- et al. 2014), an alternative strategy is to assess the inter-
ment, validation studies must be performed to determine action of a ligand in the binding pocket and identify
the sensitivity and specificity of a model against an external compounds displaying similar interaction patterns (Wasz-
test set (Vuorinen and Schuster 2015). kowycz et al. 2011). The Automatic analysis of Poses using
In the absence of both receptor 3D information and a set Self-Organizing Map (AuPosSOM) (Bouvier et al. 2010)
of active ligands, it is possible to create sequence-derived can be employed for pose clustering and ranking of virtual
3D pharmacophore models. From the concept similar screening hits. Their strategy is based on the assumption
receptors can bind with similar ligands (Klabunde et al. that specific contacts between an active compound and its
2009; Vuorinen and Schuster 2015), Pharma3D can use target are necessary to display the desired activity. In
homology models and 3D crystal structures to detect addition, clustering is also often performed to search for
common sequence motifs for ligand biomolecular recog- common scaffolds among the hit compounds and to select a
nition in protein families and create a single-feature phar- representative compound per cluster. Evaluation of struc-
macophore database. This approach has been successfully turally diverse compounds is a more cost- and time-effi-
applied in virtual screening for GPCR (family A). cient manner to investigate a large chemical space
Sequences displaying the same motifs can theoretically (Vuorinen and Schuster 2015).
distinguish the same ligand functionality at an analogous
spatial location. With this, the sequence motifs linked to Specificity
specific single-feature pharmacophores are identified and
used to generate the 3D pharmacophore model. In spite of Target specificity is a vital criterion in the search of effi-
its limitations, this is an attractive technique for drug cient drugs. Nevertheless, frequent occurrence of false

123
1694 S. J. Y. Macalino et al.

positives due to aggregation, ligand promiscuity, and ChemBioServer (Athanasiadis et al. 2012) and Free
compound reactivity is still observed during the experi- ADMET Filtering-Drugs2 (FAF-Drugs2) (Lagorce et al.
mental evaluation of lead candidates, impeding drug dis- 2008, 2011) can be used to filter a large compound data-
covery and development. The use of surfactants in base or a list of potential leads. ChemBioServer has the
screening assays minimizes compound aggregation (Ryan capability to display compounds and graph molecular
et al. 2003; Feng and Shoichet 2006), while reactive properties, filter compounds based on different chemical
compounds are identified by using reactive group filters to qualities, steric clashes, and toxicity, perform substructure
improve hit list quality (Roche et al. 2002). Additionally, search, cluster compounds, and recommend a representa-
pan-assay interference compounds (PAINS) filter (Baell tive for each group (Athanasiadis et al. 2012). Alterna-
and Holloway 2010; Erlanson 2015) can take out com- tively, FAF-Drugs2 features various pre-defined filters that
pounds that are frequent hitters in various HTS experi- the user can choose from, like the ones mentioned above,
ments, and thus cannot be used as specific inhibitors. along with others such as central nervous system (CNS)
However, the use of PAINS should be considered carefully Filter (Jeffrey and Summerfield 2010) and reactive group
since this filter was developed based on a limited number filter. In addition to these, pharmacophore models gener-
of HTS data and may not be applicable to other screening ated from inhibitors that cause toxicity can also be used to
experiments (Vuorinen and Schuster 2015). identify compounds with unfavourable moieties (Anan-
thula et al. 2008). To address the issue of drug metabolism,
Absorption, distribution, metabolism, excretion, reactivity models such as those implemented in SMART-
and toxicity (ADMET) Cyp are helpful. SMARTCyp is a free web service and
downloadable program that predicts sites in 2D compound
High attrition rates due to poor pharmacokinetic profiles structures that are likely to undergo Phase I CYP450-me-
produced the need to determine the ADMET properties of diated metabolism. It calculates the reactivity of ligand
leads in the early stages of drug screening. However, fragments using quantum chemical computations and the
experimental evaluation of pharmacokinetic properties of accessibility of atoms in the molecule to determine pos-
millions of compounds is not a viable option in terms of sible sites of metabolism (Rydberg et al. 2010). Alterna-
money and time. Thus, in order to quickly assess the drug- tively, MetaSite (Cruciani et al. 2005) also makes use of a
likeness of a lead compound prior to extensive experi- similar algorithm to identify potential metabolic reactivity
mental testing, virtual screening can be employed to filter sites, but with 3D configuration of the compound as the
hits and eliminate compounds with undesirable qualities query input. Other ADMET filters and tools are listed in
(Bajorath 2002). Similar to QSAR, in silico ADMET filters Table 3.
are derived from chemical or molecular descriptors, and In the application of these in silico ADMET models, we
are used to predict drug-like characteristics of compounds. should keep in mind that these tools are more helpful in the
The simplest and most well-known models include Lip- qualitative analysis of hits or compound sets rather than
inski Rule of Five (Lipinski et al. 2001), Rule of Three for accurately predicting the quantitative values. These meth-
fragments (Congreve et al. 2003), and Veber rules (Veber ods are beneficial in the prioritization of an identified class
et al. 2002). Publicly available web servers like of compounds for in vitro or in vivo assessment or

Table 3 Programs for prediction of ADMET properties

Function Program/server Free/commercial Description Institute/company Websites

ADMET QikProp Commercial Rapid identification of ADME Schrödinger http://www.schrodinger.com/

properties properties of drug candidates
ADMET Commercial Estimates the number of ADMET Simulations Plus, http://www.simulations-plus.com/
predictor properties from query molecular Inc. Products.aspx?grpID=1&cID=
structure 11&pID=13
ADMET and Commercial Predicts ADMET properties Biovia (formerly http://accelrys.com/products/
predictive Accelrys) discovery-studio/admet.html
toxicology
FAF-Drugs2 Free Subjects compounds to in silico University of http://www.mti.univ-paris-diderot.
ADMET filters Paris Diderot fr/recherche/plateformes/
logiciels#
Most of the information about the listed programs can be found at http://www.click2drug.org/

