The Combination of Computational and Biosensing ...downloads.hindawi.com › journals › bmri ›...

Research ArticleThe Combination of Computational andBiosensing Technologies for Selecting Aptamer againstProstate Specific Antigen

Pi-Chou Hsieh1 Hui-Ting Lin2 Wen-Yih Chen3 Jeffrey J P Tsai1 andWen-Pin Hu14

1Department of Bioinformatics and Medical Engineering Asia University Taichung City 41354 Taiwan2Department of Physical Therapy I-Shou University Kaohsiung City 82445 Taiwan3Department of Chemical and Materials Engineering National Central University Jhongli 32001 Taiwan4Department of Medical Laboratory Science and Biotechnology China Medical University Taichung City 40402 Taiwan

Correspondence should be addressed to Wen-Pin Hu wenpinhuasiaedutw

Received 24 November 2016 Revised 10 March 2017 Accepted 19 March 2017 Published 28 March 2017

Academic Editor Hicham Fenniri

Copyright copy 2017 Pi-Chou Hsieh et alThis is an open access article distributed under the Creative Commons Attribution Licensewhich permits unrestricted use distribution and reproduction in any medium provided the original work is properly cited

Herein we report a method of combining bioinformatics and biosensing technologies to select aptamers against prostate specificantigen (PSA)Themain objective of this study is to select DNAaptamers with higher binding affinity for PSA by using the proposedmethod Based on the five known sequences of PSA-binding aptamers we adopted the functions of reproduction and crossover inthe genetic algorithm to produce next-generation sequences for the computational and experimental analysis RNAfold web serverwas utilized to analyze the secondary structures and the 3-dimensional molecular models of aptamer sequences were generatedby using RNAComposer web server ZRANK scoring function was used to rerank the docking predictions from ZDOCK Thebiosensors the quartz crystal microbalance (QCM) and a surface plasmon resonance (SPR) instrument were used to verify thebinding ability of selected aptamer for PSA By carrying out the simulations and experiments after two generations we obtain oneaptamer that can have the highest binding affinity with PSA which generates almost 2-fold and 3-fold greater measured signalsthan the responses produced by the best known DNA sequence in the QCM and SPR experiments respectively

1 Introduction

Prostate specific antigen (PSA) is a glycoprotein and has amolecular weight of 33-34 kDa It is primarily created byprostate epithelial cells and is a serum marker commonlyused to diagnose prostate cancer Patients with prostate can-cer generally have a large amount and a high concentrationof PSA released into their circulatory system as a resulttheir serum PSA concentration is 105 times higher thanthat of regular healthy people Thus measuring serum PSAconcentration is commonly used in the early diagnosis andfor subsequent monitoring of patients with prostate cancerPatients who undergo an examination and show a serumPSA concentration ge 40 ngmL [1 2] are suspected of havingprostate cancer after which doctors will ask the patients tohave regular follow-up examinations If the PSA concen-trations of patients are over 10 ngmL the doctors will ask

the patients to take a biopsy to confirm whether they trulyhave prostate cancer PSA testing is crucial for determiningwhether an individual has prostate cancer and most assaysfor PSA detection are based on the use of antibodies as therecognition elements [3]

In 1990 several research teams independently presentedan in vitro selection technique Such technique can findsome certain nucleic acid sequences that can bind to tar-get molecules of nonnucleic acids in which the productsdemonstrate high affinity and specificity [3ndash5]The techniqueis called the systematic evolution of ligands by exponentialenrichment (SELEX) Concerning the SELEX technique itis an iterative process and consists of three major steps(1) binding (2) partitioning (3) amplification [6 7] Theprocess starts from a pool of single-stranded oligonu-cleotide sequences incubated with the target molecule someoligonucleotides exhibiting binding ability towards the target

HindawiBioMed Research InternationalVolume 2017 Article ID 5041683 11 pageshttpsdoiorg10115520175041683

2 BioMed Research International

molecule can be obtained In the partitioning step the target-bound nucleic acids are separated from unbound sequencesBound nucleic acids are then amplified to generate a newpool of nucleic acid sequences for use in the next round ofselection After repeating 8sim20 cycles nucleic acid aptamerswith high affinity and specificity for the target moleculecan be obtained In essence aptamers are short nucleic acidsequences with a length of 12 to 80 nucleic acids Aptamersthat have been identified by current research include RNAand DNA molecules and RNA-type aptamers account forthe majority of aptamers Aptamers can be used to reactwith various target molecules such as metal ions shortpeptides pathogenic microorganisms micromolecules suchas amino acids andmacromolecules like proteins Comparedwith antibodies aptamers are manufactured by chemicalsynthesis therefore batch-to-batch variation can be greatlyminimized The economical high-accuracy large-scale pro-duction of aptamers makes aptamers very suitable for clinicalapplications [8] In addition nucleic acid aptamers featureadvantages not found in common antibodies which are listedas follows small molecules high stability easy productionpossibility for reuse and unchanged affinity after simplechemical modifications [8] Because of these advantagesaptamers can be used in various biotechnology-related fieldssuch as environment and food quality tests therapies drugdevelopments purification processes and diagnoses

Previous studies have shown that the use of appropriatemolecular simulation software is highly beneficial for design-ing molecules with high affinity and selectivity [9] Advancesin computer software and hardware over the past severaldecades as well as developments in mathematical modelshave resulted in increasing researchers using computersto simulate and predict the reactions of biomolecules andchemical molecules Anderson and Mecozzi [10] employedthe molecular simulation method in addition to conductingtraditional experiments to investigate the interaction of the35-mer aptamer and cofactor flavin mononucleotide (FMN)Their study successfully reduced the length of the 35-meraptamer to 14mers while retaining the binding ability forflavin mononucleotide

In 2011 Bini et al [11] proposed a computationally assistedmethod to study and evaluate aptamer-protein interaction forselecting new aptamer sequencesThey performedmolecularsimulations by using a molecular model and conducted anexperiment to verify the prediction results of the molecularsimulations Bini et al initially used a well-known 15-merthrombin binding aptamer (TBA) and then they modifiedbase sequences at certain locations to generate a varietyof mutant aptamer sequences After docking simulationsexperiments were performed by using a surface plasmonresonance (SPR) biosensor The results indicated a mutantsequence could have a slightly better performance than thewell-known TBA and they found that experimental resultswere in agreement with the simulation findings Lupoldet al [12] reported the specific binding between an RNAaptamer and a prostate specific membrane antigen (PSMA)the RNA aptamer demonstrated the potential to be usedas a drug carrier to treat prostate cancer In 2010 Jeonget al [13] researched and found an RNA aptamer that can

be used to identify PSA the RNA aptamer featured a 51015840-end to 31015840-end sequence of CCGUAGGUCACGGCAGCG-AAGCUCUAGGCGCGGCCAGUUGC Savory et al [14]selected five DNA sequences by using the SELEX techniqueand performed subsequent in silico analyses With the use ofselected five DNA sequences as parent sequences the geneticalgorithm (GA) was adopted to generate next-generationsequences and those generated sequences were synthesizedand assayed in vitro One of the fourth-generation DNAaptamers exhibited a 48-fold higher PSA-binding ability thanthe parent sequences

Some scholars have identified RNA aptamers as a possiblealternative to antibodies for testing serum PSA concentration[12 13] Nevertheless the long length of RNA aptamerscould cause the difficulty in the commercial synthesis andlimit the applications [3] By contrast DNA-based PSAaptamers had been used to detect PSA by using differentsensing technologies [15 16] Herein we reported a studythat employed computational approaches including GA theanalysis of DNA secondary structure and the molecularsimulation to evaluate the aptamer-protein interactions andbiosensing technologies to verify the PSA-binding ability ofselected aptamers The original five sequences of aptamersreported in the study of Savory et al [14] were chosen asthe parent sequences and the genetic algorithm was usedto generate new nucleic acid aptamers for the selection Byusing the proposed strategy of selection a newDNA aptamerexhibited stronger binding capability with PSA which wasselected in the end of selection procedure We think theproposed strategy of combining computational approaches(GA and molecular simulations) in the selection of aptamercan streamline the number of experiments carried out forthe selection and the results show that it is possible tocomplement SELEX for the selection of aptamer

