In Silico De Novo Synthesis, Screening, and ADME/T Profiling of

DNA-pA104R Inhibitors as Potential African Swine Fever

Therapeutics

Kim Rafaelle E. Reyes

, Timothy Jen R. Roxas

, Marineil C. Gomez

and Lemmuel L. Tayo

School of Chemical, Biological, and Materials Engineering and Sciences, Mapúa University,

658 Muralla Street, Intramuros, Manila, Philippines

Keywords: African Swine Fever virus, ADME/T, In Silico, De Novo, Drug-likeness, pA104R, Molecular Docking,

Ligand-based Virtual Screening.

Abstract: African Swine Fever Virus (ASFV) is a dsDNA virus causative of the African Swine Fever (ASF) in wild and

domestic hogs. ASF is characterized by hemorrhagic fever, high mortality, and transmissibility. The binding

of the DNA to the pA104R protein mediates viral replication and genome packaging. In the present study, we

generated nine (9) reference compounds that exhibited high docking affinities through de novo computer-

aided drug design (CADD). These compounds were then used as query molecules to find commercially

available drug-like compounds using ligand-based virtual screening (VS). We were able to retrieve 900 hit

compounds that exhibited the same pharmacophoric activities. Then, these hit compounds were subjected to

drug-likeness filtration to identify the most likely to be developed as commercial drugs based on established

parameters. We identified sixty-two (62) drug-like molecules. Molecular docking was then performed to

determine the top five compounds with the highest binding affinities against the target protein. ADME/T

profiling was done on these compounds to assess their pharmacokinetic properties. Compound 8.45 performed

best based on our devised scoring system. This paper shall serve as a good reference in the discovery and

development of anti-ASFV therapeutics.

1 INTRODUCTION

The African Swine Fever Virus (ASFV) is a highly

transmissible virus causative of the African Swine

Fever (ASF) in wild and domestic hogs. Apart from

its swift spread, ASF is characterized by high

mortality rates, to which death is usually observed a

week after the onset of the disease. The identification

of the viral infection is of little difficulty due to the

readily observable symptoms in infected pigs that

include (1) high fever, (2) reduced locomotor

movements, (3) lack of appetite, (4) huddling, (5)

conjunctivitis, (6) diarrhea and vomiting, (7)

somnolence, (8) dyspnea, (9) seizures, and (10) skin

hemorrhages (Blome et al., 2020). This virus's

https://orcid.org/0000-0002-2662-209X

https://orcid.org/0000-0002-6009-5334

https://orcid.org/0000-0001-6238-5709

https://orcid.org/0000-0002-0869-2131

transmission and promulgation rely on vector species

such as ticks that primarily target boars found in the

wilderness. ASFV has evolved from a very mild

strain into a highly transmissible virus that threatens

today's swine population (Chen et al., 2021).

Although tremendously virulent to hogs, there is no

risk of this virus being transmitted to humans and

cause the same threats that it poses for pigs. The virus

indirectly impacts society through the economy since

the meat of the domesticated pigs is often a central

ingredient in making food from all ranges of cuisine.

It is therefore of great importance to develop

therapeutics that could eliminate this virus. Up to this

date, there is no commercially available vaccine or

drug to combat ASF in infected animals. The

scientific community has only relied on control

Reyes, K., Roxas, T., Gomez, M. and Tayo, L.

In Silico De Novo Synthesis, Screening, and ADME/T Proﬁling of DNA-pA104R Inhibitors as Potential African Swine Fever Therapeutics.

DOI: 10.5220/0010774500003123

In Proceedings of the 15th International Joint Conference on Biomedical Engineering Systems and Technologies (BIOSTEC 2022) - Volume 3: BIOINFORMATICS, pages 15-26

ISBN: 978-989-758-552-4; ISSN: 2184-4305

strategies to confine the virus and prevent its further

spread.

Regarding its structure, ASFV is classified as a

nucleocytoplasmic large DNA virus (NCLDV),

having a genome size of 180,916 base pairs (Alonso

et al., 2018) and an overall virion diameter of 175-215

nm with a 70-100 nm diameter nucleoprotein core

(Blome et al., 2020). The core is surrounded by: (1)

an internal lipid layer, (2) an icosahedral capsid that

is composed of 1892-2172 capsomeres, and (3) the

dispensable lipid envelope (Alonso et al., 2018). This

virus is highly stable in environmental settings and

thrives in most and protein-rich areas. Furthermore, it

can survive in raw meat products at variable

durations. The structure and architecture of the virus

have not been fully elucidated. However, the

pA104R protein of the ASFV has successfully been

studied (Urbano & Ferreira, 2020), and its

crystallographic structure is available (Liu et al.,

2020). This macromolecule is a DNA-binding protein

essential for viral replication and transcription

(Frouco et al., 2017b). ASFV enters the host cell

through endocytosis and micropinocytosis. Then, the

virus uses the pA104R to interact with the host cell's

DNA, leading to the mass manufacture of the viral

parts via gene editing and the completion of the

replicative cycle of the ASFV (Galindo & Alonso,

2017). The structure of the pA104R is characterized

as histone-like since it shares conserved sequence

homology with the other histone-like proteins derived

from bacteria (Frouco et al., 2017a).

In a study by Liu et al., the researchers

successfully elucidated the structure and determined

the binding properties of the pA104R to the host

DNA (i.e., pA104R-DNA complex). Their findings

illustrated the unique binding pattern of pA104R as it

uses its β-ribbon arm and inserts it in the major

groove of the DNA. Furthermore, the researchers

evaluated the ability of the stilbene derivatives, SD1

and SD4, to inhibit viral replication by disrupting the

pA104R-DNA binding in swine macrophages (Liu et

al., 2020). Zhu et al. utilized protein-protein

interaction (PPI) networks to determine the ASFV-

interacting proteins and assessed some commercially

available drugs, such as Polaprezinc and

Geldanamycin, that could potentially bind to some

viral proteins to inhibit the action of the ASFV (Zhu

et al., 2020). A similar study by Mottola et al. helped

unmask the antiviral activity of fluoroquinolones

against the virus (Zhu et al., 2019). Recently, there

have been explorations on the potential of

antimicrobial peptides (AMPs) and their effect on

porcine viruses, including their mechanism of action

(Pen et al., 2020); however, the AMPs used were

already existing ones, and therefore, further

exploration on more efficient peptides must be done.

Although the research on ASF has reached great

strides in the past years, there is still no potential

candidate to eliminate or inhibit the effects of the

virus in question. It is therefore imperative to devise

new strategies that could identify compounds that

could neutralize ASFV.

Drug discovery and development is defined as the

process of identifying chemical entities that have

potential therapeutic effects (Mohs & Greig, 2017).

Over the years, this process has undergone radical

changes with the further integration of biology,

chemistry, physics, mathematics, and computer

science (Umashankar & Gurunathan, 2015). The

pipeline (i.e., drug discovery and development)

involves a multistage process that should be strictly

followed before a novel chemical entity is

commercially available for public consumption

(Mohs & Greig, 2017). Until the late 1980s, drug

discovery was solely based on blind screening and

serendipity (Kiriiri et al., 2020). This was changed

upon introducing high-throughput screening (HTS)

and combinatorial chemistry, allowing scientists to

discover and synthesize many compounds

(Umashankar & Gurunathan, 2015). However, the

methods above were very costly and could be

described as "brute force" approach as finding lead

candidates is dependent on the initial library of

compounds (Polanski, 2020). A further refinement of

the pipeline has emerged with the introduction of the

in silico approach. Such a strategy uses computational

methods to predict the binding properties of the

compound of interest to the biological target. The

Boston Consulting Group estimated that integrating

in silico practices in the drug discovery pipeline could

save 14% of the total cost (Agarwal et al., 2017).