123
Role of computer-aided drug design in modern drug discovery 1695

evaluation of a particular descriptor and SAR (Gleeson and While both SBDD and LBDD were successfully utilized
Montanari 2012). individually in the studies mentioned above, relying solely
on one approach greatly limits the probability of finding a
plausible lead. Integration of these two methods in a drug
Applications of computational drug design discovery study can provide better and more extensive
information in the modeling of innovative drug candidates
Several CADD studies have been reported in the past years. against various diseases. This can be observed in a recent
Here, we briefly illustrate selected studies that apply virtual computational study involving the transient receptor
screening tools in drug discovery. potential vanilloid type 1 (TRPV1), wherein the combi-
In 2014, a study by Gao (2014) highlighted the structure- nation of homology modeling, pharmacophore filtering,
based rational drug design against Tip60 histone acetyl- and molecular docking identified a number of potential
transferase, an attractive target for cancer drug discovery. In antagonists. Initially, they generated a pharmacophore
their study, the 3D structure of the human Tip60 acetyl- model from known TRPV1 antagonists and used it to filter
transferase domain was retrieved from PDB for modeling the NCI database before performing docking experiments
experiments. However, since several key residues were against the hTRPV1 homology model. From the docking
missing in the structure, homology modeling was employed results, they selected the top-scoring compounds for further
using the human Tip60 sequence for the target and the experimental testing, yielding novel antagonists for
incomplete human Tip60 crystal structure as the template. hTRPV1 (Feng et al. 2015). In addition to assisting in the
They used Alpha Site Finder to find the binding site into selection of potential leads for a given target, computa-
which Pentamidine (PNT, known inhibitor) and acetyl-CoA tional tools can also provide viable interaction hypotheses
(natural substrate) were later docked. Subsequently, PNT for experimentally validated inhibitors of a target disease.
derivatives were computationally generated and docked to As an example, our group conducted molecular modeling
the Tip60 model to determine their binding affinities. MD studies of potent DNA methyltransferase (DNMT) inhibi-
simulations were also implemented to account for explicit tors, SGI-1027 and CBC12, to propose the binding modes
water molecules and flexibility. Finally, they identified and of these compounds and give a possible explanation for
experimentally validated TH1834 as a potential lead for the their observed activity in vitro. Remarkably, the binding
inhibition of Tip60 activity. As for studies focused on scores obtained for SGI-1027 were in excellent agreement
ligand-based techniques, our laboratory performed 3D- with the published experimental results, validating the
QSAR analyses of heterocyclic quinones to investigate their docking models. Moreover, the docking result of CBC12
inhibitory activity on vascular smooth muscle cell (SMC) corroborates the proposed inhibitory mechanism for
proliferation (Ryu et al. 2008) and their cytostatic activity DNMTs which suggests the use of ‘‘long’’ scaffolds in the
(Lee et al. 2009). In both studies, we used genetic algorithm design of DNMT inhibitors (Yoo et al. 2013).
with linear assignment of hypermolecular alignment of
database (GALAHAD) to generate and refine the pharma-
cophore model for molecule alignment, and CoMFA and Integration of drug discovery tools
CoMSIA to relate the molecular properties of the com-
pounds with their observed activities. In the first study, our In spite of the apparent advantages of virtual screening, it
group determined the inhibitory activity of known hetero- must be noted that it is not without its own pitfalls (Scior
cyclic quinone inhibitors for SMC proliferation and utilized et al. 2012; Kar and Roy 2013). Assimilating both SBDD
3D-QSAR in order to obtain and study their 3D molecular and LBDD may be necessary to satisfy all the practical
contour maps. The information acquired from this research requirements in identifying a promising lead. Moreover,
can then be employed for the design of more potent SMC combining virtual screening and HTS tools can be used to
proliferation inhibitors (Ryu et al. 2008). For the second form an efficient drug discovery workflow to overcome the
study, we examined the cytostatic activity of heterocyclic limitations presented by either approach. Different inte-
quinones by utilizing semi-empirical calculations and the gration workflows (Polgar and Keseru 2011; Tanrikulu
resulting LUMO energy to optimize the structures and et al. 2013) include: (a) Parallel integration consists of
compute the potential for the one-electron reduction of either simultaneous use of multiple virtual screening pro-
quinones, respectively. Our study provided reasonable tocols before HTS and experimental validation or
correlation between the cytotoxic activity of the compounds employing both virtual screening and HTS protocols in
with their calculated reduction potential, yielding 3D- parallel. This method can provide enriched hit rates since it
QSAR models and contour maps that can be used to design was often observed that different hits (from a different
compounds with reduced cytotoxic activity (Lee et al. chemical space) are obtained from different virtual
2009). screening protocols (Clark et al. 2004; Pirard et al. 2005;

123
1696 S. J. Y. Macalino et al.