2 Materials and Methods

21The Overview of Research Process Figure 1 shows all stepsof aptamer selection in this study First five sequences thatcan bind to PSA (obtained by Savory et al by using the SELEXprocedure [14]) were set as the parent sequences Savory et alnamed the five sequences as PSap4-3 PSap4-4 PSap4-11PSap4-9 and PSap4-6 To enable the genetic algorithm tocalculate the four bases used in this study (ie A T G andC) such bases were coded bases A T G and C for the fivesequences were numbered 00 01 10 and 11 respectively Thegenetic algorithm toolbox for MATLAB [17] was utilized toperform reproduction and crossover operations for generat-ing 20 next-generation sequences The sequence information(in numerical format) was then converted back to the originalbase sequence format by using a C program written byus The detailed information of generating new sequenceswas provided in the Section S11 of Supplementary Materialavailable online at httpsdoiorg10115520175041683 Thecomputational processes of using genetic algorithm were thesame in the generation of two next-generation sequencesNext an RNAfold web server [18] was used to predict andanalyze the DNA secondary structure of the 20 sequences

BioMed Research International 3

(1) Five sequences from [14] are set as the parent sequences

(2) Twenty new sequences are created using the genetic algorithm

(3) RNAfold and RNAComposer web server are used

(4) Protein-DNA simulations are performed using the DiscoveryStudio soware

(5) Eight sequences are selected from the simulation results to perform the QCM experiments

(6) Four sequences are selected from the experiment results to generate the next-generation sequences

(7) Twelve new sequences are produced from the genetic algorithm

(8) Procedures 3 and 4 are implemented

(9) Four sequences are selected from the simulation results and experiments are conducted using the QCM and SPR

(10) DNA aptamers that show high anity with the PSA are identied

Figure 1 Flow chart of the study

in which sequences that failed to form a clear secondarystructure were removed The RNAfold web server is suit-able for predicting secondary structures of DNA and RNAsequences Then an RNAComposer web server [19] wasutilized to build three-dimensional molecular models for theremaining DNA aptamers The method used by the RNA-Composer web server is based on the machine translationprinciple and needs the information of secondary structureprovided by RNAfold and it operates on the RNA FRABASEdatabase After getting the RNA model generated by theRNAComposer web server we used Accelrys DiscoveryStudio (DS) 41 to edit the pyrimidine bases in the structurefile to make uracil become thymine These DNA modelswere used for subsequent docking calculations in molecularsimulations We obtained the three-dimensional structure ofPSA from a molecular structure file numbered ldquo3QUMrdquo inthe Protein Data Bank (PDB) The 3QUM contained notonly the molecular structure of PSA but also that of Fabfragments of two monoclonal antibodies (antibodies werenumbered as ldquo5D5A5rdquo and ldquo5D3D11rdquo) DS 41 was utilized tocarry out molecular simulations The DS software was usedto read the molecular structure files (3QUM) and struc-tures of two antibodies were removed and the structure ofPSA was retained for subsequent simulations This studyused a docking program ZDOCK and the ZRANK scor-ing function to assess the interactions between the DNAaptamers and PSA According to the simulation resultsof first-generation sequences eight DNA aptamers wereselected and synthesized for the quartz crystal microbalance

(QCM) experiments QCM experiments were performedto evaluate the real binding situations between the eightselected sequences of the first round and PSA On the basisof the experiment results four sequences were chosen outof the eight as the parent sequences to produce second-generation sequences A genetic algorithm was utilized againto perform reproduction and crossover operations whichgenerated 12 second-generation sequencesTheRNAfold webserver was used to analyze the secondary structure for eachsecond-generation sequence and then an RNAComposerweb server was employed to build 3D structural models ofselected aptamers for molecular simulations Finally threeaptamer sequences selected from the simulations were usedto perform theQCMand SPR experiments for the assessmentof aptamer-PSA interactions Besides an aptamer named asΔPSap45 in the Savory et alrsquos study [14] showed the highestPSA-binding ability was used as the control groups in theQCM and SPR experiments All QCM and SPR experimentsfor each aptamer were done in triplicate

22 Molecular Simulations The ZDOCK simulation func-tion offered by the DS software was used to assess theinteractions between the DNA aptamers and PSA ZDOCK isa docking method commonly used in protein-protein inter-action simulations The ZDOCK scoring function evaluatesthe protein-protein interactions by taking shape complemen-tary (SC) electrostatics and pairwise atomic potential intoconsideration According to our previous studies [20 21]ZRANK score was more suitable for using in the evaluations


of the interactions betweenproteins and aptamerswith longersequences in length which was due to the results of ZRANKscores that were closer to those obtained from experimentsThis ZRANK program with an optimized energy functioncan significantly improve the success rate of prediction fromthe initial ZDOCK [22] Regarding the equipment used toperform the calculations in this study it comprised HP Z620desktop workstation two Intel Xeon processors (containing24 computing cores) a memory of 36GB and a 64-bitWindows 7 Professional Operating System

23 Reagents and Biological Molecules We purchased thepoly(ethylene glycol) thiol (Thiol-PEG4-Alcohol) fromBroadpharm Inc (San Diego California USA) All ofthe DNA aptamers were synthesized and purchased fromMDBio Inc (Taipei City Taiwan) and each sequence wasmodifiedwith 1-hexanethiol (C6SH) at the 51015840 endThe humanprostate specific antigen was obtained fromMP BiomedicalsLLC (Santa Ana California USA) Sodium dihydrogenphosphate monohydrate (KH2PO4) and sodium chloridewere purchased from Sigma-Aldrich Co LLC (St LouisMissouri USA) Dibasic sodium phosphate (Na2HPO4) andpotassium chloride were acquired from JT Baker (CenterValley Pennsylvania USA) Phosphate-buffered saline (PBS)solution with 1x concentration was used in the experimentswhich contains 10mM dibasic sodium phosphate 137mMsodium chloride 2mM potassium dihydrogen phosphateand 27mM potassium chloride (adjusted to pH 74 withHCl) All other chemicals used in this study were of reagentgrade

24 Quartz Crystal Microbalance (QCM) Instrument Afterstructural analysis and simulations were completed QCMexperiments were used to evaluate the binding reactionsbetween the DNA aptamers and PSA The amount of changein frequency signals caused by the binding reactions was usedto identify which DNA aptamers displayed favorable bindingability with PSA The QCM instrument is produced by CHInstruments Inc (USA) and the model is CHI 410C Thevariations in the crystal oscillator frequency of theQCMwereprimarily related to changes in themass on the sensor surfaceThe relationship formula between oscillation frequency andadsorption quality is expressed in

Δ119891 = minus211989102Δ119898

[119860sqrt (120583120588)] (1)

where 1198910 is the fundamental resonant frequency of crystal 119860is the area of the gold disk on the crystal (0196 cm2) 120588 is thedensity of crystal (2684 gcm3) and 120583 is the shear modulusof quartz (2947 times 1011 gcm lowast s2)

25 Surface Plasmon Resonance (SPR) Instrument The SPRimaging (SPRi) platform used in this study is developed bythe Institute of Photonics andElectronics (IPE Prague CzechRepublic) [23 24] This SPR platform is very sensitive andis capable of detecting the change of refractive index unit(RIU) occurring on the sensing surface better than 10minus6