Herein, we identified potential drug-like

compounds that can be utilized to treat African Swine

Fever (ASF) mainly through the inhibition of the

pA104R-DNA binding. To accomplish this, de novo

methods and a ligand-based virtual screening

approach were employed. The binding affinities of

the generated and retrieved compounds were

determined through molecular docking studies.

Finally, the pharmacokinetic properties of the

identified drug-like compounds were ADME/T

studies. This study shall only entail the identification

and pharmacokinetic characterization of the possible

hit compounds.

BIOINFORMATICS 2022 - 13th International Conference on Bioinformatics Models, Methods and Algorithms

2 METHODOLOGY

2.1 Protein Preparation

The crystallographic structure of ASFV pA104R in

complex with dsDNA was obtained from Protein

Data Bank (accession number: 6LMH). Protein

visualization and refinement were conducted in the

Biovia discovery studio visualizer (DASSAULT

SYSTEMES, 2020). pA104R contains two chains: A

and B (i.e., AHR and BDR, respectively). In this

study, only the AHR was considered, and so the chain

B was removed. The heteroatoms (i.e., dsDNA and

water molecules) were deleted and polar hydrogen

was added for the subsequent analyses. The pA104R-

DNA complex active site residues were shown in

Table 1.

Figure 1: In silico experimental design.

2.2 In Silico De Novo Synthesis of

Reference Molecules

To our knowledge, there has been no definitive report

of possible drug-like DNA-binding inhibitors of the

African swine fever virus. As such, we opted to

perform a de novo drug design approach to generate

potential ligands. This strategy uses computational

algorithms to build molecules that exhibit specified

properties. e-LEA3D (https://chemoinfo.ipmc.cnrs

.fr/LEA3D/index.html) is an online tool that enables

users to perform computer-aided de novo drug design

efficiently. This web server creates diverse

molecules through a genetic algorithm that evolves

fragments based on mutation and crossover operators

(Douguet, 2010).

The prepared protein structure was loaded to the

server. Then, the desired molecular properties of the

output molecules were selected using default

parameters. e-LEA3D uses the PLANTS docking

program to assign scores in the generated molecules.

The parameters used in this function were as follows:

binding site radius = 40, binding site residue =

LYS89, weight in final score = 1. The server returned

10 'reference' compounds that have a high binding

affinity towards pA104R. The experimental design

employed in this study was shown in Figure 1.

2.3 Ligand-Based Virtual Screening

(LBVS)

The de novo approach has one major flaw: molecules

generated through this strategy are hard to synthesize

(Mouchlis et al., 2021). To make up for this

drawback, we adapted ligand-based virtual screening

that could be used to search for commercially

available active compounds from several enormous

libraries of molecules. This approach is usually done

when there is no prior knowledge of the 3D structure

of the target protein (Hamza et al., 2012).

Nonetheless, ligand-based VS could also be

employed when searching for new ligands with

similar chemical and biological activities.

In this study, the USR-VS web server

(http://usr.marseille.inserm.fr/) was used. This tool

implements Ultrafast Shape Recognition (USR) and

Ultrafast Shape Recognition with CREDO Atom

Types (USRCAT) algorithms to screen a library for

similar compounds relative to the pharmacophoric

properties of the query molecule (Schreyer &

Blundell, 2012). Currently, the USR-VS screening

library is comprised of 23 million molecules with

over 94 million low-energy conformers. To conduct

the virtual screening, the structure data files (SDF) of

the nine (9) reference molecules were retrieved from

the e-LEA3D webserver.

In Silico De Novo Synthesis, Screening, and ADME/T Proﬁling of DNA-pA104R Inhibitors as Potential African Swine Fever Therapeutics

The screening process was straightforward. The

SDF files were uploaded into the webserver. Then,

the desired scoring algorithm (i.e. USR and

USRCAT) was selected. We used USRCAT since it

has been reported that this algorithm outperformed

USR in retrospective screening (Schreyer & Blundell,

2012). Clicking 'submit' initiates the virtual screening

that will only take milliseconds. Each query (i.e.

reference) molecule generated 100 hits. Once the

screening was finished, the results were displayed in

a separate internet tab and the SDF for each hit

compound was downloaded.

2.4 Drug-likeness Filtration

Drug-likeness filters are important aspects of drug

discovery. These parameters determined the

likelihood of a compound to exhibit therapeutic

effects while being biologically safe based on its

molecular structure. Moreover, applying filters to a

large library of compounds could eliminate the non-

drug-like molecules, thus saving operation time and

cost (Shen et al., 2012). In this study, a total of 900

hit compounds were identified through ligand-based

virtual screening. Five drug-like filters were

employed for the analysis (Table S1). The filtration

process was conducted using the SwissADME

website (http://www.swissadme.ch/index.php). This

tool is a versatile free-of-charge webserver for

determining the physicochemical properties,

lipophilicity, water-solubility, pharmacokinetics,

drug-likeness, and medicinal chemistry of small

molecules (Daina et al., 2017).

SwissADME only accepts chemical structures in

SMILES format (i.e. .smi extension). Thus, the hit

compounds (i.e., in SDF format) must be converted

into the acceptable file format before executing the

filtration process. To do this, the OpenBabel widget

of the PyRx version 0.8 (The Scripps Research

Institute, 2008) was used. After the said conversion,

the SMILES were uploaded to the webserver. All

compounds that passed the five filtration parameters

without any violations were selected for the

subsequent analyses. Sixty-two (62) unique drug-like

compounds were identified.

2.5 Molecular Docking Studies

The 62 drug-like compounds were subjected to

molecular docking studies to determine their affinity

for binding against pA104R. DockThor

(https://dockthor.lncc.br/v2/) is a user-friendly

webserver for receptor-ligand docking developed by

the GMMSB group (Santos et al., 2020). This tool

performs molecular docking through flexible-ligand

and rigid-receptor strategies based on the MMFF94S

force field (Guedes, Costa, et al., 2021). The docking

procedure is a three-step process. First, the prepared

protein (i.e. protonated) in PDB format was uploaded

to the server. Since no cofactors were considered in

this study, the 'do not use cofactor' function was

selected. Then, the 62 drug-like compounds in SDF

format were docking program requires the user to

upload the protonated version of the ligands. For

convenience, DockThor is embedded with an 'add H'

function. The submitted protein and ligand structure

were processed after clicking 'send'. A checkmark

appeared which indicated that the input molecules

were valid and recognized by the force field. The final

step involves setting up the docking configuration.

The user could choose from blind docking or user-

defined docking. Since the binding site was already

determined (Table 1), we performed user-defined

docking. DockThor utilizes a genetic algorithm to

determine the optimal poses for flexible ligand

docking (De Magalhães et al., 2014). Furthermore,

the platform allows the user to customize the

algorithm parameters, but the 'standard algorithm'

was selected for this study. Table S2 shows the

different parameters used in the docking experiment.