Tomori et al. 2012), as well as from HTS (Doman et al. evaluation is necessary, LY-517717 remains a promising
2002; Polgar et al. 2005). However, this approach may not candidate. In another study, parallel integration of virtual
be appropriate for some studies as the list of compounds and high-throughput screening was applied in the discovery
obtained through this method is often large in number. of potential inhibitors against the HIV-1 Nef protein.
(b) Iterative or Sequential integration consists of combin- Computational studies (i.e. drug-like filtering, docking, and
ing computational and experimental tools to continually pharmacophore screening) were performed in conjunction
improve the selectivity of the screening workflow. with high-throughput screening assays for the Diversity
Sequential virtual screening methods use successive filters compound library. In both methods, PubChem CID 308963
in the research process to shrink the number of compounds was identified as the most promising inhibitor (Betzi et al.
before experimental evaluation (Prathipati et al. 2007; 2007).
Drwal and Griffith 2013; Kumar and Zhang 2015). Here,
the last step before biological testing of a compound is
mostly done by visual inspection (includes ligand binding Conclusion
and conformation in the active site or the shape of the
pharmacophore) and compound selection (Kumar and Computer aided drug design is a powerful tool in the search
Zhang 2015). Another way is subjecting virtual hits to of promising drug candidates, particularly when used in
experimental validation, after which the information tandem with current chemical biology screening tech-
acquired from in vitro screening will be used to optimize niques. Despite the fact that CADD makes use of several
the in silico model (Hofmann et al. 2008; Zander et al. restrictions and approximations, this knowledge-driven
2010). Theoretically, this will produce more potent hits approach has become an essential part in the drug design
against the chosen target. (c) Focused integration uses process due to its ability to fast-track drug discovery by
computational methods as a pre-filtering technique to utilizing existing knowledge and theories on receptor-li-
eliminate incompatible and unfavorable molecules and gand interactions, energy and structural optimization, and
create a focused library for experimental screening (Kiss synthesis. Nonetheless, there is still room for further
et al. 2008). Several other methods can be applied upon improvements, especially in the case of scoring functions,
integration of virtual screening tools, such as interaction- incorporating molecular flexibility and solvent effects,
based, pharmacophore-based, and similarity-based targeting receptors with little to no structural information,
approaches (Wilson and Lill 2011). Certainly, the incor- and increasing computational efficiency. To enhance cur-
poration of computational drug design methods in any rent in silico tools, continuous developments in the field of
stage of the drug discovery process allows great informa- chemical and structural biology, bioinformatics, and com-
tion evolution that can lead to better and more desirable putational technology is necessary. With this, the numerous
drug candidates. shortcomings of virtual drug discovery methods can be
The discovery and development of LY-517717 (Eli alleviated and the full potential of computational drug
Lilly) for the prevention of venous thromboembolism in design can be achieved.
hip or knee surgery patients is a useful example of drug
discovery integration, combining in silico de novo drug Acknowledgments This work was supported by the grant from
National Leading Research Lab (NLRL) program (2011-0028885)
design, structure-based lead optimization, iterative syn- through the Ministry of Science, ICT & Future Planning (MSIP) and
thesis, and assay studies (Jones et al. 2001; Liebeschuetz National Research Foundation of Korea (NRF).
et al. 2002; Devabhakthuni et al. 2015). Benzamidine
analogs were identified as potent factor Xa inhibitors and
subsequent optimization of this compound class provided
essential structure-based information. Similar to synthetic
References
combinatorial chemistry, a series of factor Xa focused Abel, R., T. Young, R. Farid, B.J. Berne, and R.A. Friesner. 2008.
libraries were generated using de novo method in the Role of the active-site solvent in the thermodynamics of factor
PRO_SELECT program (Liebeschuetz et al. 2002). Based Xa ligand binding. Journal of the American Chemical Society
on the identified critical binding site interactions, the 130: 2817–2831.
Agnelli, G., S. Haas, J.S. Ginsberg, K.A. Krueger, A. Dmitrienko, and
benzamidine analogs were optimized by changing and J.T. Brandt. 2007. A phase II study of the oral factor Xa inhibitor
adding substituents while measuring the fit to the protein LY517717 for the prevention of venous thromboembolism after
active site. In the end, the benzamidine moiety was hip or knee replacement. Journal of Thrombosis and Haemosta-
replaced with an indole group to improve pharmacokinetic sis 5: 746–753.
Amaro, R.E., R. Baron, and J.A. Mccammon. 2008. An improved
properties, yielding LY-517717 with a Ki value of 5 nM. relaxed complex scheme for receptor flexibility in computer-
Phase II clinical trial studies for this compound has already aided drug design. Journal of Computer-Aided Molecular Design
finished (Agnelli et al. 2007), and while continued 22: 693–705.