The p-polarized light beam with a central emission wave-length of 750 nm illuminates on the chip and excites surfaceplasmon waves at the metal-dielectric interface For thisSPR instrument the amount of biomolecular interactions onthe sensing surface can be expressed as biomolecular surfacecoverageThe smallest amount of biomolecular surface cover-age that can be detected is 002 ngcm2The SPRi can performtraditional measurements or carry out a high-throughputmeasurement by using an array chip The SPR chips werefabricated by precoating a thin layer of chromium (thicknessapprox 2 nm) on the BK7 glass substrate and a layer of goldfilm (thickness approx 48 nm) was coated via an evaporationdeposition process afterward All SPR experiments wereperformed at a controlled temperature of 25∘Cwith a constantflow rate of 50 120583lmin The total amount of flow channelsin the SPRi is 6 and one of the channels is utilized as thereference channel After subtracting the data of referencechannel experimental data of aptamer-protein interactionscould be obtained from other five channels

26 Surface Functionalization and Binding ExperimentsPrior to performing the experiment the quartz crystal chipunderwent a modification treatment PEG thiols with an OHfunctional group and synthesized DNAs were all dissolved ina 1M KH2PO4 solution to the concentrations of 19mM and02 120583M respectively Appropriate amounts of OH-PEG thioland the synthesized DNA were taken and mixed together tomake the DNA have a final concentration of 100 nM ThePEG thiol in the mixed solution had a concentration of 5120583Mand the molar ratio of DNAPEG was 1 50 Next 03ml ofDNAPEG mixed solution was applied to cover the gold filmon the surface of the quartz crystal chip The chip was placedin a Petri dish and sealed with parafilm to react for 24 h ina 4∘C environment After the reaction a DNAPEG self-assembledmonolayer (SAM) formed on the chip surfaceThePEG resisted and lowered the nonspecific adsorption of pro-teins as well as providing enough space for preventing three-dimensional steric hindrances duringmolecular interactionsThe modified chip was initially installed inside a chamberin the QCM measurement Initially 33ml of 10mM PBS(pH 74) solution was added to the QCM chamber After theoscillation frequency of the chip stabilized 05ml of PSAsolution (2 ngml) was injected into the device to observe thebinding reaction of the PSA and the aptamer immobilized onthe chip surface

Concerning the experiment performed on the surfaceplasmon resonance biosensor the gold film on the chipwas treated in a manner similar to that on the QCM chipThe surface functionalization of gold film on the chip wasaccomplished by directly immersing in the DNAPEGmixedsolution for 24 h at 4∘C The chips were then rinsed with DIwater and blown dry with nitrogen Afterwards the SPRchip was installed in the SPR instrument and 10mM PBSwas introduced to flow channels through a peristaltic pumpWhen the instrument displayed the SPR signal reaching astable state the solution containing 10mMPBS and PSAwiththe concentration of 2 ngml was injected for 30min Subse-quently PBS solution was again introduced to flow channels


for 13min to obtain the final reaction amount from the spe-cific binding of the aptamer and PSAThe affinity and kineticparameters of aptamer-protein interactions were obtainedby using two mathematical equations based on the first-order kinetics model for fitting SPR sensorgrams [20 25]From the calculation the association rate constant 119896119886 and thedissociation rate constant 119896119889 can be determined In additionthe binding affinity 119870119860 is defined in accordance with therelationship119870119860 = 119896119886119896119889

3 Results and Discussion

31 First-Generation Sequence Calculations and MolecularSimulation Results Calculations were made using a geneticalgorithm which produced 20 sequences as shown in Table 1The third and fourth column of Table 1 are minimumfree energy (kcalmol) and dot-bracket format respectivelywhich were obtained by performing analyses from theRNAfold web server Sequences with a dot-bracket formatcomprising exclusively periods (eg PSAG16) signified thatsuch sequences did not display a clear secondary structureIn other words they did not have a clear stem-loop structureand were not appropriate aptamer candidates For suchsequences no next-stage molecular simulation analysis wasperformed For other sequences 3D structural models werebuilt using the RNAComposer web server and these modelswere then edited in the DS software Graphic results of 3Dstructural models built by the RNAComposer web server andedited by the DS software are shown in Figure S1 (in theSupplementary Material) The structural models of aptamersand the PSA then underwent ZDOCK simulations witheach molecular simulation calculation lasting approximately15 h long The simulations produced ZDOCK and ZRANKscores for these sequences in which those with a lowerZRANK score indicated a superior docking result Accordingto the molecular simulations first-generation aptamers thatshowed the most favorable binding capability with PSAlisted in descending order were PSAG11 PSAG114 PSAG119PSAG118 PSAG112 PSAG117 PSAG15 and PSAG17

32 QCM Experimental Results of the First-GenerationSequences The eight sequences described in Section 31(ie PSAG11 PSAG15 PSAG17 PSAG112 PSAG114 PSAG117PSAG118 and PSAG119) were sent to a vendor to synthesizebefore QCM experiments were conductedThe experimentalresults are shown in Table 2 and indicate that the eightsequences can bind to PSA resulting in relatively largerchanges to the QCM oscillation frequency (ie drops inoscillation frequency) The changes in QCM oscillationfrequency generated by the sequences listed in descendingorder were PSAG15 PSAG119 PSAG117 PSAG112 PSAG114PSAG17 PSAG118 and PSAG11 However most of rankingresults from experimental and simulation studies are notconsistent Concerning the ranking of the eight sequences inexperiments only PSAG112 and PSAG119 showed a rankingthat closely matched those obtained from the molecularsimulations One of the reasons is that we supposed thatthe docking method used in this study is not the optimal

algorithm for the evaluation of aptamer-protein interactionSo far much fewer algorithms for protein-DNA dockinghave been specifically developedThe fast Fourier correlationtechniques calculate shape complementarity and are mainlyused in the protein-protein docking field which are appliedin the computational study of proteinDNA complexes [26]TheZDOCKused in this study is also a fast Fourier transformbased docking algorithm that searches all possible bindingmodes of complexes and includes the pairwise shape comple-mentarity in the evaluation Although the results of simula-tion and experiment are not fully consistent the experimentalresults show that most of these selected aptamer sequencescan react with PSA and produce significant signal changes inthe frequency of QCM chip In the study reported by Binirsquosgroup [11] they found the computational approach partiallyconfirmed results observed in SELEX that the TBA bindingscore corresponded to only 813 of the best candidateWe think that the information provided from molecularsimulations is still valuable to combine with experiments forthe selection of aptamers

In order to obtain better simulation results of protein-nucleic acid interactions some researchers are dedicatedto developing new docking algorithms for protein-DNAcomplexes Banitt andWolfson [27] reported a novel protein-DNA docking algorithm named as ParaDock for dockingshort DNA fragments to the protein based on geometriccomplementarity in 2011 Another web server for protein-nucleic acid docking called NPDock used specific protein-nucleic acid statistical potentials for scoring and selection ofmodeled complexes [28] A distance dependent knowledge-based coarse grained force field is recently developed by Setnyet al [29] for evaluating protein-DNAdockingThe force fieldcan improve the quality of predictions in the protein-DNAdocking and they find shape complementarity and sequence-dependent DNA internal energy have great contribution tothe specific protein-DNA interaction

Besides we supposed there were another two reasonsto explain the inconsistency between the simulation andexperimental results (1) The modeling methods used herecan build acceptable and reasonable models of single-strandDNAs but the DNA model may not fully present theactual situation (2) In the measurements of biosensors theimmobilization of aptamer on sensor surface may influencethe interaction between the aptamer and the target proteinFor the first point the RNAComposer web server couldgenerate the model of RNA automatically combined withthe information of secondary structure of DNA sequencepredicted by RNAfold web server In order to build themodel of DNA aptamer we had ever considered using othersoftware or web servers like the 3D-DART web server [30]However users needed to input some structural parametersfor building the model of single-stranded DNA such as thenucleic type (A-form or B-form) global and local bend-angle and location of bend-angle The technique for fullyautomated prediction of DNA 3D structures is still absent sofar Although we edited the atoms of models generated fromRNAComposer web server to make the RNA model becomea DNA model we supposed that the DNA model comparedwith the three-dimensional structure of DNA in real situation