The webserver ranked the 62 drug-like compounds

based on their binding affinities. The chemical

structure (i.e., in .mol2 format) of the best docking

pose for each input molecule was obtained. The 3D

and 2D protein-ligand interactions were visualized

using Biovia discovery studio.

2.6 ADME/T Studies

Compounds must undergo ADME/T studies to

determine their pharmacokinetic properties and

safety level. The 62 drug-like compounds and the 9

reference molecules were then subjected to ADME/T

studies. To do so, we used pkCSM

(biosig.unimelb.edu.au/pkcsm/prediction), a web-

based tool commonly used to calculate the

pharmacokinetic properties of small-molecule drugs,

such as the compounds involved in this study. This

application allows for the fast development of

predictive models of central ADMET properties via

graph-based signatures (Pires et al., 2015). Since the

pkCSM webserver only accepts entries in the

SMILES format, we first converted the available files

to the .smi format via the OpenBabel widget of the

PyRx software, similar to the earlier method.

Subsequently, the compounds were uploaded to the

web-based server for the prediction of their

pharmacokinetic properties

BIOINFORMATICS 2022 - 13th International Conference on Bioinformatics Models, Methods and Algorithms

3 RESULTS AND DISCUSSION

3.1 The pA104R: A Therapeutic Target

for ASF

DNA packaging is a vital process in the life cycle of

double-stranded DNA (dsDNA) viruses. Packaging

proteins bind with the DNA and initiates

conformational changes that cause it to bend and be

organized into densely packed chromatin structures

(Urbano & Ferreira, 2020). Failure of these proteins

to promote condensation and packaging will

inevitably cause DNA damage, ultimately leading to

apoptosis (J. Y. J. Wang, 2001). Therefore, targeting

the proteins involved in the said process is an

attractive approach to design therapeutics against

viruses. DNA-packaging proteins have been reported

for a wide range of organisms. For instance, the

packaging proteins in bacteria are the histone-like

proteins belonging to the HU/IHF superfamily

(Swinger & Rice, 2004). In relevance to the ASFV,

p10 and pA104R are the major DNA-packaging

proteins in mature ASFV (Andrés et al., 2002).

However, a more recent study using small interfering

RNA (siRNA) has shown that pA104R has a

profound role in DNA replication, transcription, and

packaging of ASFV (Frouco et al., 2017a).

Figure 2: The structure of (A) DNA-pA104R complex

(PDB accession code:6LMJ) and (B) apo-pA104R (PDB

accession: 6LMH).

pA104R is a homodimer that significantly

resembles other bacterial HU/IHF homologs (Liu et

al., 2020). The crystallographic structures of the apo-

pA104R and DNA-bound pA104R were shown in

Figure 2. The protein folds into two domains, namely

the "body" AHR(i.e., alpha-helical region) and the

"arms" BDR (i.e., β-strand DNA binding region). As

shown in Fig 2B, the DNA interacts predominantly

with pA104R via the BDR arm. The surface of this

region is saturated with positively charged, making it

an attractive binding site for the negatively charged

DNA molecule (Liu et al., 2020). Thus, the

subsequent analyses were simplified by focusing on

the BDR. To design an effective inhibitor, the key

amino acid residues within the binding site must be

identified. Thus, the active site amino acid residues in

the DNA-pA104R complex were determined using

the Biovia Discovery Studio (Table 1). As

hypothesized, the key amino acid residues in the BDR

arm are mostly positively charged at physiological

pH. HIS78, LYS89, and LYS91 are all positively

charged. Meanwhile, PRO80 is an aliphatic amino

acid making it nonpolar and hydrophobic. This

residue interacts with the DNA strand, but the nature

of its interaction is not electrostatic since it is

nonpolar at physiological pH. Further studies are

encouraged to uncover the linkages between this

residue and the target DNA strand.

Table 1: Interacting amino acid of DNA-pA104R complex.

Only the AAs in the BDR were considered.

A104R domain Active residues

AHR

LYS 63

LYS98

ARG100

LEU102

LYS103

BDR

HIS78

PRO80

LYS89

LYS91

3.2 De Novo CADD of ASFV DNA

Binding Inhibitors

There is only a handful of literature dedicated to

searching for possible DNA-binding inhibitors in

AFSV. Liu et al. reported that SD1 and SD4 (i.e.,

stilbene derivatives) had inhibitory effects on the

DNA-pA104R binding. Such results were evident by

their ability to repress the ASFV replication (Liu et

al., 2020). To our knowledge, these are the only

molecules known to have therapeutic potential

against African swine fever. It is of great importance

to increase the number of viable ligands. De novo

computer-aided drug design (CADD) approach in

drug discovery enables the generation of novel

ligands based on defined scoring functions (Douguet,

2010). To that end, e-LEA3D, a de novo drug design

tool, was employed in this study. This program uses

a genetic algorithm that evolves molecular fragments

and optimizes the combination of these fragments

(Douguet et al., 2005). Once a library of optimized

molecules is generated, they are assigned a score

based on docking fitness calculated by the PLANTS

program. As shown in Table 2, e-LEA3D generated

nine molecules. ref1 has the highest score (i.e.,

91.25%), implying that this compound has the best

docking conformation from all the generated

molecules using the program's algorithm. However,

In Silico De Novo Synthesis, Screening, and ADME/T Proﬁling of DNA-pA104R Inhibitors as Potential African Swine Fever Therapeutics

this does not bear weight on the binding affinity of the

ligand to pA104R since its primary purpose is to rank

the solutions.

Table 2: Reference molecules generated using e-LEA3D.

Code

Molecular

Formula

Weight

mol

-1

)

Score

(

)

ref1 C

1066.19 91.25

ref2 C

1411.66 87.07

ref3 C

984.11 84.89

ref4 C

1397.64 84.31

ref5 C

1451.65 83.75

ref6 C

1121.29 83.31

ref7 C

1335.57 83.29

ref8 C

1335.57 83.23

ref9 C

1141.25 82.94

3.3 Ligand-Based Virtual Screening

(LBVS) of Molecules Based on

Pharmacophoric Activity

The de novo design strategy of molecules does not

guarantee their ability to be developed into

therapeutic agents. As mentioned, the synthetic

accessibility of the generated compounds is one of the

major challenges in the de novo approach (Mouchlis

et al., 2021). A high binding affinity serves no

purpose if the molecule is hard to synthesize. To

address this problem, a screening process within a

library of commercially available molecules may be

performed. Virtual screening (VS) of prospective

drug compounds has become the norm in the early

stages of drug discovery. It is often regarded as the in

silico counterpart of the tedious and expensive high-

throughput screening (HTS) (Polgar & M. Keseru,

2011).

The screening process is divided into ligand-based

and structure-based approaches. The latter aims to

determine the best ligand that will bind to the receptor

based on surface complementarity (Maia et al., 2020).

The pre-requisite for this type of analysis is the

availability of the 3D protein structure. Meanwhile,

ligand-based VS uses the pharmacophoric properties

of a query molecule to retrieve compounds that

exhibit similar biological and chemical activities

from a database(Singh et al., 2021).