123
Role of computer-aided drug design in modern drug discovery 1697

Amaro, R.E., and W.W. Li. 2010. Emerging methods for ensemble- Cheng, L.S., R.E. Amaro, D. Xu, W.W. Li, P.W. Arzberger, and J.A.
based virtual screening. Current Topics in Medicinal Chemistry Mccammon. 2008. Ensemble-based virtual screening reveals
10: 3–13. potential novel antiviral compounds for avian influenza neu-
Ananthula, R.S., M. Ravikumar, A.B. Pramod, K.K. Madala, and S.K. raminidase. Journal of Medicinal Chemistry 51: 3878–3894.
Mahmood. 2008. Strategies for generating less toxic P-selectin Cheng, T., Q. Li, Z. Zhou, Y. Wang, and S.H. Bryant. 2012.
inhibitors: Pharmacophore modeling, virtual screening and Structure-based virtual screening for drug discovery: A problem-
counter pharmacophore screening to remove toxic hits. Journal centric review. The AAPS Journal 14: 133–141.
of Molecular Graphics and Modelling 27: 546–557. Cheng, T., X. Li, Y. Li, Z. Liu, and R. Wang. 2009. Comparative
Anderson, A.C. 2003. The process of structure-based drug design. assessment of scoring functions on a diverse test set. Journal of
Chemistry and Biology 10: 787–797. Chemical Information and Modeling 49: 1079–1093.
Andricopulo, A.D., L.B. Salum, and D.J. Abraham. 2009. Structure- Cherkasov, A., E.N. Muratov, D. Fourches, A. Varnek, I.I. Baskin, M.
based drug design strategies in medicinal chemistry. Current Cronin, J. Dearden, P. Gramatica, Y.C. Martin, R. Todeschini, V.
Topics in Medicinal Chemistry 9: 771–790. Consonni, V.E. Kuz’min, R. Cramer, R. Benigni, C.H. Yang, J.
Athanasiadis, E., Z. Cournia, and G. Spyrou. 2012. ChemBioServer: Rathman, L. Terfloth, J. Gasteiger, A. Richard, and A. Tropsha.
A web-based pipeline for filtering, clustering and visualization of 2014. QSAR modeling: Where have you been? Where are you
chemical compounds used in drug discovery. Bioinformatics 28: going to? Journal of Medicinal Chemistry 57: 4977–5010.
3002–3003. Clark, D.E., C. Higgs, S.P. Wren, H.J. Dyke, M. Wong, D. Norman,
Baell, J.B., and G.A. Holloway. 2010. New substructure filters for P.M. Lockey, and A.G. Roach. 2004. A virtual screening
removal of pan assay interference compounds (PAINS) from approach to finding novel and potent antagonists at the melanin-
screening libraries and for their exclusion in bioassays. Journal concentrating hormone 1 receptor. Journal of Medicinal Chem-
of Medicinal Chemistry 53: 2719–2740. istry 47: 3962–3971.
Bajorath, J. 2002. Integration of virtual and high-throughput screen- Congreve, M., R. Carr, C. Murray, and H. Jhoti. 2003. A ‘rule of
ing. Nature Reviews Drug Discovery 1: 882–894. three’ for fragment-based lead discovery? Drug Discovery Today
Betzi, S., A. Restouin, S. Opi, S.T. Arold, I. Parrot, F. Guerlesquin, X. 8: 876–877.
Morelli, and Y. Collette. 2007. Protein protein interaction Corbeil, C.R., P. Englebienne, and N. Moitessier. 2007. Docking
inhibition (2P2I) combining high throughput and virtual screen- ligands into flexible and solvated macromolecules. 1. Develop-
ing: Application to the HIV-1 Nef protein. Proceedings of the ment and validation of FITTED 1.0. Journal of Chemical
National Academy of Sciences of the United States of America Information and Modeling 47: 435–449.
104: 19256–19261. Corbeil, C.R., P. Englebienne, C.G. Yannopoulos, L. Chan, S.K. Das,
Bolia, A., Z.N. Gerek, and S.B. Ozkan. 2014. BP-Dock: A flexible D. Bilimoria, L. L’heureux, and N. Moitessier. 2008. Docking
docking scheme for exploring protein-ligand interactions based ligands into flexible and solvated macromolecules. 2. Develop-
on unbound structures. Journal of Chemical Information and ment and application of FITTED 1.5 to the virtual screening of
Modeling 54: 913–925. potential HCV polymerase inhibitors. Journal of Chemical
Bostrom, J., A. Hogner, and S. Schmitt. 2006. Do structurally similar Information and Modeling 48: 902–909.
ligands bind in a similar fashion? Journal of Medicinal Corbeil, C.R., and N. Moitessier. 2009. Docking ligands into flexible
Chemistry 49: 6716–6725. and solvated macromolecules. 3. Impact of input ligand confor-
Bouvier, G., N. Evrard-Todeschi, J.P. Girault, and G. Bertho. 2010. mation, protein flexibility, and water molecules on the accuracy
Automatic clustering of docking poses in virtual screening of docking programs. Journal of Chemical Information and
process using self-organizing map. Bioinformatics 26: 53–60. Modeling 49: 997–1009.
Brady Jr, G.P., and P.F. Stouten. 2000. Fast prediction and Cramer, R.D. 2003. Topomer CoMFA: A design methodology for
visualization of protein binding pockets with PASS. Journal of rapid lead optimization. Journal of Medicinal Chemistry 46:
Computer-Aided Molecular Design 14: 383–401. 374–388.
Brylinski, M., and W. Feinstein. 2013. eFindSite: Improved predic- Cramer, R.D., D.E. Patterson, and J.D. Bunce. 1988. Comparative
tion of ligand binding sites in protein models using meta- molecular field analysis (CoMFA). 1. Effect of shape on binding
threading, machine learning and auxiliary ligands. Journal of of steroids to carrier proteins. Journal of the American Chemical
Computer-Aided Molecular Design 27: 551–567. Society 110: 5959–5967.
Burdine, L., and T. Kodadek. 2004. Target identification in chemical Cross, J.B., D.C. Thompson, B.K. Rai, J.C. Baber, K.Y. Fan, Y.B. Hu,
genetics: The (often) missing link. Chemistry and Biology 11: and C. Humblet. 2009. Comparison of several molecular docking
593–597. programs: Pose prediction and virtual screening accuracy.
Capra, J.A., R.A. Laskowski, J.M. Thornton, M. Singh, and T.A. Journal of Chemical Information and Modeling 49: 1455–1474.
Funkhouser. 2009. Predicting protein ligand binding sites by Cruciani, G., E. Carosati, B. De Boeck, K. Ethirajulu, C. Mackie, T.
combining evolutionary sequence conservation and 3D structure. Howe, and R. Vianello. 2005. MetaSite: Understanding meta-
PLoS Computational Biology 5: e1000585. bolism in human cytochromes from the perspective of the
Cartmell, J., S. Enoch, D. Krstajic, and D.E. Leahy. 2005. Automated chemist. Journal of Medicinal Chemistry 48: 6970–6979.
QSPR through competitive workflow. Journal of Computer- Damale, M.G., S.N. Harke, F.A. Kalam Khan, D.B. Shinde, and J.N.
Aided Molecular Design 19: 821–833. Sangshetti. 2014. Recent advances in multidimensional QSAR
Cavagnaro, J.A. 2002. Preclinical safety evaluation of biotechnology- (4D-6D): A critical review. Mini Reviews in Medicinal Chem-
derived pharmaceuticals. Nature Reviews Drug Discovery 1: istry 14: 35–55.
469–475. Datar, P.A., S.A. Khedkar, A.K. Malde, and E.C. Coutinho. 2006.
Chen, K., M.J. Mizianty, J. Gao, and L. Kurgan. 2011. A critical Comparative residue interaction analysis (CoRIA): A 3D-QSAR
comparative assessment of predictions of protein-binding sites approach to explore the binding contributions of active site
for biologically relevant organic compounds. Structure 19: residues with ligands. Journal of Computer-Aided Molecular
613–621. Design 20: 343–360.
Chen, Y.P.P., and F. Chen. 2008. Identifying targets for drug Davis, A.M., and D.J. Wood. 2013. Quantitative structure-activity
discovery using bioinformatics. Expert Opinion on Therapeutic relationship models that stand the test of time. Molecular
Targets 12: 383–389. Pharmaceutics 10: 1183–1190.

123
1698 S. J. Y. Macalino et al.