Table1First-g

enerationsequ

encesa

ndresults

from

thea

nalysis

andcalculations

Nam

eSequ

ence

(51015840-31015840)

Minim

umfre

eenergy(kcalm

ol)

Dot-bracketform

atZD

OCK

score

ZRANKscore

Rank

PSAG

11TT

TTTC

TGTT

GCC

CGGAAC

GTC

GTG

GCC

CTTT

minus56

(((((())))))

1734

minus8144

1PS

AG12

TTTT

TGTG

GTG

TTTA

TTGTT

TACT

GTC

CCTT

Tminus02

(((())))

4927

minus6091

PSAG

13TT

TTTC

TGGTG

TTTA

TTCC

ATCA

AAT

ATCT

TTminus31

((((((()))))))

4003

minus6033

PSAG

14TT

TTTA

ATAT

CAAC

TTGGTT

TACT

GTC

CCTT

Tminus02

(())

4393

minus7008

PSAG

15TT

TTTC

TGGAAT

GAT

TTCC

CGGTT

GTC

TCTT

Tminus33

(((((())))))

3878

minus724

7PS

AG16

TTTT

TACT

TTGTT

TATT

GTT

TACT

GTC

CCTT

T0

NA

NA

PSAG

17TT

TTTC

TGGTC

CGGGTA

CGTT

TTTT

GGCC

TTT

minus48

((((((()))))))

3973

minus7149

8PS

AG18

TTTT

TCCG

CAGTT

TATT

GTT

TACT

GTC

CCTT

Tminus35

((((()))))

4005

minus6755

PSAG

19TT

TTTG

TGTT

GCC

CGGAAC

GTC

GTA

TATC

TTT

minus08

(((((())))))

4816

minus6889

PSAG

110TT

TTTA

ATAT

CAAC

TTGCC

ATCA

AGGCC

CTTT

minus33

((()))

4501

minus6888

PSAG

111

TTTT

TGTG

TTGCC

ATTT

CCCG

GTT

GTC

TCTT

Tminus13

((()))

486

minus6473

PSAG

112

TTTT

TACT

TAAT

GCG

GAAC

GTC

GTG

GCC

CTTT

minus14

(((())))

4258

minus7689

5PS

AG113

TTTT

TGTG

TTGCC

CGGAAC

TTTT

TTGGCC

TTT

minus3

(((((())))))

3977

minus6701

PSAG

114TT

TTTC

CGCA

CCGGGTA

CGGTC

GTG

GCC

CTTT

minus108

(((((((())))))))

4345

minus7891

2PS

A G115

TTTT

TAAT

ATCA

ACTT

GCC

AGGTT

GTC

TCTT

Tminus27

(((((())))))

4436

minus6184

PSAG

116TT

TTTA

CTTA

ATGAT

TTCC

CTCA

AAT

ATCT

TT0

NA

NA

PSAG

117TT

TTTA

ATAT

CCGGGTA

CGTT

TTTT

GGCC

TTT

minus24

(((((())))))

4752

minus7457

6PS

AG118

TTTT

TCCG

CACA

ACTT

GCC

ATCA

AAT

ATCT

TTminus11

((()))

4613

minus769

4PS

AG119

TTTT

TACG

CACC

GGGTA

CGTT

TTTT

GGCC

TTT

minus24

(((((())))))

4041

minus7784

3PS

AG120

TTTT

TCCT

TAAT

GAT

TTCC

CGGTT

GTC

TCTT

T0

NA

NA


Table 2 Selected first-generation sequences and their bindingreactions with the PSA as measured by the QCM

Name

Average andstandard

deviation valuesof frequencychanges

Ranked resultsin experiments

Ranked resultsin simulations

PSAG11 192 plusmn 17 8 1PSAG15 1082 plusmn 14 1 7PSAG17 373 plusmn 19 6 8PSAG112 717 plusmn 13 4 5PSAG114 63 plusmn 54 5 2PSAG117 829 plusmn 58 3 6PSAG118 249 plusmn 39 7 4PSAG119 913 plusmn 97 2 3

might be slightly different Nevertheless we consider thisapproach is able to build an acceptable and reasonable modelof single-strand DNA based on the currently available toolsThe limitation caused by the correctness of DNA model maybe a possible reason to explain that the results of simulationstudies are not fully consistent with experimental results

The second point is an unavoidable issue between theexperiment and simulation The orientation of immobilizedaptamer on the sensor surface may influence the interactionwith the target protein In our previous study [22] we usedthe RNAComposer web server to generate models of RNAsequences for studying the RNA-protein interaction Wealso found that our experimental findings on the biosensorwere not fully consistent with the computational results Onthe other hand the detection scheme on biosensors is alsocommonly used in the enzyme-linked immunosorbent assay(ELISA) except that no labeled secondary probe agent is usedThemeasured signals on biosensors can present the potentialsof aptamers in the medical applications

33 Second-Generation Sequence Calculations and MolecularSimulation Results On the basis of the QCM experimentalresults of the first-generation sequences we selected PSAG15PSAG112 PSAG117 and PSAG119 as the parent sequencesto generate second-generation sequences Similar to thecalculations on the first-generation sequences the processfor producing second-generation sequences was made usingthe genetic algorithm which produced 12 sequences (asshown in Table 3) The secondary structure of the sequenceswas analyzed using the RNAfold web server According tothe analysis results of the sequences PSAG27 and PSAG211were removed For other sequences DNA molecular modelswere built using the RNAComposer web server and theresults obtained from docking simulations by using ZDOCKwere shown in Table 3 The three sequences with the mostfavorable ZRANK scores listed in descending order werePSAG28 PSAG212 and PSAG24These three sequences weresubsequently selected for the next stage of the experiment ADNA aptamer sequence ΔPSap45 introduced in a study byAnderson andMecozzi [10] was used as the baseline sequence

PSAG24PSAG28PSAG212

ΔPSap45

minus200

minus150

minus100

minus50

0

50

Freq

uenc

y ch

ange

s (H

z)

2 4 6 8 10 12 14 16 18 200Time (minute)

Figure 2TheQCMexperiments for the second-generation aptamersequences Frequency variations produced by the binding reactionsbetween second-generation sequences and PSA in the QCM exper-iments

for comparisons Simulations and experiments were alsoperformed for the sequence ΔPSap45 and this sequencereceived a ZDOCK and ZRANK score of 451 and minus70 in thesimulations respectively

34 QCM Experimental Results of the Second-GenerationSequences The second-generation sequences selected weresynthesized for a QCM experiment which produced theresults shown in Figure 2 The experiment data are compiledin Table 4 The experiment results showed the amount ofsignals generated by the second-generation sequences duringthe experiment in which only PSAG212 generated a signalsmaller than that generated by ΔPSap45 Concerning theother two selected sequences they produced clear signalswhen interacting with PSA of an identical concentrationIn particular PSAG28 produced an amount of signal whichapproximately was 2 times higher than that produced by thebest sequence ΔPSap45 introduced in an existing literatureBesides PSAG24 could generate an amount of signal that wasalmost 17 times higher than that of ΔPSap45

35 Measuring the Binding Reactions between Second-Generation Nucleic Acid Aptamer Sequences and PSA byUsingthe SPR Biosensor To verify the binding reactions betweensecond-generation nucleic acid aptamer sequences and PSAwe performed an experiment by using the SPR biosensorThe SPR experiment results are shown in Figure 3 whichreveals that the largest binding amount was produced bythe interaction between PSA and PSAG28 immobilized onthe sensor surface The amount of change in SPR signalwas equivalent to a biomolecular surface coverage of 255 plusmn023 ngmm2 on the sensor surface A largest binding amountwas also generated when PSA reacted with PSAG24 fixed on