In this study, we applied ligand-based virtual

screening to obtain commercially available molecules

based on the pharmacophores of the reference

compounds. USR-VS is a webserver that uses

Ultrafast Shape Recognition (USR), and Ultrafast

Shape Recognition with CREDO Atom Types

(USRCAT) algorithms for effective pharmacophore

search and retrieval (Li et al., 2016). USR predicts the

molecular shape by analyzing the relative positions of

bonded atoms. As implied by its name, USR enables

the user to search for molecules with a similar three-

dimensional shape at incredible speed. USRCAT is

an extension of USR integrated with the CREDO, a

structural interactomics database (Schreyer &

Blundell, 2013). This algorithm works similarly with

USR, but it uses pharmacophoric constraints for a

more effective similarity search. Therefore, the

USRCAT algorithm was used in this analysis. The

nine (9) de novo designed compounds were used as

query molecules to the USR-VS webserver. The

similarity search covered 23 million molecules and 94

million low-energy conformers from the ZINC

database. Each run returned 100 hit compounds.

Therefore, the nine reference molecules generated

900 hits.

3.4 Drug-likeness Filtration of Hit

Compounds

It is estimated that only 40% of hit compounds can

transition from the pre-clinical to first-in-humans

stage due to their poor physical and chemical

properties (Venkatesh and Lipper, 2000). Drug-

likeness filtration is one of the barriers a compound

must overcome to advance in the late phases of drug

discovery (Hu et al., 2018). This assesses the

probability of a compound to be manufactured as a

therapeutic drug based on some physicochemical

parameters. The method of applying the drug-likeness

filter has been an integral step in the drug discovery

pipeline because any chemical compound may

exhibit an excellent therapeutic effect. Still, not all

could be transformed into viable drug.

To eliminate the hit compounds with undesirable

properties, drug-likeness filtration was performed

using the SwissADME webserver. This web tool has

been used in 2100 in silico analyses (i.e., as per the

number of citations of the published paper of the

developers (Daina et al., 2017) ). SwissADME uses

five filters to assess the drug-like properties.

Violation in any of the filters (i.e., Lipinski (Lipinski,

2004), Ghose (Ghose et al., 1999), Egan (Egan et al.,

2000), Veber (Veber et al., 2002), and Muegge

(Muegge et al., 2001)) disqualifies the compound

from further analysis. By adhering to this selection

criterion, one could ascertain the excellent drug-like

properties of successful compounds.

BIOINFORMATICS 2022 - 13th International Conference on Bioinformatics Models, Methods and Algorithms

Table 3: Drug-likeness filtration of the 900 compounds.

(Note: For simplification, only results from five compounds

were shown).

Compound

no.

Code Formula

No. of filter

violations

(i.e., Lipinski, Ghose,

Veber, Egan, and

Muegge)

1.27 C

2.59 C

9.19 C

S 0

9.43 C

9.92 C

Table 3 shows the summary of the results. After

screening 900 compounds, only 62 were drug-like.

This translates to a 6.89% success rate from hit

identification to drug-like filtration. Lipinski's rule of

five (Ro5) was primarily developed to assess the

druggability of new molecular entities (Lipinski,

2000). If a molecule fails one of the parameters of

Ro5, then the absorption and permeability properties

are put into question. However, Lipinski et al. stated

that molecules that violate at least one of the said

parameters should not be necessarily removed from

the selection process (Petit et al., 2012). Instead, such

molecules should be given low priority in the drug

discovery process. Nonetheless, satisfying the Ro5

without violation is an indicator of excellent drug-

likeness.

3.5 Molecular Docking Studies

Sixty-two (62) identified drug-like molecules

underwent further screening to determine the

compounds that exhibit high binding interactions

with the target protein pA104R. The analysis was

conducted through molecular docking, a structure-

based virtual screening strategy. Molecular docking

is a computational approach to screen for ligands that

fit the protein's ligand-binding site with high

complementarity (i.e., geometrically and

energetically) (Azam & Abbasi, 2013). Docking

tools use search algorithms to predict a ligand's best

docking pose (Sanchez, 2013). Then, a scoring

function calculates the binding free energy of the

protein-ligand complex (Bissantz et al., 2000).

In this study, DockThor, a web server for highly

flexible ligand-docking, was used. This tool utilizes a

dynamic genetic algorithm as a search method. Such

an algorithm allows the intensive survey of the energy

hypersurface to generate multiple minima solutions

(De Magalhães et al., 2014). DockThor uses the

DockTScore as a scoring function based on the

MMFF94S force field (Guedes, Barreto, et al., 2021).

The scoring function considers the intermolecular

interactions, torsional entropy, lipophilic interaction,

polar solvation, and nonpolar solvation. As shown in

Table 4, the binding affinities achieved range from -

7.790 to -6.158 kcal mol

-1

. Compound 8.45 ranked

first with a binding affinity of -7.790 kcal mol

-1

Table 4: Docking results of the 62 drug-like compounds.

(Note: For simplification, only results from five compounds

were shown).

Rank

Compound

Code

Binding

affinity

(kcal mol

-1

)

Total

energy

(kcal mol

-1

)

vdW

Energy

Elec.

energy

1 8.45 -7.790 12.623 -15.214 -11.853

2 2.21 -7.705 36.308 -13.926 -11.001

52 2.59 -6.817 18.362 -4.727 -19.801

51 9.19 -6.841 -1.480 -8.951 -16.563

62 6.45 -6.158 -7.153 -2.613 -19.790

There are two amino acids critical for the binding

of 8.45 with ASFV. GLN76 (i.e., glutamine at

position 76) formed three hydrogen bonds, two of

which are conventional, while the remaining is a pi-

donor hydrogen bond. The first hydrogen bond is

formed by the interaction of the oxygen atom of the

GLN76 to the hydrogen atom of the amino group in

8.45. Then, the hydrogen from the GLN76 interacts

with the carbonyl oxygen of the compound.

Meanwhile, the glutamine's nitrogen atom forms a pi-

donor hydrogen bonding (Figure 3). HIS78 creates a

pi-alkyl interaction with the said molecule. The

amino acid, HIS78, is one of the active site amino acid

residues (Table 1). Such interactions might explain

why compound 8.45 had the highest binding affinity

among all drug-like molecules that underwent

molecular docking. Therefore, 8.45 could be a

Figure 3: Molecular interactions of compound 8.45 to (top)

GLN76 and (bottom) HIS78 of the ASFV.

In Silico De Novo Synthesis, Screening, and ADME/T Proﬁling of DNA-pA104R Inhibitors as Potential African Swine Fever Therapeutics

potential DNA-binding inhibitor of pA104R solely

based on the docking results. However, further tests

must be conducted to determine this compound's

potential as a therapeutic agent.

3.6 ADME/T Profiling

Determination of pharmacokinetic characteristics is

one of the most critical steps to ensure that the drug

being developed is safe to be administered during

animal and clinical trials. The results for the ADME/T

of the unique compounds with the highest binding

affinities are shown in Table 5. For the absorption

parameters, it can be observed that the intestinal

absorption% for the drug-like compounds have a very

high positive value, ranging from 69.985% to

95.646%. Such results match the existing literature

since adhering to Lipinski's Rule of Five entails that

the drug-like compounds are likely to be well-

absorbed in the intestine (Zhao et al., 2001). Further

supporting this idea, the range derived for the human

intestinal absorption (HIA) % was above the optimum

level of 30%, as shown by Wang (2016) (N. N. Wang

et al., 2017). All the values for skin permeability of

both the unique compounds indicate skin

permeability because all the values were lower than -

2.5 (Hassan et al., 2018). The results are favorable

since they signify that the drugs can be applied

through skin contact and promote the elimination of

these drugs to prevent the accumulation of chemicals

in the body (Osborne & Musakhanian, 2018). This

finding provides an alternative route of

administration for the proposed compounds.