Davis, I.W., and D. Baker. 2009. RosettaLigand docking with full Gao, M., and J. Skolnick. 2013. APoc: Large-scale identification of
ligand and receptor flexibility. Journal of Molecular Biology similar protein pockets. Bioinformatics 29: 597–604.
385: 381–392. Gleeson, M.P., and D. Montanari. 2012. Strategies for the generation,
Devabhakthuni, S., C.H. Yoon, and K.J. Pincus. 2015. Review of the validation and application of in silico ADMET models in lead
target-specific oral anticoagulants in development for the generation and optimization. Expert Opinion on Drug Metabo-
treatment and prevention of venous thromboembolism. The lism and Toxicology 8: 1435–1446.
Journal of Pharmacy Practice. doi:10.1177/0897190014568388. Gohlke, H., and G. Klebe. 2002. DrugScore meets CoMFA:
Dimasi, J.A., R.W. Hansen, and H.G. Grabowski. 2003. The price of Adaptation of fields for molecular comparison (AFMoC) or
innovation: New estimates of drug development costs. Journal of how to tailor knowledge-based pair-potentials to a particular
Health Economics 22: 151–185. protein. Journal of Medicinal Chemistry 45: 4153–4170.
Ding, F., and N.V. Dokholyan. 2013. Incorporating backbone Goncearenco, A., S. Mitternacht, T.P. Yong, B. Eisenhaber, F.
flexibility in MedusaDock improves ligand-binding pose predic- Eisenhaber, and I.N. Berezovsky. 2013. SPACER: Server for
tion in the CSAR2011 docking benchmark. Journal of Chemical predicting allosteric communication and effects of regulation.
Information and Modeling 53: 1871–1879. Nucleic Acids Research 41: W266–W272.
Dixon, S.L., A.M. Smondyrev, E.H. Knoll, S.N. Rao, D.E. Shaw, and Grinter, S.Z., and X. Zou. 2014. Challenges, applications, and recent
R.A. Friesner. 2006a. PHASE: A new engine for pharmacophore advances of protein-ligand docking in structure-based drug
perception, 3D QSAR model development, and 3D database design. Molecules 19: 10150–10176.
screening: 1. Methodology and preliminary results. Journal of Halgren, T.A. 2009. Identifying and characterizing binding sites and
Computer-Aided Molecular Design 20: 647–671. assessing druggability. Journal of Chemical Information and
Dixon, S.L., A.M. Smondyrev, and S.N. Rao. 2006b. PHASE: A Modeling 49: 377–389.
novel approach to pharmacophore modeling and 3D database Hartshorn, M.J., M.L. Verdonk, G. Chessari, S.C. Brewerton, W.T.
searching. Chemical Biology and Drug Design 67: 370–372. Mooij, P.N. Mortenson, and C.W. Murray. 2007. Diverse, high-
Doman, T.N., S.L. Mcgovern, B.J. Witherbee, T.P. Kasten, R. quality test set for the validation of protein-ligand docking
Kurumbail, W.C. Stallings, D.T. Connolly, and B.K. Shoichet. performance. Journal of Medicinal Chemistry 50: 726–741.
2002. Molecular docking and high-throughput screening for Hofmann, B., L. Franke, E. Proschak, Y. Tanrikulu, P. Schneider, D.
novel inhibitors of protein tyrosine phosphatase-1B. Journal of Steinhilber, and G. Schneider. 2008. Scaffold-hopping cascade
Medicinal Chemistry 45: 2213–2221. yields potent inhibitors of 5-lipoxygenase. ChemMedChem 3:
Doppelt-Azeroual, O., F. Delfaud, F. Moriaud, and A.G. De Brevern. 1535–1538.
2010. Fast and automated functional classification with MED- Hongjian L, Kwong-Sak L, Man-Hon W. idock: A multithreaded
SuMo: An application on purine-binding proteins. Protein virtual screening tool for flexible ligand docking. Computational
Science 19: 847–867. Intelligence in Bioinformatics and Computational Biology
Drwal, M.N., and R. Griffith. 2013. Combination of ligand- and (CIBCB), 2012 IEEE Symposium on, 9–12 May 2012, pp 77–84.
structure-based methods in virtual screening. Drug Discovery Houston, D.R., and M.D. Walkinshaw. 2013. Consensus docking:
Today: Technologies 10: e395–e401. improving the reliability of docking in a virtual screening
Dundas, J., Z. Ouyang, J. Tseng, A. Binkowski, Y. Turpaz, and J. context. Journal of Chemical Information and Modeling 53:
Liang. 2006. CASTp: Computed atlas of surface topography of 384–390.
proteins with structural and topographical mapping of function- Huang, B.D., and M. Schroeder. 2006. LIGSITE(csc): Predicting
ally annotated residues. Nucleic Acids Research 34: W116– ligand binding sites using the Connolly surface and degree of
W118. conservation. BMC Structural Biology 6(1): 19.
Englebienne, P., and N. Moitessier. 2009a. Docking ligands into Huang, S.Y., S.Z. Grinter, and X.Q. Zou. 2010. Scoring functions and
flexible and solvated macromolecules. 4. Are popular scoring their evaluation methods for protein-ligand docking: recent
functions accurate for this class of proteins? Journal of Chemical advances and future directions. Physical Chemistry Chemical
Information and Modeling 49: 1568–1580. Physics 12: 12899–12908.
Englebienne, P., and N. Moitessier. 2009b. Docking ligands into Huang, W.K., S.Y. Lu, Z.M. Huang, X.Y. Liu, L.K. Mou, Y. Luo,
flexible and solvated macromolecules. 5. Force-field-based Y.L. Zhao, Y.Q. Liu, Z.J. Chen, T.J. Hou, and J. Zhang. 2013.
prediction of binding affinities of ligands to proteins. Journal Allosite: A method for predicting allosteric sites. Bioinformatics
of Chemical Information and Modeling 49: 2564–2571. 29: 2357–2359.
Erlanson, D.A. 2015. Learning from PAINful lessons. Journal of Hung, C.L., and C.C. Chen. 2014. Computational approaches for drug
Medicinal Chemistry 58: 2088–2090. discovery. Drug Development Research 75: 412–418.
Feixas, F., S. Lindert, W. Sinko, and J.A. Mccammon. 2014. Jeffrey, P., and S. Summerfield. 2010. Assessment of the blood-brain
Exploring the role of receptor flexibility in structure-based drug barrier in CNS drug discovery. Neurobiology of Diseases 37:
discovery. Biophysical Chemistry 186: 31–45. 33–37.
Feng, B.Y., and B.K. Shoichet. 2006. A detergent-based assay for the Jhoti, H., S. Rees, and R. Solari. 2013. High-throughput screening and
detection of promiscuous inhibitors. Nature Protocols 1: structure-based approaches to hit discovery: Is there a clear
550–553. winner? Expert Opinion on Drug Discovery 8: 1449–1453.
Feng, Z.W., L.V. Pearce, X.M. Xu, X.L. Yang, P. Yang, P.M. Jiang, F., and S.H. Kim. 1991. ‘‘Soft docking’’: Matching of
Blumberg, and X.Q. Xie. 2015. Structural insight into tetrameric molecular surface cubes. Journal of Molecular Biology 219:
hTRPV1 from homology modeling, molecular docking, molec- 79–102.
ular dynamics simulation, virtual screening, and bioassay Jones, S.D., J.W. Liebeschuetz, P.J. Morgan, C.W. Murray, A.D.
validations. Journal of Chemical Information and Modeling Rimmer, J.M.E. Roscoe, B. Waszkowycz, P.M. Welsh, W.A.
55: 572–588. Wylie, S.C. Young, H. Martin, J. Mahler, L. Brady, and K.
Gao, C., E. Bourke, M. Scobie, M.A. Famme, T. Koolmeister, T. Wilkinson. 2001. The design of phenylglycine containing
Helleday, L.A. Eriksson, N.F. Lowndes, and J.A. Brown. 2014. benzamidine carboxamides as potent and selective inhibitors of
Rational design and validation of a Tip60 histone acetyltrans- factor Xa. Bioorganic and Medicinal Chemistry Letters 11:
ferase inhibitor. Scientific Reports 4: 5372. 733–736.