Table3Second

-generationsequ

encesa

ndanalysisandcalculationresults

Nam

eSequ

ence

(51015840-31015840)

Minim

umfre

eenergy(kcalm

ol)

Dot-bracketform

atZD

OCK

score

ZRANKscore

Rank

PSAG

21TT

TTTC

TGGAAT

GAT

TAAC

GTC

GTG

GCC

CTTT

minus33

((((((()))))))

4344

minus6868

PSAG

22TT

TTTA

CTTA

ATGCG

GTC

CCGGGGGGCT

CTTT

minus49

((((()))))

4577

minus6916

PSAG

23TT

TTTC

TGGAAT

GAT

TACG

TTTT

TTGGCC

TTT

minus1

(((((())))))

4161

minus5612

PSAG

24TT

TTTA

ATAT

CCGGGTT

CCCG

GTT

GTC

TCTT

Tminus67

((((((()))))))

1754

minus7091

3PS

AG25

TTTT

TCTG

GAAT

GAT

TTCC

CGGTT

TGGCC

TTT

minus43

(((())))((()))

4569

minus5504

PSAG

26TT

TTTA

CGCA

CCGGGTA

CGTT

TTTG

TCTC

TTT

minus26

(((((())))))

4463

minus6012

PSAG

27TT

TTTA

CTTA

ATGCG

TACG

TTTT

TTGGCC

TTT

0

NA

NA

PSAG

28TT

TTTA

ATAT

CCGGGGAAC

GTC

GTG

GCC

CTTT

minus41

((((()))))

4264

minus86

1PS

AG29

TTTT

TACT

TAAT

GCG

GAC

GTT

TTTT

GGCC

TTT

minus17

(((())))

4451

minus6583

PSAG

210

TTTT

TACG

CACC

GGGTA

ACGTC

GTG

GCC

CTTT

minus54

(((((())))))

4261

minus685

PSAG

211

TTTT

TACT

TAAT

GAAT

ACGTT

TTTT

GGCC

TTT

0

NA

NA

PSAG

212

TTTT

TACG

CACC

GGGTA

CGTT

TTTT

GGCC

TTT

minus24

(((((())))))

4041

minus7784

2


Table 4 Selected second-generation sequences and their bindingreactions with the PSA in QCMmeasurements

Name of aptamer

Average and standarddeviation values offrequency changes

(Hz)

Ranked results inexperiments

PSAG24 1313 plusmn 49 2PSAG28 1662 plusmn 67 1PSAG212 595 plusmn 58 4ΔPSap45lowast 761 plusmn 21 3lowastA sequence introduced in a study by Savory et al [14] that showed highaffinity with PSA

Table 5 Kinetic parameters of aptamer-PSA interactions by fittingSPR sensorgrams

Name of aptamer 119896119886

(times103Mminus1sminus1) 119896119889 (times10minus3 sminus1) 119870

119860

(times106Mminus1)PSAG24 337 plusmn 062 1234 plusmn 058 027 plusmn 004PSAG28 662 plusmn 077 67 plusmn 037 099 plusmn 006PSAG212 311 plusmn 05 1123 plusmn 08 028 plusmn 003ΔPSap45 42 plusmn 062 708 plusmn 023 059 plusmn 007

the surface the amount of change in SPR signal was equiv-alent to a change of 089 plusmn 02 ngcm2 in the biomolecularsurface coverage RegardingΔPSap45 it induced a change of077plusmn 019 ngcm2 in the experiments PSAG212 displayed theworst performance generating a change of 065plusmn 018 ngcm2in the biomolecular surface coverage In essence the SPRexperiment results were consistent with those obtained usingthe QCM

The kinetic parameters of these binding reactions werecalculated from global fitting of the SPR sensorgram dataRepresentative SPR curves for these aptamer-PSA interac-tions are shown in Figure 4 It was worth noting that PSAG24could produce a large and apparent signal in the associationphase of reaction in the sensorgram but the signal decreaseddramatically in the dissociation stage Table 5 shows thekinetic parameters for these four kinds of aptamer-PSAinteractions According to the values binding affinity (119870119860)the sequences had the most favorable binding affinity withPSA listed in descending order being PSAG28 PSap45PSAG24 and PSAG212 PSAG24 could produce a slightlylarger amount of change in the average SPR signal thanthat of ΔPSap45 (shown in Figure 3) yet the bindingability of ΔPSap45 to PSA was better in the viewpointof kinetic parameters Concerning the amount of SPR sig-nal created by the binding reactions that of PSAG28 wasalmost 3 times higher than that of ΔPSap45 PSAG28 evenshowed superior kinetic parameters comparedwith other twoaptamers selected from the second-generation sequences andΔPSap45 Thus we successfully selected an aptamer withhigh affinity for PSA by using the combination of com-putational approaches and biosensing technologies in thisstudy We consider that PSAG28 is a suitable aptamer for

PSAG24 PSAG28 PSAG2120

05

1

15

2

25

3

Adso

rptio

n pe

r are

a (ng

cm

2)

ΔPSap45

Figure 3 The data of SPR experiments for the second-generationaptamer sequences Binding reactions between different nucleic acidaptamer sequences and PSAmeasured by using the surface plasmonresonance biosensor

00

2

4

6

8

10

12

14

Adso

rptio

n pe

r are

a (ng

cm

2)

3020 40 5010Time (min)

PSAG24PSAG28

PSAG212ΔPSap45

Figure 4 SPR sensorgrams Representative SPR curves for thesefour kinds of aptamer-PSA interactions

molecular recognition element in the aptasensors and it alsohas potential in other biomedical applications

4 Conclusions

In this study structural analyses molecular simulationsand biosensor experiments were used to identify DNAaptamers that featured high binding affinity with PSA Byusing nucleic acid aptamer sequence analysis tools suchas RNAfold web server sequences that did not form aclear secondary structure could be eliminated Furthermoremolecular simulation results gave us information to excludeaptamers that were predicted with poor binding to PSA fromaptamer candidates which effectively reduced the number of


experiments in the aptamer selectionThrough the screeningwe successfully found one sequence (ie PSAG28) that couldproduce a relatively superior binding reaction with PSAThisaptamer showed almost 3-fold higher binding signal in theSPR experiment than that of the best PSA-binding aptamerreported previously ΔPSap45 Besides kinetic parametersfor the evaluation of interaction between the aptamer andPSA also demonstrated that PSAG28 was the best aptamerThis study reveals that the computational approach is valuablefor use in the post-SELEX screening procedure for facilitatingthe preselection of aptamer candidates that will be tested inthe further screening experiments For getting much bettersimulation results the tool that can predict andmodel the 3Dstructures of single-stranded DNA sequences precisely andautomatically is still a strong demand

Conflicts of Interest

The authors declare that there are no conflicts of interestregarding the publication of this paper

Acknowledgments

The authors gratefully acknowledge the financial supportprovided by Ministry of Science and Technology Taiwanunder Contract nos MOST 103-2221-E-468-019 MOST 102-2632-E-468 -001-MY3 and MOST 105-2221-E-468-017

References

[1] B Lojanapiwat W Anutrakulchai W Chongruksut and CUdomphot ldquoCorrelation and diagnostic performance of theprostate-specific antigen level with the diagnosis aggressive-ness and bonemetastasis of prostate cancer in clinical practicerdquoProstate International vol 2 no 3 pp 133ndash139 2014

[2] W J Catalona D S Smith andD K Ornstein ldquoProstate cancerdetection in men with serum PSA concentrations of 26 to40 ngmL and benign prostate examination enhancement ofspecificity with free PSA measurementsrdquo Journal of the Ameri-can Medical Association vol 277 no 18 pp 1452ndash1455 1997