Caco-2 permeability is considered the final

absorption parameter. It makes use of the Caco-2 cell,

or the human colon adenocarcinoma, to model the

intestinal absorption of many compounds

(Matsumoto et al., 2018). The Caco-2 permeability

values for the unique group were varying. The unique

and reference groups yielded acceptable results as all

values were above the reported threshold for optimum

Caco-2 permeability value (i.e., -5.15) (N. N. Wang

et al., 2016). This finding reinforces the results given

by the HIA% that the compounds under study have

adequate intestinal absorption.

We observed that the VD

values of the unique

group varied but were negative for all reference

compounds for the distribution parameters. A higher

value entails better distribution into the tissues

than in the plasma (Yates & Arundel, 2008).

Compounds 2.2 and 9.19 had unfavorable VD

values because they were close to the minimum range

(i.e., -0.15) (Firdausy et al., 2020). On the other hand,

the drug compounds 8.45, 2.21, 8.40, 7.21, and 2.59

displayed moderate VD

values because their values

were between the range reported (i.e., -0.15 to 0.45)

(Firdausy et al., 2020).

The blood-brain barrier permeability was varying

for the unique group but all negative for the reference

compounds. Nevertheless, all the unique compounds

were unable to penetrate the blood-brain barrier (i.e.,

< 0.3) (Firdausy et al., 2020). Such a result is a

positive indication since the expected target of the

compounds is not found within the brain. Regarding

CNS permeability, compound 7.21 can effectively

penetrate the central nervous system (i.e., > -2),

whereas the remaining unique compounds could only

poorly penetrate the CNS (i.e., < -3) (Pires et al.,

2015). However, even if the CNS permeability

values were favorable for all the compounds, they

would still not penetrate the CNS due to the blood-

brain barrier (Carpenter et al., 2014).

The metabolism of the compounds being studied

was dictated by their capacity to become either

CYP2D6 or CYP3A4 (i.e., the two main subtypes of

cytochrome P450) inhibitors (Firdausy et al., 2020).

All of the unique compounds were not CYP2D6

inhibitors. Meanwhile, compounds 2.2, 8.40, and

7.21 were known to be CYP3A4 inhibitors. A

negative result from these tests could suggest the

excellent metabolism of the proposed drug-like

compounds in the human body; the presence of

inhibitors poses a threat for the body since decreased

metabolism leads to the accumulation of the

compounds and will thus increase the toxicity of that

potential drug (Niel et al., 1992).

For the excretion parameter, total clearance was

considered. This parameter measures the compound's

ability to be cleared from all tissues (i.e., the

combination of renal and hepatic clearances. The

CLtot values for compounds considered were within

the range -0.278 to 1.449 log ml min

-1

. It was

found that the highest total clearance was achieved by

compound 8.45, which suggests that it has the highest

bioavailability (Firdausy et al., 2020). Meanwhile,

compounds 7.21 and 9.19 had negative values

indicating their poor systemic clearance.

Finally, the toxicity of the proposed drugs was

evaluated. The Ames test is a preliminary evaluation

to determine the mutagenicity of drug candidates

(Mortelmans et al., 2016). Based on the results, only

compound 2.2 was characterized as a mutagen. There

is a high correlation between carcinogenicity and

mutagenicity (ca. 90%). This indicates that

compound 2.2 could induce mutations leading to

cancer (Mortelmans et al., 2016). It is therefore

essential to perform other tests to determine its

genotoxicity.

BIOINFORMATICS 2022 - 13th International Conference on Bioinformatics Models, Methods and Algorithms

Table 5: ADME/T Results of the unique (i.e., drug-like) group. The top 5 compounds based on their binding affinities from

the molecular docking studies were used. Compounds 9.19 and 2.59 were also included for the ADME/T profiling due to

their low hepatoxicity.

Code

Absorption Distribution Metabolism Excretion Toxicity

Intestina

l abs.

Skin

permeability

Caco-2

permeability

BBB

permeabili

CNS

permeabi

lity

CYP 2D6

inhibitor

CYP 3A4

inhibitor

Total

Clearance

Ames

Toxicity

Hepato

toxicity

LD50

(mg kg

-1

)

8.45 89.927 -2.849 0.794 0.361 -0.569 -2.641 No No 1.022 No Yes 695.29

2.21 91.148 -3.022 0.811 0.395 -0.802 -2.653 No No 0.808 No Yes 911.54

2.2 82.316 -2.741 0.158 -0.182 -1.053 -2.42 No Yes 0.109 Yes Yes 1056.52

8.40 92.54 -3.069 0.662 0.221 -0.955 -2.974 No Yes 0.248 No Yes 1081.99

7.21 92.942 -3.246 0.703 0.126 -0.183 -1.953 No Yes -0.317 No Yes 793.17

9.19 91.444 -3.869 0.878 -0.087 -0.375 -2.918 No No -0.278 No No 390.81

2.59 88.437 -2.863 0.987 0.105 -0.836 -2.44 No No 0.438 No No 2583.83

Meanwhile, a hepatotoxicity test was also

performed to determine whether the drug could cause

significant liver injury. This stage of the development

process greatly impedes the translation of a substance

into a commercial drug (Björnsson, 2016). Based on

the results, only compounds 9.19 and 2.59 were not

hepatoxic. This result signifies that these are the only

compounds from the unique group that causes

minimal harm to the human liver.

The final parameter considered is the rat oral acute

toxicity (LD

) of the proposed drug candidates. This

parameter determines the amount of the substance

that could kill 50% of the test animal population

(Adamson, 2016). The higher the LD

value of the

compound, the less toxic the substance is when taken

orally by the individual. The resulting LD50 of the

unique group compounds ranged from 2.121 to 2.985.

Based on the Hodge and Sterner scale, all compounds

except 9.19 are considered only slightly toxic, with a

toxicity rating of 4 (ca. 500-5000 mg kg

-1

) (Ahmed,

2015). Meanwhile, compound 9.19 is considered

moderately toxic since its LD

value falls under a

toxicity rating of 3 (ca. 50-500 mg kg

-1

). The

remaining compounds had relatively low LD

that is

indicative of their high toxicity. Therefore, caution

must be exercised when deriving the optimum dosage

of these drug candidates.

To identify the most suitable compounds among

the unique group, we devised a scoring system that

consisted of the molecular docking rank and the

ADME/T score. The top five compounds from the

molecular docking studies were analyzed for

ADME/T profiling. However, we also considered

compounds 9.19 and 2.59 even though they ranked

and 52

, respectively. These two were the only

non-hepatoxic compounds; therefore, we opted to

include them in the scoring system. For the ADME/T

scoring, each parameter violation was awarded one

point. The compound with the lowest score was

deemed to have the most favorable ADME/T

properties. Based on the results, compound 2.59

performed best on the ADME/T studies (Table 6).