123
Role of computer-aided drug design in modern drug discovery 1699

Kalyaanamoorthy, S., and Y.P. Chen. 2011. Structure-based drug Li, H., K.S. Leung, P.J. Ballester, and M.H. Wong. 2014. istar: A web
design to augment hit discovery. Drug Discovery Today 16: platform for large-scale protein-ligand docking. PLoS ONE 9:
831–839. e85678.
Kar, S., and K. Roy. 2013. How far can virtual screening take us in Lie, M.A., R. Thomsen, C.N. Pedersen, B. Schiott, and M.H.
drug discovery? Expert Opinion on Drug Discovery 8: 245–261. Christensen. 2011. Molecular docking with ligand attached water
Kaya, C., A. Armutlulu, S. Ekesan, and T. Haliloglu. 2013. MCPath: molecules. Journal of Chemical Information and Modeling 51:
Monte Carlo path generation approach to predict likely allosteric 909–917.
pathways and functional residues. Nucleic Acids Research 41: Liebeschuetz, J.W., S.D. Jones, P.J. Morgan, C.W. Murray, A.D.
W249–W255. Rimmer, J.M. Roscoe, B. Waszkowycz, P.M. Welsh, W.A.
Kiss, R., B. Kiss, A. Konczol, F. Szalai, I. Jelinek, V. Laszlo, B. Wylie, S.C. Young, H. Martin, J. Mahler, L. Brady, and K.
Noszal, A. Falus, and G.M. Keseru. 2008. Discovery of novel Wilkinson. 2002. PRO_SELECT: Combining structure-based
human histamine H4 receptor ligands by large-scale structure- drug design and array-based chemistry for rapid lead discovery.
based virtual screening. Journal of Medicinal Chemistry 51: 2. The development of a series of highly potent and selective
3145–3153. factor Xa inhibitors. Journal of Medicinal Chemistry 45:
Klabunde, T., C. Giegerich, and A. Evers. 2009. Sequence-derived 1221–1232.
three-dimensional pharmacophore models for g-protein-coupled Lin, Y., S. Yoo, and R. Sanchez. 2012. SiteComp: A server for ligand
receptors and their application in virtual screening. Journal of binding site analysis in protein structures. Bioinformatics 28:
Medicinal Chemistry 52: 2923–2932. 1172–1173.
Korb, O., T.S. Olsson, S.J. Bowden, R.J. Hall, M.L. Verdonk, J.W. Lionta, E., G. Spyrou, D.K. Vassilatis, and Z. Cournia. 2014.
Liebeschuetz, and J.C. Cole. 2012. Potential and limitations of Structure-based virtual screening for drug discovery: Principles,
ensemble docking. Journal of Chemical Information and applications and recent advances. Current Topics in Medicinal
Modeling 52: 1262–1274. Chemistry 14: 1923–1938.
Koshland, D.E. 1958. Application of a theory of enzyme specificity to Lipinski, C.A., F. Lombardo, B.W. Dominy, and P.J. Feeney. 2001.
protein synthesis. Proceedings of the National Academy of Experimental and computational approaches to estimate solubil-
Sciences of the United States of America 44: 98–104. ity and permeability in drug discovery and development settings.
Kumar, A., and K.Y. Zhang. 2015. Hierarchical virtual screening Advanced Drug Delivery Reviews 46: 3–26.
approaches in small molecule drug discovery. Methods 71: Liu, J., X. He, and J.Z. Zhang. 2013. Improving the scoring of
26–37. protein-ligand binding affinity by including the effects of
Kutchukian, P.S., and E.I. Shakhnovich. 2010. De novo design: structural water and electronic polarization. Journal of Chemical
Balancing novelty and confined chemical space. Expert Opinion Information and Modeling 53: 1306–1314.
on Drug Discovery 5: 789–812. Liu, J., and R. Wang. 2015. Classification of current scoring
Lagorce, D., J. Maupetit, J. Baell, O. Sperandio, P. Tuffery, M.A. functions. Journal of Chemical Information and Modeling 55:
Miteva, H. Galons, and B.O. Villoutreix. 2011. The FAF-Drugs2 475–482.
server: A multistep engine to prepare electronic chemical Macarron, R., M.N. Banks, D. Bojanic, D.J. Burns, D.A. Cirovic, T.
compound collections. Bioinformatics 27: 2018–2020. Garyantes, D.V. Green, R.P. Hertzberg, W.P. Janzen, J.W.
Lagorce, D., O. Sperandio, H. Galons, M.A. Miteva, and B.O. Paslay, U. Schopfer, and G.S. Sittampalam. 2011. Impact of
Villoutreix. 2008. FAF-Drugs2: Free ADME/tox filtering tool to high-throughput screening in biomedical research. Nature
assist drug discovery and chemical biology projects. BMC Reviews Drug Discovery 10: 188–195.
Bioinformatics 9: 396. May, A., F. Sieker, and M. Zacharias. 2008. How to efficiently
Lauck, F., C.A. Smith, G.F. Friedland, E.L. Humphris, and T. include receptor flexibility during computational docking. Cur-
Kortemme. 2010. RosettaBackrub–a web server for flexible rent Computer-Aided Drug Design 4: 143–153.
backbone protein structure modeling and design. Nucleic Acids Melo, C.C., R.C. Braga, and C.H. Andrade. 2014. 3D-QSAR
Research 38: W569–W575. approaches in drug design: Perspectives to generate reliable
Laurie, A.T., and R.M. Jackson. 2005. Q-SiteFinder: An energy-based CoMFA models. Current Computer-Aided Drug Design 10:
method for the prediction of protein-ligand binding sites. 148–159.
Bioinformatics 21: 1908–1916. Morris, G.M., R. Huey, W. Lindstrom, M.F. Sanner, R.K. Belew, D.S.
Laurie, A.T., and R.M. Jackson. 2006. Methods for the prediction of Goodsell, and A.J. Olson. 2009. AutoDock4 and AutoDockTool-
protein-ligand binding sites for structure-based drug design and s4: Automated docking with selective receptor flexibility.
virtual ligand screening. Current Protein and Peptide Science 7: Journal of Computational Chemistry 30: 2785–2791.
395–406. Najmanovich, R., J. Kuttner, V. Sobolev, and M. Edelman. 2000.
Lavecchia, A., and C. Di Giovanni. 2013. Virtual screening strategies Side-chain flexibility in proteins upon ligand binding. Proteins-
in drug discovery: A critical review. Current Medicinal Chem- Structure Function and Bioinformatics 39: 261–268.
istry 20: 2839–2860. Panjkovich, A., and X. Daura. 2014. PARS: A web server for the
Lavi, A., C.H. Ngan, D. Movshovitz-Attias, T. Bohnuud, C. Yueh, D. prediction of protein allosteric and regulatory sites. Bioinfor-
Beglov, O. Schueler-Furman, and D. Kozakov. 2013. Detection matics 30: 1314–1315.
of peptide-binding sites on protein surfaces: The first step toward Paul, S.M., D.S. Mytelka, C.T. Dunwiddie, C.C. Persinger, B.H.
the modeling and targeting of peptide-mediated interactions. Munos, S.R. Lindborg, and A.L. Schacht. 2010. How to improve
Proteins-Structure Function and Bioinformatics 81: 2096–2105. R&D productivity: The pharmaceutical industry’s grand chal-
Le Guilloux, V., P. Schmidtke, and P. Tuffery. 2009. Fpocket: An lenge. Nature Reviews Drug Discovery 9: 203–214.
open source platform for ligand pocket detection. BMC Bioin- Phillips, J.C., R. Braun, W. Wang, J. Gumbart, E. Tajkhorshid, E.
formatics 10: 168. Villa, C. Chipot, R.D. Skeel, L. Kale, and K. Schulten. 2005.
Lee, Y., S. Kim, H.K. Rhee, K.E. Doh, J. Park, C.O. Lee, S. Choi, and Scalable molecular dynamics with NAMD. Journal of Compu-
H.Y.P. Choo. 2009. 3D-QSAR studies of cytotoxic heterocyclic tational Chemistry 26: 1781–1802.
quinones using calculated reduction potential. Drug Develop- Pirard, B., J. Brendel, and S. Peukert. 2005. The discovery of Kv1.5
ment Research 70: 438–444. blockers as a case study for the application of virtual screening