[3] P Jolly N Formisano and P Estrela ldquoDNA aptamer-baseddetection of prostate cancerrdquo Chemical Papers vol 69 no 1 pp77ndash89 2015

[4] C Tuerk and L Gold ldquoSystematic evolution of ligands byexponential enrichment RNA ligands to bacteriophage T4DNApolymeraserdquo Science vol 249 no 4968 pp 505ndash510 1990

[5] A D Ellington and J W Szostak ldquoIn vitro selection of RNAmolecules that bind specific ligandsrdquo Nature vol 346 no 6287pp 818ndash822 1990

[6] D L Robertson and G F Joyce ldquoSelection in vitro of an RNAenzyme that specifically cleaves single-stranded DNArdquo Naturevol 344 no 6265 pp 467ndash468 1990

[7] C L A Hamula J W Guthrie H Zhang X-F Li andX C Le ldquoSelection and analytical applications of aptamersrdquoTrACmdashTrends in Analytical Chemistry vol 25 no 7 pp 681ndash691 2006

[8] V Thiviyanathan and D G Gorenstein ldquoAptamers and thenext generation of diagnostic reagentsrdquo ProteomicsmdashClinicalApplications vol 6 no 11-12 pp 563ndash573 2012

[9] R Rajamani andACGood ldquoRanking poses in structure-basedlead discovery and optimization current trends in scoringfunction developmentrdquo Current Opinion in Drug Discovery andDevelopment vol 10 no 3 pp 308ndash315 2007

[10] P C Anderson and S Mecozzi ldquoIdentification of a 14mer RNAthat recognizes and binds flavin mononucleotide with highaffinityrdquo Nucleic Acids Research vol 33 no 22 pp 6992ndash69992005

[11] A Bini M Mascini M Mascini and A P F Turner ldquoSelec-tion of thrombin-binding aptamers by using computationalapproach for aptasensor applicationrdquo Biosensors and Bioelec-tronics vol 26 no 11 pp 4411ndash4416 2011

[12] S E Lupold B J Hicke Y Lin and D S Coffey ldquoIdentificationand characterization of nuclease-stabilized RNA moleculesthat bind human prostate cancer cells via the prostate-specificmembrane antigenrdquo Cancer Research vol 62 no 14 pp 4029ndash4033 2002

[13] S Jeong S R Han Y J Lee and S-W Lee ldquoSelection of RNAaptamers specific to active prostate-specific antigenrdquo Biotech-nology Letters vol 32 no 3 pp 379ndash385 2010

[14] N Savory K Abe K Sode and K Ikebukuro ldquoSelectionof DNA aptamer against prostate specific antigen using agenetic algorithm and application to sensingrdquo Biosensors andBioelectronics vol 26 no 4 pp 1386ndash1391 2010

[15] B Liu L Lu E Hua S Jiang and G Xie ldquoDetection of thehuman prostate-specific antigen using an aptasensor with goldnanoparticles encapsulated by graphitizedmesoporous carbonrdquoMicrochimica Acta vol 178 no 1-2 pp 163ndash170 2012

[16] Z Chen Y Lei X Chen ZWang and J Liu ldquoAn aptamer basedresonance light scattering assay of prostate specific antigenrdquoBiosensors and Bioelectronics vol 36 no 1 pp 35ndash40 2012

[17] A J Chipperfield and P J Fleming ldquoThe MATLAB geneticalgorithm toolboxrdquo in Proceedings of the IEE Colloquium onApplied Control Techniques Using MATLAB pp 101ndash104January 1995

[18] RNAfold WebServer httprnatbiunivieacatcgi-binRNA-WebSuiteRNAfoldcgi

[19] M Popenda M Szachniuk M Antczak et al ldquoAutomated 3Dstructure composition for large RNAsrdquo Nucleic Acids Researchvol 40 no 14 article e112 2012

[20] B Pierce and Z Weng ldquoZRANK Reranking protein dockingpredictions with an optimized energy functionrdquo Proteins vol67 no 4 pp 1078ndash1086 2007

[21] W-P Hu J V Kumar C-J Huang and W-Y Chen ldquoCompu-tational selection of RNA aptamer against angiopoietin-2 andexperimental evaluationrdquo BioMed Research International vol2015 Article ID 658712 8 pages 2015

[22] J V Kumar J J P Tsai W P Hu andW Y Chen ldquoComparativemolecular simulation method for Ang2aptamers with in vitrostudiesrdquo International Journal of Pharma Medicine and Biologi-cal Sciences vol 4 no 1 pp 2ndash5 2015

[23] M Piliarik and J Homola ldquoSelf-referencing SPR imagingfor most demanding high-throughput screening applicationsrdquoSensors and Actuators B Chemical vol 134 no 2 pp 353ndash3552008

[24] M Piliarik M Bockova and J Homola ldquoSurface plasmon reso-nance biosensor for parallelized detection of protein biomarkersin diluted blood plasmardquo Biosensors and Bioelectronics vol 26no 4 pp 1656ndash1661 2010

[25] W-P Hu G-L Chang S-J Chen and Y-M Kuo ldquoKineticanalysis of 120573-amyloid peptide aggregation induced by metal


ions based on surface plasmon resonance biosensingrdquo Journalof Neuroscience Methods vol 154 no 1-2 pp 190ndash197 2006

[26] P Aloy G Moont H A Gabb E Querol F X Aviles and M JE Sternberg ldquoModelling repressor proteins docking to DNArdquoProteins Structure Function andGenetics vol 33 no 4 pp 535ndash549 1998

[27] I Banitt and H J Wolfson ldquoParaDock a flexible non-specific DNAmdashrigid protein docking algorithmrdquo Nucleic AcidsResearch vol 39 no 20 article e135 2011

[28] I Tuszynska M Magnus K Jonak W Dawson and J MBujnicki ldquoNPDock a web server for proteinndashnucleic aciddockingrdquo Nucleic Acids Research vol 43 no 1 pp W425ndashW430 2015

[29] P Setny R P Bahadur and M Zacharias ldquoProtein-DNAdocking with a coarse-grained force fieldrdquo BMC Bioinformaticsvol 13 no 1 article 228 2012

[30] M vanDijk andAM J J Bonvin ldquo3D-DART aDNA structuremodelling serverrdquoNucleic Acids Research vol 37 supplement 2pp W235ndashW239 2009

Submit your manuscripts athttpswwwhindawicom

Stem CellsInternational

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014


MEDIATORSINFLAMMATION

of


Behavioural Neurology

EndocrinologyInternational Journal of



Disease Markers


BioMed Research International

OncologyJournal of



Oxidative Medicine and Cellular Longevity


PPAR Research

The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014

Immunology ResearchHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Journal of

ObesityJournal of



Computational and Mathematical Methods in Medicine

OphthalmologyJournal of


Diabetes ResearchJournal of



Research and TreatmentAIDS


Gastroenterology Research and Practice


Parkinsonrsquos Disease

Evidence-Based Complementary and Alternative Medicine

Volume 2014Hindawi Publishing Corporationhttpwwwhindawicom


molecule can be obtained In the partitioning step the target-bound nucleic acids are separated from unbound sequencesBound nucleic acids are then amplified to generate a newpool of nucleic acid sequences for use in the next round ofselection After repeating 8sim20 cycles nucleic acid aptamerswith high affinity and specificity for the target moleculecan be obtained In essence aptamers are short nucleic acidsequences with a length of 12 to 80 nucleic acids Aptamersthat have been identified by current research include RNAand DNA molecules and RNA-type aptamers account forthe majority of aptamers Aptamers can be used to reactwith various target molecules such as metal ions shortpeptides pathogenic microorganisms micromolecules suchas amino acids andmacromolecules like proteins Comparedwith antibodies aptamers are manufactured by chemicalsynthesis therefore batch-to-batch variation can be greatlyminimized The economical high-accuracy large-scale pro-duction of aptamers makes aptamers very suitable for clinicalapplications [8] In addition nucleic acid aptamers featureadvantages not found in common antibodies which are listedas follows small molecules high stability easy productionpossibility for reuse and unchanged affinity after simplechemical modifications [8] Because of these advantagesaptamers can be used in various biotechnology-related fieldssuch as environment and food quality tests therapies drugdevelopments purification processes and diagnoses