Meanwhile, compound 2.2 had the most number of

parameter violations.

Table 6: Summary of docking and ADME/T performance.

For ADME/T profiling, violation in any of the parameters

is rewarded one point. The final score is the average of the

docking and ADME/T scores.

Code

Docking

Rank/Score

ADME/T Score Final Score

8.45 1 2 1.5

2.21 2 2 2

2.2 3 4 3.5

8.40 4 3 3.5

7.21 5 3 4

9.19 51 2 26.5

2.59 52 1 26.5

The docking score and the ADME/T score were

averaged to calculate the final score. Compound 8.45

had the lowest score (ca. 1.5), signifying its excellent

binding affinity against pA104R and favorable

pharmacokinetic properties. Although compound

2.59 ranked first in the ADME/T studies, it fell short

of its molecular docking ranking resulting in a low

final score (ca. 26.5). Compounds 2.21, 2.2, 8.40, and

7.21 had relatively good final scores. Such results

implied their superior properties similar to compound

8.45. Additional tests must be conducted on these

compounds to determine their capabilities as

therapeutic agents against ASF. Particularly,

bioassays such as the haemadsorption test (HAT) are

useful in exploring the efficiency of the compounds

as ASFV therapeutics (Fischer et al., 2020).

In Silico De Novo Synthesis, Screening, and ADME/T Proﬁling of DNA-pA104R Inhibitors as Potential African Swine Fever Therapeutics

4 CONCLUSION

In this paper, the identification and characterization

of potential drug candidates for the treatment of ASF

were conducted. We were able to characterize the

structure of the pA104R protein with visualization

software. The DNA binding site of the pA104R was

determined. Nine (9) de novo reference compounds

were generated. Of these compounds, 900

commercially available drug-like small molecules

were retrieved through ligand-based virtual screening

using pharmacophoric similarities.

Drug-likeness filtration was done to determine the

compounds with excellent druggability properties.

Sixty-two (62) drug-like compounds were subjected

to molecular docking and ADME/T studies. Of these

filtered drug-like molecules, compound 8.45

achieved exceptional docking rank (ca. 1) and

ADME/T score (ca. 2), earning the lowest final score.

The other drug-like molecules (i.e., 2.21, 2.2, 8.40,

and 7.21) also performed well. Compounds 9.19 and

2.59 had the best ADME/T profile but performed

poorly in the molecular docking studies. Further

experiments must be performed to identify their

potential as anti-ASF therapeutics.

REFERENCES

Adamson, R. H. (2016). The acute lethal dose 50 (LD50) of

caffeine in albino rats. Regulatory Toxicology and

Pharmacology, 80, 274–276. https://doi.org/10.1016/

j.yrtph.2016.07.011

Agarwal, D., Udoji, M., & Trescot, A. (2017). Genetic

Testing for Opioid Pain Management: A Primer. Pain

and Therapy, 6, 1–13. https://doi.org/10.1007/s40122-

017-0069-2

Ahmed, M. (2015). Acute Toxicity (Lethal Dose 50

Calculation) of Herbal Drug Somina in Rats and Mice.

Pharmacology & Pharmacy, 06(03), 185–189.

https://doi.org/10.4236/pp.2015.63019

Alonso, C., Borca, M., Dixon, L., Revilla, Y., Rodriguez,

F., & Escribano, J. M. (2018). ICTV virus taxonomy

profile: Asfarviridae. Journal of General Virology,

99(5), 613–614. https://doi.org/10.1099/jgv.0.001049

Andrés, G., Alejo, A., Salas, J., & Salas, M. L. (2002).

African Swine Fever Virus Polyproteins pp220 and

pp62 Assemble into the Core Shell. Journal of

Virology, 76(24), 12473–12482. https://doi.org/10.112

8/jvi.76.24.12473-12482.2002

Azam, S. S., & Abbasi, S. W. (2013). Molecular docking

studies for the identification of novel melatoninergic

inhibitors for acetylserotonin-O-methyltransferase

using different docking routines. Theoretical Biology

and Medical Modelling, 10(1), 1–16. https://doi.org/

10.1186/1742-4682-10-63

Bissantz, C., Folkers, G., & Rognan, D. (2000). Protein-

based virtual screening of chemical databases. 1.

Evaluation of different docking/scoring combinations.

Journal of Medicinal Chemistry, 43(25), 4759–4767.

https://doi.org/10.1021/jm001044l

Björnsson, E. S. (2016). Hepatotoxicity by drugs: The most

common implicated agents. International Journal of

Molecular Sciences, 17(2). https://doi.org/10.3390/ijms

17020224

Blome, S., Franzke, K., & Beer, M. (2020). African swine

fever – A review of current knowledge. Virus Research,

287(April), 198099. https://doi.org/10.1016/j.virusres.

2020.198099

Carpenter, T. S., Kirshner, D. A., Lau, E. Y., Wong, S. E.,

Nilmeier, J. P., & Lightstone, F. C. (2014). A Method

to Predict Blood-Brain Barrier Permeability of Drug-

Like Compounds Using Molecular Dynamics

Simulations. Biophysical Journal, 107(3), 630–641.

https://doi.org/10.1016/j.bpj.2014.06.024

Chen, S., Zhang, X., Nie, Y., Li, H., Chen, W., Lin, W.,

Chen, F., & Xie, Q. (2021). African Swine Fever Virus

Protein E199L Promotes Cell Autophagy through the

Interaction of PYCR2. Virologica Sinica, 36(2), 196–

206. https://doi.org/10.1007/s12250-021-00375-x

Daina, A., Michielin, O., & Zoete, V. (2017). SwissADME:

A free web tool to evaluate pharmacokinetics, drug-

likeness and medicinal chemistry friendliness of small

molecules. Scientific Reports, 7(January), 1–13.

https://doi.org/10.1038/srep42717

De Magalhães, C. S., Almeida, D. M., Barbosa, H. J. C., &

Dardenne, L. E. (2014). A dynamic niching genetic

algorithm strategy for docking highly flexible ligands.

Information Sciences, 289(1), 206–224.

https://doi.org/10.1016/j.ins.2014.08.002

Douguet, D. (2010). e-LEA3D: A computational-aided

drug design web server. Nucleic Acids Research,

38(SUPPL. 2), 615–621.

https://doi.org/10.1093/nar/gkq322

Douguet, D., Munier-Lehmann, H., Labesse, G., & Pochet,

S. (2005). LEA3D: A computer-aided ligand design for

structure-based drug design. Journal of Medicinal

Chemistry, 48(7), 2457–2468. https://doi.org/10.1021

/jm0492296

Egan, W. J., Merz, K. M., & Baldwin, J. J. (2000).

Prediction of drug absorption using multivariate

statistics. Journal of Medicinal Chemistry, 43(21),

3867–3877. https://doi.org/10.1021/jm000292e

Firdausy, A. F., Muti’ah, R., & Rahmawati, E. K. (2020).