123
1700 S. J. Y. Macalino et al.

approaches. Journal of Chemical Information and Modeling 45: biology and drug discovery. Nature Chemical Biology 9:
477–485. 232–240.
Polgar, T., A. Baki, G.I. Szendrei, and G.M. Keseru. 2005. Schmidtke, P., A. Bidon-Chanal, F.J. Luque, and X. Barril. 2011.
Comparative virtual and experimental high-throughput screening MDpocket: Open-source cavity detection and characterization on
for glycogen synthase kinase-3 beta inhibitors. Journal of molecular dynamics trajectories. Bioinformatics 27: 3276–3285.
Medicinal Chemistry 48: 7946–7959. Scior, T., A. Bender, G. Tresadern, J.L. Medina-Franco, K. Martinez-
Polgar, T., and G.M. Keseru. 2011. Integration of virtual and high Mayorga, T. Langer, K. Cuanalo-Contreras, and D.K. Agrafiotis.
throughput screening in lead discovery settings. Combinatorial 2012. Recognizing pitfalls in virtual screening: A critical review.
Chemistry and High Throughput Screening 14: 889–897. Journal of Chemical Information and Modeling 52: 867–881.
Pottel, J., E. Therrien, J.L. Gleason, and N. Moitessier. 2014. Docking Shoichet, B.K., S.L. Mcgovern, B.Q. Wei, and J.J. Irwin. 2002. Lead
ligands into flexible and solvated macromolecules. 6. Develop- discovery using molecular docking. Current Opinion in Chem-
ment and application to the docking of HDACs and other zinc ical Biology 6: 439–446.
metalloenzymes inhibitors. Journal of Chemical Information and Silverman, R.B. 2004. The organic chemistry of drug design and drug
Modeling 54: 254–265. action, 2nd ed. Amsterdam: Elsevier Academic Press.
Prathipati, P., A. Dixit, and A.K. Saxena. 2007. Computer-aided drug Somarowthu, S., and M.J. Ondrechen. 2012. POOL server: machine
design: Integration of structure-based and ligand-based learning application for functional site prediction in proteins.
approaches in drug design. Current Computer-Aided Drug Bioinformatics 28: 2078–2079.
Design 3: 133–148. Song, C.M., S.J. Lim, and J.C. Tong. 2009. Recent advances in
Pronk, S., S. Pall, R. Schulz, P. Larsson, P. Bjelkmar, R. Apostolov, computer-aided drug design. Brief Bioinform 10: 579–591.
M.R. Shirts, J.C. Smith, P.M. Kasson, D. Van Der Spoel, B. Sotriffer, C.A. 2011. Accounting for induced-fit effects in docking:
Hess, and E. Lindahl. 2013. GROMACS 4.5: A high-throughput What is possible and what is not? Current Topics in Medicinal
and highly parallel open source molecular simulation toolkit. Chemistry 11: 179–191.
Bioinformatics 29: 845–854. Tanrikulu, Y., B. Kruger, and E. Proschak. 2013. The holistic
Putz, M.V., and A.M. Lacrama. 2007. Introducing spectral structure integration of virtual screening in drug discovery. Drug Discov
activity relationship (S-SAR) analysis. Application to ecotoxi- Today 18: 358–364.
cology. International Journal of Molecular Sciences 8: 363–391. Therrien, E., N. Weill, A. Tomberg, C.R. Corbeil, D. Lee, and N.
Reynolds, C.H. 2014. Impact of computational structure-based Moitessier. 2014. Docking ligands into flexible and solvated
methods on drug discovery. Current Pharmaceutical Design macromolecules. 7. Impact of protein flexibility and water
20: 3380–3386. molecules on docking-based virtual screening accuracy. Journal
Roberts, B.C., and R.L. Mancera. 2008. Ligand-protein docking with of Chemical Information and Modeling 54: 3198–3210.
water molecules. Journal of Chemical Information and Modeling Thilagavathi, R., and R.L. Mancera. 2010. Ligand-protein cross-
48: 397–408. docking with water molecules. Journal of Chemical Information
Roche, O., P. Schneider, J. Zuegge, W. Guba, M. Kansy, A. Alanine, and Modeling 50: 415–421.
K. Bleicher, F. Danel, E.M. Gutknecht, M. Rogers-Evans, W. Tomori, T., I. Hajdu, L. Barna, Z. Lorincz, S. Cseh, and G. Dorman.
Neidhart, H. Stalder, M. Dillon, E. Sjogren, N. Fotouhi, P. 2012. Combining 2D and 3D in silico methods for rapid selection of
Gillespie, R. Goodnow, W. Harris, P. Jones, M. Taniguchi, S. potential PDE5 inhibitors from multimillion compounds’ reposi-
Tsujii, W. Von Der Saal, G. Zimmermann, and G. Schneider. tories: Biological evaluation. Molecular Diversity 16: 59–72.
2002. Development of a virtual screening method for identifi- Totrov, M., and R. Abagyan. 2008. Flexible ligand docking to
cation of ‘‘frequent hitters’’ in compound libraries. Journal of multiple receptor conformations: A practical alternative. Current
Medicinal Chemistry 45: 137–142. Opinion in Structural Biology 18: 178–184.
Rodrigues, T., and G. Schneider. 2014. Flashback forward: Reaction- Trabuco, L.G., S. Lise, E. Petsalaki, and R.B. Russell. 2012. PepSite:
driven de novo design of bioactive compounds. Synlett 25: Prediction of peptide-binding sites from protein surfaces.
170–178. Nucleic Acids Research 40: W423–W427.
Rossato, G., B. Ernst, A. Vedani, and M. Smiesko. 2011. AcquaAlta: Trott, O., and A.J. Olson. 2010. AutoDock Vina: Improving the speed
a directional approach to the solvation of ligand-protein com- and accuracy of docking with a new scoring function, efficient
plexes. Journal of Chemical Information and Modeling 51: optimization, and multithreading. Journal of Computational
1867–1881. Chemistry 31: 455–461.
Ryan, A.J., N.M. Gray, P.N. Lowe, and C.W. Chung. 2003. Effect of Tuffery, P., and P. Derreumaux. 2012. Flexibility and binding affinity
detergent on ‘‘promiscuous’’ inhibitors. Journal of Medicinal in protein-ligand, protein-protein and multi-component protein
Chemistry 46: 3448–3451. interactions: Limitations of current computational approaches.
Rydberg, P., D.E. Gloriam, J. Zaretzki, C. Breneman, and L. Olsen. Journal of the Royal Society, Interface 9: 20–33.
2010. SMARTCyp: A 2D method for prediction of cytochrome Vainio, M.J., and M.S. Johnson. 2005. McQSAR: A multiconforma-
P450-mediated drug metabolism. ACS Medicinal Chemistry tional quantitative structure-activity relationship engine driven
Letters 1: 96–100. by genetic algorithms. Journal of Chemical Information and
Ryu, C.K., Y. Lee, S.G. Park, H.J. You, R.Y. Lee, S.Y. Lee, and S. Modeling 45: 1953–1961.
Choi. 2008. 3D-QSAR studies of heterocyclic quinones with Veber, D.F., S.R. Johnson, H.Y. Cheng, B.R. Smith, K.W. Ward, and
inhibitory activity on vascular smooth muscle cell proliferation K.D. Kopple. 2002. Molecular properties that influence the oral
using pharmacophore-based alignment. Bioorganic and Medic- bioavailability of drug candidates. Journal of Medicinal Chem-
inal Chemistry 16: 9772–9779. istry 45: 2615–2623.
Saladin, A., J. Rey, P. Thevenet, M. Zacharias, G. Moroy, and P. Vuorinen, A., and D. Schuster. 2015. Methods for generating and
Tuffery. 2014. PEP-SiteFinder: a tool for the blind identification applying pharmacophore models as virtual screening filters and
of peptide binding sites on protein surfaces. Nucleic Acids for bioactivity profiling. Methods 71: 113–134.
Research 42: W221–W226. Wang, R., Y. Lu, and S. Wang. 2003. Comparative evaluation of 11
Schenone, M., V. Dancik, B.K. Wagner, and P.A. Clemons. 2013. scoring functions for molecular docking. Journal of Medicinal
Target identification and mechanism of action in chemical Chemistry 46: 2287–2303.