Previous studies have shown that the use of appropriatemolecular simulation software is highly beneficial for design-ing molecules with high affinity and selectivity [9] Advancesin computer software and hardware over the past severaldecades as well as developments in mathematical modelshave resulted in increasing researchers using computersto simulate and predict the reactions of biomolecules andchemical molecules Anderson and Mecozzi [10] employedthe molecular simulation method in addition to conductingtraditional experiments to investigate the interaction of the35-mer aptamer and cofactor flavin mononucleotide (FMN)Their study successfully reduced the length of the 35-meraptamer to 14mers while retaining the binding ability forflavin mononucleotide

In 2011 Bini et al [11] proposed a computationally assistedmethod to study and evaluate aptamer-protein interaction forselecting new aptamer sequencesThey performedmolecularsimulations by using a molecular model and conducted anexperiment to verify the prediction results of the molecularsimulations Bini et al initially used a well-known 15-merthrombin binding aptamer (TBA) and then they modifiedbase sequences at certain locations to generate a varietyof mutant aptamer sequences After docking simulationsexperiments were performed by using a surface plasmonresonance (SPR) biosensor The results indicated a mutantsequence could have a slightly better performance than thewell-known TBA and they found that experimental resultswere in agreement with the simulation findings Lupoldet al [12] reported the specific binding between an RNAaptamer and a prostate specific membrane antigen (PSMA)the RNA aptamer demonstrated the potential to be usedas a drug carrier to treat prostate cancer In 2010 Jeonget al [13] researched and found an RNA aptamer that can

be used to identify PSA the RNA aptamer featured a 51015840-end to 31015840-end sequence of CCGUAGGUCACGGCAGCG-AAGCUCUAGGCGCGGCCAGUUGC Savory et al [14]selected five DNA sequences by using the SELEX techniqueand performed subsequent in silico analyses With the use ofselected five DNA sequences as parent sequences the geneticalgorithm (GA) was adopted to generate next-generationsequences and those generated sequences were synthesizedand assayed in vitro One of the fourth-generation DNAaptamers exhibited a 48-fold higher PSA-binding ability thanthe parent sequences

Some scholars have identified RNA aptamers as a possiblealternative to antibodies for testing serum PSA concentration[12 13] Nevertheless the long length of RNA aptamerscould cause the difficulty in the commercial synthesis andlimit the applications [3] By contrast DNA-based PSAaptamers had been used to detect PSA by using differentsensing technologies [15 16] Herein we reported a studythat employed computational approaches including GA theanalysis of DNA secondary structure and the molecularsimulation to evaluate the aptamer-protein interactions andbiosensing technologies to verify the PSA-binding ability ofselected aptamers The original five sequences of aptamersreported in the study of Savory et al [14] were chosen asthe parent sequences and the genetic algorithm was usedto generate new nucleic acid aptamers for the selection Byusing the proposed strategy of selection a newDNA aptamerexhibited stronger binding capability with PSA which wasselected in the end of selection procedure We think theproposed strategy of combining computational approaches(GA and molecular simulations) in the selection of aptamercan streamline the number of experiments carried out forthe selection and the results show that it is possible tocomplement SELEX for the selection of aptamer

2 Materials and Methods

21The Overview of Research Process Figure 1 shows all stepsof aptamer selection in this study First five sequences thatcan bind to PSA (obtained by Savory et al by using the SELEXprocedure [14]) were set as the parent sequences Savory et alnamed the five sequences as PSap4-3 PSap4-4 PSap4-11PSap4-9 and PSap4-6 To enable the genetic algorithm tocalculate the four bases used in this study (ie A T G andC) such bases were coded bases A T G and C for the fivesequences were numbered 00 01 10 and 11 respectively Thegenetic algorithm toolbox for MATLAB [17] was utilized toperform reproduction and crossover operations for generat-ing 20 next-generation sequences The sequence information(in numerical format) was then converted back to the originalbase sequence format by using a C program written byus The detailed information of generating new sequenceswas provided in the Section S11 of Supplementary Materialavailable online at httpsdoiorg10115520175041683 Thecomputational processes of using genetic algorithm were thesame in the generation of two next-generation sequencesNext an RNAfold web server [18] was used to predict andanalyze the DNA secondary structure of the 20 sequences




















Δ119891 = minus211989102Δ119898

[119860sqrt (120583120588)] (1)















Table1First-g

enerationsequ

encesa

ndresults

from

thea

nalysis

andcalculations

Nam

eSequ

ence

(51015840-31015840)

Minim

umfre

eenergy(kcalm

ol)

Dot-bracketform

atZD

OCK

score

ZRANKscore

Rank

PSAG

11TT

TTTC

TGTT

GCC

CGGAAC

GTC

GTG

GCC

CTTT

minus56

(((((())))))

1734

minus8144

1PS

AG12

TTTT

TGTG

GTG

TTTA

TTGTT

TACT

GTC

CCTT

Tminus02

(((())))

4927

minus6091

PSAG

13TT

TTTC

TGGTG

TTTA

TTCC

ATCA

AAT

ATCT

TTminus31

((((((()))))))

4003

minus6033

PSAG

14TT

TTTA

ATAT

CAAC

TTGGTT

TACT

GTC

CCTT

Tminus02

(())

4393

minus7008

PSAG

15TT

TTTC

TGGAAT

GAT

TTCC

CGGTT

GTC

TCTT

Tminus33

(((((())))))

3878

minus724

7PS

AG16

TTTT

TACT

TTGTT

TATT

GTT

TACT

GTC

CCTT

T0

NA

NA

PSAG

17TT

TTTC

TGGTC

CGGGTA

CGTT

TTTT

GGCC

TTT

minus48

((((((()))))))

3973

minus7149

8PS

AG18

TTTT

TCCG

CAGTT

TATT

GTT

TACT

GTC

CCTT

Tminus35

((((()))))

4005

minus6755

PSAG

19TT

TTTG

TGTT

GCC

CGGAAC

GTC

GTA

TATC

TTT

minus08

(((((())))))

4816

minus6889

PSAG

110TT

TTTA

ATAT

CAAC

TTGCC

ATCA

AGGCC

CTTT

minus33

((()))

4501

minus6888

PSAG

111

TTTT

TGTG

TTGCC

ATTT

CCCG

GTT

GTC

TCTT

Tminus13

((()))

486

minus6473

PSAG

112

TTTT

TACT

TAAT

GCG

GAAC

GTC

GTG

GCC

CTTT

minus14

(((())))

4258

minus7689

5PS

AG113

TTTT

TGTG

TTGCC

CGGAAC

TTTT

TTGGCC

TTT

minus3

(((((())))))

3977

minus6701

PSAG

114TT

TTTC

CGCA

CCGGGTA

CGGTC

GTG

GCC

CTTT

minus108

(((((((())))))))

4345

minus7891

2PS

A G115

TTTT

TAAT

ATCA

ACTT

GCC

AGGTT

GTC

TCTT

Tminus27

(((((())))))

4436

minus6184

PSAG

116TT

TTTA

CTTA

ATGAT

TTCC

CTCA

AAT

ATCT

TT0

NA

NA

PSAG

117TT

TTTA

ATAT

CCGGGTA

CGTT

TTTT

GGCC

TTT

minus24

(((((())))))

4752

minus7457

6PS

AG118

TTTT

TCCG

CACA

ACTT

GCC

ATCA

AAT

ATCT

TTminus11

((()))

4613

minus769

4PS

AG119

TTTT

TACG

CACC

GGGTA

CGTT

TTTT

GGCC

TTT

minus24

(((((())))))