Predicting Pharmacokinetic Profiles of Sunflower’s

(Helianthus annuus L.) Active Compounds using in

Silico Approach. Journal of Islamic Medicine, 4(1), 1–

7. https://doi.org/10.18860/jim.v4i1.8840

Fischer, M., Hühr, J., Blome, S., Conraths, F. J., & Probst,

C. (2020). Stability of African swine fever virus in

carcasses of domestic pigs and wild boar

experimentally infected with the ASFV “Estonia 2014”

isolate. Viruses, 12(10). https://doi.org/10.3390/v12

101118

Frouco, G., Freitas, F. B., Coelho, J., Leitão, A., Martins,

C., & Ferreira, F. (2017a). crossm. 91(12), 1–14.

BIOINFORMATICS 2022 - 13th International Conference on Bioinformatics Models, Methods and Algorithms

Frouco, G., Freitas, F. B., Coelho, J., Leitão, A., Martins,

C., & Ferreira, F. (2017b). DNA-Binding Properties of

African Swine Fever Virus pA104R, a Histone-Like

Protein Involved in Viral Replication and

Transcription. Journal of Virology, 91(12), 1–14.

https://doi.org/10.1128/jvi.02498-16

Galindo, I., & Alonso, C. (2017). African swine fever virus:

A review. Viruses, 9(5). https://doi.org/10.3390/v905

0103

Ghose, A. K., Viswanadhan, V. N., & Wendoloski, J. J.

(1999). A knowledge-based approach in designing

combinatorial or medicinal chemistry libraries for drug

discovery. 1. A qualitative and quantitative

characterization of known drug databases. Journal of

Combinatorial Chemistry, 1(1), 55–68.

https://doi.org/10.1021/cc9800071

Guedes, I. A., Barreto, A. M. S., Marinho, D., Krempser,

E., Kuenemann, M. A., Sperandio, O., Dardenne, L. E.,

& Miteva, M. A. (2021). New machine learning and

physics-based scoring functions for drug discovery.

Scientific Reports, 11(1), 1–19.

https://doi.org/10.1038/s41598-021-82410-1

Guedes, I. A., Costa, L. S. C., dos Santos, K. B., Karl, A. L.

M., Rocha, G. K., Teixeira, I. M., Galheigo, M. M.,

Medeiros, V., Krempser, E., Custódio, F. L., Barbosa,

H. J. C., Nicolás, M. F., & Dardenne, L. E. (2021). Drug

design and repurposing with DockThor-VS web server

focusing on SARS-CoV-2 therapeutic targets and their

non-synonym variants. Scientific Reports, 11(1), 1–20.

https://doi.org/10.1038/s41598-021-84700-0

Hamza, A., Wei, N. N., & Zhan, C. G. (2012). Ligand-based

virtual screening approach using a new scoring

function. Journal of Chemical Information and

Modeling, 52(4), 963–974. https://doi.org/10.1021/ci

200617d

Hassan, M., Shahzadi, S., Seo, S. Y., Alashwal, H., Zaki,

N., & Moustafa, A. A. (2018). Molecular docking and

dynamic simulation of AZD3293 and solanezumab

effects against BACE1 to treat alzheimer’s disease.

Frontiers in Computational Neuroscience, 12(June), 1–

11. https://doi.org/10.3389/fncom.2018.00034

Hu, Q., Feng, M., Lai, L., & Pei, J. (2018). Prediction of

Drug-Likeness Using Deep Autoencoder Neural

Networks. Frontiers in Genetics, 9(November), 1–8.

https://doi.org/10.3389/fgene.2018.00585

Kiriiri, G. K., Njogu, P. M., & Mwangi, A. N. (2020).

Exploring different approaches to improve the success

of drug discovery and development projects: a review.

Future Journal of Pharmaceutical Sciences, 6(1), 27.

https://doi.org/10.1186/s43094-020-00047-9

Li, H., Leung, K. S., Wong, M. H., & Ballester, P. J. (2016).

USR-VS: a web server for large-scale prospective

virtual screening using ultrafast shape recognition

techniques. Nucleic Acids Research, 44(W1), W436–

W441. https://doi.org/10.1093/nar/gkw320

Lipinski, C. A. (2000). Drug-like properties and the causes

of poor solubility and poor permeability. Journal of

Pharmacological and Toxicological Methods,

44(1),

235–249. https://doi.org/10.1016/S1056-8719(00)0010

7-6

Lipinski, C. A. (2004). Lead- and drug-like compounds:

The rule-of-five revolution. Drug Discovery Today:

Technologies, 1(4), 337–341.

https://doi.org/10.1016/j.ddtec.2004.11.007

Liu, R., Sun, Y., Chai, Y., Li, S., Li, S., Wang, L., Su, J.,

Yu, S., Yan, J., Gao, F., Zhang, G., Qiu, H.-J., Gao, G.

F., Qi, J., & Wang, H. (2020). The structural basis of

African swine fever virus pA104R binding to DNA and

its inhibition by stilbene derivatives. Proceedings of the

National Academy of Sciences of the United States of

America, 117(20), 11000–11009.

https://doi.org/10.1073/pnas.1922523117

Maia, E. H. B., Assis, L. C., de Oliveira, T. A., da Silva, A.

M., & Taranto, A. G. (2020). Structure-Based Virtual

Screening: From Classical to Artificial Intelligence.

Frontiers in Chemistry, 8(April).

https://doi.org/10.3389/fchem.2020.00343

Matsumoto, T., Kaifuchi, N., Mizuhara, Y., Warabi, E., &

Watanabe, J. (2018). Use of a Caco-2 permeability

assay to evaluate the effects of several Kampo

medicines on the drug transporter P-glycoprotein.

Journal of Natural Medicines, 72(4), 897–904.

https://doi.org/10.1007/s11418-018-1222-x

Mohs, R. C., & Greig, N. H. (2017). Drug discovery and

development: Role of basic biological research.

Alzheimer’s & Dementia (New York, N. Y.), 3(4), 651–

657. https://doi.org/10.1016/j.trci.2017.10.005

Mortelmans, K., Mortelmans, K., & Zeiger, E. (2016). The

Ames Salmonella / microsome mutagenicity assay The

Ames Salmonella / microsome mutagenicity assay.

5107(December 2000), 29–60.

Mouchlis, V. D., Afantitis, A., Serra, A., Fratello, M.,

Papadiamantis, A. G., Aidinis, V., Lynch, I., Greco, D.,

& Melagraki, G. (2021). Advances in de novo drug

design: From conventional to machine learning

methods. International Journal of Molecular Sciences,

22(4), 1–22. https://doi.org/10.3390/ijms22041676

Muegge, I., Heald, S. L., & Brittelli, D. (2001). Simple

selection criteria for drug-like chemical matter. Journal

of Medicinal Chemistry, 44(12), 1841–1846.

https://doi.org/10.1021/jm015507e

Niel, N., Rechencq, E., Vidal, J. P., Escale, R., Durand, T.,

Girard, J. P., Rossi, J. C., Muller, A., & Bonne, C.

(1992). Synthesis and contractile activity of new

pseudopeptido and thioaromatic analogues of

leukotriene D4. Prostaglandins, 43(1), 45–54.

https://doi.org/10.1016/0090-6980(92)90063-Y

Osborne, D. W., & Musakhanian, J. (2018). Skin

Penetration and Permeation Properties of

Transcutol®—Neat or Diluted Mixtures. AAPS

PharmSciTech, 19(8), 3512–3533.

https://doi.org/10.1208/s12249-018-1196-8

Pen, G., Yang, N., Teng, D., Mao, R., Hao, Y., & Wang, J.