123
Role of computer-aided drug design in modern drug discovery 1701

Warren, G.L., C.W. Andrews, A.M. Capelli, B. Clarke, J. Lalonde, Yang, Y., F.C. Lightstone, and S.E. Wong. 2013. Approaches to
M.H. Lambert, M. Lindvall, N. Nevins, S.F. Semus, S. Senger, efficiently estimate solvation and explicit water energetics in
G. Tedesco, I.D. Wall, J.M. Woolven, C.E. Peishoff, and M.S. ligand binding: The use of WaterMap. Expert Opinion on Drug
Head. 2006. A critical assessment of docking programs and Discovery 8: 277–287.
scoring functions. Journal of Medicinal Chemistry 49: Yoo, J., S. Choi, and J.L. Medina-Franco. 2013. Molecular modeling
5912–5931. studies of the novel inhibitors of DNA methyltransferases SGI-
Waszkowycz, B., D.E. Clark, and E. Gancia. 2011. Outstanding 1027 and CBC12: Implications for the mechanism of inhibition
challenges in protein-ligand docking and structure-based virtual of DNMTs. PLoS One 8: e62152.
screening. Wiley Interdisciplinary Reviews-Computational Zander, J., M. Hartenfeller, V. Hahnke, E. Proschak, S. Besier, T.A.
Molecular Science 1: 229–259. Wichelhaus, and G. Schneider. 2010. Multistep virtual screening
Wilson, G.L., and M.A. Lill. 2011. Integrating structure-based and for rapid and efficient identification of non-nucleoside bacterial
ligand-based approaches for computational drug design. Future thymidine kinase inhibitors. Chemistry 16: 9630–9637.
Medicinal Chemistry 3: 735–750. Zhang, Z.M., Y. Li, B.Y. Lin, M. Schroeder, and B.D. Huang. 2011.
Wolber, G., and T. Langer. 2005. LigandScout: 3-D pharmacophores Identification of cavities on protein surface using multiple
derived from protein-bound ligands and their use as virtual computational approaches for drug binding site prediction.
screening filters. Journal of Chemical Information and Modeling Bioinformatics 27: 2083–2088.
45: 160–169. Zhao, Y., and M.F. Sanner. 2007. FLIPDock: Docking flexible
Xu, W., A.J. Lucke, and D.P. Fairlie. 2015. Comparing sixteen ligands into flexible receptors. Proteins-Structure Function and
scoring functions for predicting biological activities of ligands Bioinformatics 68: 726–737.
for protein targets. Journal of Molecular Graphics and
Modelling 57: 76–88.

123