(2020). A review on the use of antimicrobial peptides

to combat porcine viruses.

Antibiotics, 9(11), 1–18.

https://doi.org/10.3390/antibiotics9110801

Petit, J., Meurice, N., Kaiser, C., & Maggiora, G. (2012).

Softening the Rule of Five - Where to draw the line?

Bioorganic and Medicinal Chemistry, 20(18), 5343–

5351. https://doi.org/10.1016/j.bmc.2011.11.064

In Silico De Novo Synthesis, Screening, and ADME/T Proﬁling of DNA-pA104R Inhibitors as Potential African Swine Fever Therapeutics

Pires, D. E. V., Blundell, T. L., & Ascher, D. B. (2015).

pkCSM: Predicting small-molecule pharmacokinetic

and toxicity properties using graph-based signatures.

Journal of Medicinal Chemistry, 58(9), 4066–4072.

https://doi.org/10.1021/acs.jmedchem.5b00104

Polanski, J. (2020). 4.26 - Chemoinformatics☆. In S.

Brown, R. Tauler, & B. Walczak (Eds.),

Comprehensive Chemometrics (Second Edition) (pp.

635–676). Elsevier. https://doi.org/https://doi.org/10.

1016/B978-0-12-409547-2.14327-6

Polgar, T., & M. Keseru, G. (2011). Integration of Virtual

and High Throughput Screening in Lead Discovery

Settings. Combinatorial Chemistry & High Throughput

Screening, 14(10), 889–897.

https://doi.org/10.2174/138620711797537148

Sanchez, G. (2013). Las instituciones de ciencia y

tecnología en los procesos de aprendizaje de la

producción agroalimentaria en Argentina. El Sistema

Argentino de Innovación: Instituciones, Empresas y

Redes. El Desafío de La Creación y Apropiación de

Conocimiento., 26(February), 15–26. https://doi.org/

10.1002/prot

Santos, K. B., Guedes, I. A., Karl, A. L. M., & Dardenne,

L. E. (2020). Highly Flexible Ligand Docking:

Benchmarking of the DockThor Program on the

LEADS-PEP Protein-Peptide Data Set. Journal of

Chemical Information and Modeling, 60(2), 667–683.

https://doi.org/10.1021/acs.jcim.9b00905

Schreyer, A. M., & Blundell, T. (2012). USRCAT: Real-

time ultrafast shape recognition with pharmacophoric

constraints. Journal of Cheminformatics, 4(11), 1.

https://doi.org/10.1186/1758-2946-4-27

Schreyer, A. M., & Blundell, T. L. (2013). CREDO: A

structural interactomics database for drug discovery.

Database, 2013, 1–9.

https://doi.org/10.1093/database/bat049

Shen, M., Tian, S., Li, Y., Li, Q., Xu, X., Wang, J., & Hou,

T. (2012). Drug-likeness analysis of traditional Chinese

medicines: 1. property distributions of drug-like

compounds, non-drug-like compounds and natural

compounds from traditional Chinese medicines.

Journal of Cheminformatics, 4(1), 1–13.

https://doi.org/10.1186/1758-2946-4-31

Singh, N., Chaput, L., & Villoutreix, B. O. (2021). Virtual

screening web servers: designing chemical probes and

drug candidates in the cyberspace. Briefings in

Bioinformatics, 22(2), 1790–1818. https://doi.org/10.

1093/bib/bbaa034

Swinger, K. K., & Rice, P. A. (2004). IHF and HU: Flexible

architects of bent DNA. Current Opinion in Structural

Biology, 14(1), 28–35. https://doi.org/10.1016/j.sbi.20

03.12.003

Umashankar, V., & Gurunathan, S. (2015). Drug discovery:

An appraisal. International Journal of Pharmacy and

Pharmaceutical Sciences, 7(4), 59–66.

Urbano, A. C., & Ferreira, F. (2020). Role of the dna-

binding protein pa104r in asfv genome packaging and

as a novel target for vaccine and drug development.

Vaccines, 8(4), 1–17. https://doi.org/10.3390/vaccines

8040585

Veber, D. F., Johnson, S. R., Cheng, H. Y., Smith, B. R.,

Ward, K. W., & Kopple, K. D. (2002). Molecular

properties that influence the oral bioavailability of drug

candidates. Journal of Medicinal Chemistry, 45(12),

2615–2623. https://doi.org/10.1021/jm020017n

Venkatesh, S., & Lipper, R. A. (2000). Role of the

development scientist in compound lead selection and

optimization. Journal of Pharmaceutical Sciences,

89(2), 145–154. https://doi.org/10.1002/(SICI)1520-

6017(200002)89:2<145::AID-JPS2>3.0.CO;2-6

Wang, J. Y. J. (2001). DNA damage and apoptosis. Cell

Death and Differentiation, 8(11), 1047–1048.

https://doi.org/10.1038/sj.cdd.4400938

Wang, N. N., Dong, J., Deng, Y. H., Zhu, M. F., Wen, M.,

Yao, Z. J., Lu, A. P., Wang, J. B., & Cao, D. S. (2016).

ADME Properties Evaluation in Drug Discovery:

Prediction of Caco-2 Cell Permeability Using a

Combination of NSGA-II and Boosting. Journal of

Chemical Information and Modeling, 56(4), 763–773.

https://doi.org/10.1021/acs.jcim.5b00642

Wang, N. N., Huang, C., Dong, J., Yao, Z. J., Zhu, M. F.,

Deng, Z. K., Lv, B., Lu, A. P., Chen, A. F., & Cao, D.

S. (2017). Predicting human intestinal absorption with

modified random forest approach: a comprehensive

evaluation of molecular representation, unbalanced

data, and applicability domain issues. RSC Advances,

7(31), 19007–19018. https://doi.org/10.1039/C6RA2

8442F

Yates, J. W. T., & Arundel, P. A. (2008). On the volume of

distribution at steady state and its relationship with two-

compartmental models. Journal of Pharmaceutical

Sciences, 97(1), 111–122. https://doi.org/10.1002/jp

s.21089

Zhao, Y. H., Le, J., Abraham, M. H., Hersey, A.,

Eddershaw, P. J., Luscombe, C. N., Boutina, D., Beck,

G., Sherborne, B., Cooper, I., & Platts, J. A. (2001).

Evaluation of human intestinal absorption data and

subsequent derivation of a quantitative structure -

Activity relationship (QSAR) with the Abraham

descriptors. Journal of Pharmaceutical Sciences, 90(6),

749–784. https://doi.org/10.1002/jps.1031

Zhu, Z., Fan, Y., Cai, Z., Zhang, Z., Lu, C., Jiang, T.,

Zhang, G., & Peng, Y. (2019). Prediction of antiviral

drugs against African Swine Fever Viruses based on

protein-protein interaction analysis. BioRxiv.

https://doi.org/10.1101/599043

Zhu, Z., Fan, Y., Liu, Y., Jiang, T., Cao, Y., & Peng, Y.

(2020). Prediction of antiviral drugs against African

swine fever viruses based on protein-protein interaction

analysis. PeerJ, 2020(4), 1–19. https://doi.org/10.

7717/peerj.8855

BIOINFORMATICS 2022 - 13th International Conference on Bioinformatics Models, Methods and Algorithms