4041

minus7784

3PS

AG120

TTTT

TCCT

TAAT

GAT

TTCC

CGGTT

GTC

TCTT

T0

NA

NA



Name

Average andstandard








PSAG24PSAG28PSAG212

ΔPSap45

minus200

minus150

minus100

minus50

0

50

Freq

uenc

y ch

ange

s (H

z)

2 4 6 8 10 12 14 16 18 200Time (minute)






Table3Second

-generationsequ

encesa


Nam

eSequ

ence

(51015840-31015840)

Minim

umfre

eenergy(kcalm

ol)

Dot-bracketform

atZD

OCK

score

ZRANKscore

Rank

PSAG

21TT

TTTC

TGGAAT

GAT

TAAC

GTC

GTG

GCC

CTTT

minus33

((((((()))))))

4344

minus6868

PSAG

22TT

TTTA

CTTA

ATGCG

GTC

CCGGGGGGCT

CTTT

minus49

((((()))))

4577

minus6916

PSAG

23TT

TTTC

TGGAAT

GAT

TACG

TTTT

TTGGCC

TTT

minus1

(((((())))))

4161

minus5612

PSAG

24TT

TTTA

ATAT

CCGGGTT

CCCG

GTT

GTC

TCTT

Tminus67

((((((()))))))

1754

minus7091

3PS

AG25

TTTT

TCTG

GAAT

GAT

TTCC

CGGTT

TGGCC

TTT

minus43

(((())))((()))

4569

minus5504

PSAG

26TT

TTTA

CGCA

CCGGGTA

CGTT

TTTG

TCTC

TTT

minus26

(((((())))))

4463

minus6012

PSAG

27TT

TTTA

CTTA

ATGCG

TACG

TTTT

TTGGCC

TTT

0

NA

NA

PSAG

28TT

TTTA

ATAT

CCGGGGAAC

GTC

GTG

GCC

CTTT

minus41

((((()))))

4264

minus86

1PS

AG29

TTTT

TACT

TAAT

GCG

GAC

GTT

TTTT

GGCC

TTT

minus17

(((())))

4451

minus6583

PSAG

210

TTTT

TACG

CACC

GGGTA

ACGTC

GTG

GCC

CTTT

minus54

(((((())))))

4261

minus685

PSAG

211

TTTT

TACT

TAAT

GAAT

ACGTT

TTTT

GGCC

TTT

0

NA

NA

PSAG

212

TTTT

TACG

CACC

GGGTA

CGTT

TTTT

GGCC

TTT

minus24

(((((())))))

4041

minus7784

2



Name of aptamer


(Hz)






119860





05

1

15

2

25

3

Adso

rptio

n pe

r are

a (ng

cm

2)

ΔPSap45


00

2

4

6

8

10

12

14

Adso

rptio

n pe

r are

a (ng

cm

2)

3020 40 5010Time (min)

PSAG24PSAG28

PSAG212ΔPSap45



4 Conclusions






Acknowledgments


References






































of






Disease Markers



OncologyJournal of





PPAR Research



Journal of

ObesityJournal of



































Δ119891 = minus211989102Δ119898

[119860sqrt (120583120588)] (1)















Table1First-g

enerationsequ

encesa

ndresults

from

thea

nalysis

andcalculations

Nam

eSequ

ence

(51015840-31015840)

Minim

umfre

eenergy(kcalm

ol)

Dot-bracketform

atZD

OCK

score

ZRANKscore

Rank

PSAG

11TT

TTTC

TGTT

GCC

CGGAAC

GTC

GTG

GCC

CTTT

minus56

(((((())))))

1734

minus8144

1PS

AG12

TTTT

TGTG

GTG

TTTA

TTGTT

TACT

GTC

CCTT

Tminus02

(((())))

4927

minus6091

PSAG

13TT

TTTC

TGGTG

TTTA

TTCC

ATCA

AAT

ATCT

TTminus31

((((((()))))))

4003

minus6033

PSAG

14TT

TTTA

ATAT

CAAC

TTGGTT

TACT

GTC

CCTT

Tminus02

(())

4393

minus7008

PSAG

15TT

TTTC

TGGAAT

GAT

TTCC

CGGTT

GTC

TCTT

Tminus33

(((((())))))

3878

minus724

7PS

AG16

TTTT

TACT

TTGTT

TATT

GTT

TACT

GTC

CCTT

T0

NA

NA

PSAG

17TT

TTTC

TGGTC

CGGGTA

CGTT

TTTT

GGCC

TTT

minus48

((((((()))))))

3973

minus7149

8PS

AG18

TTTT

TCCG

CAGTT

TATT

GTT

TACT

GTC

CCTT

Tminus35

((((()))))

4005

minus6755

PSAG

19TT

TTTG

TGTT

GCC

CGGAAC

GTC

GTA

TATC

TTT

minus08

(((((())))))

4816

minus6889

PSAG

110TT

TTTA

ATAT

CAAC

TTGCC

ATCA

AGGCC

CTTT

minus33

((()))

4501

minus6888

PSAG

111

TTTT

TGTG

TTGCC

ATTT

CCCG

GTT

GTC

TCTT

Tminus13

((()))

486

minus6473

PSAG

112

TTTT

TACT

TAAT

GCG

GAAC

GTC

GTG

GCC

CTTT

minus14

(((())))

4258

minus7689

5PS

AG113

TTTT

TGTG

TTGCC

CGGAAC

TTTT

TTGGCC

TTT

minus3

(((((())))))

3977

minus6701

PSAG

114TT

TTTC

CGCA

CCGGGTA

CGGTC

GTG

GCC

CTTT

minus108

(((((((())))))))

4345

minus7891

2PS

A G115

TTTT

TAAT

ATCA

ACTT

GCC

AGGTT

GTC

TCTT

Tminus27

(((((())))))

4436

minus6184

PSAG

116TT

TTTA

CTTA

ATGAT

TTCC

CTCA

AAT

ATCT

TT0

NA

NA

PSAG

117TT

TTTA

ATAT

CCGGGTA

CGTT

TTTT

GGCC

TTT

minus24

(((((())))))

4752

minus7457

6PS

AG118

TTTT

TCCG

CACA

ACTT

GCC

ATCA

AAT

ATCT

TTminus11

((()))

4613

minus769

4PS

AG119

TTTT

TACG

CACC

GGGTA

CGTT

TTTT

GGCC

TTT

minus24

(((((())))))

4041

minus7784

3PS

AG120

TTTT

TCCT

TAAT

GAT

TTCC

CGGTT

GTC

TCTT

T0

NA

NA



Name

Average andstandard








PSAG24PSAG28PSAG212

ΔPSap45

minus200

minus150

minus100

minus50

0

50

Freq

uenc

y ch

ange

s (H

z)

2 4 6 8 10 12 14 16 18 200Time (minute)






Table3Second

-generationsequ

encesa


Nam

eSequ

ence

(51015840-31015840)

Minim

umfre

eenergy(kcalm

ol)

Dot-bracketform

atZD

OCK

score

ZRANKscore

Rank

PSAG

21TT

TTTC

TGGAAT

GAT

TAAC

GTC

GTG

GCC

CTTT

minus33

((((((()))))))

4344

minus6868

PSAG

22TT

TTTA

CTTA

ATGCG

GTC

CCGGGGGGCT

CTTT

minus49

((((()))))

4577

minus6916

PSAG

23TT

TTTC

TGGAAT

GAT

TACG

TTTT

TTGGCC

TTT

minus1

(((((())))))

4161

minus5612

PSAG

24TT

TTTA

ATAT

CCGGGTT

CCCG

GTT

GTC

TCTT

Tminus67

((((((()))))))

1754

minus7091

3PS

AG25

TTTT

TCTG

GAAT

GAT

TTCC

CGGTT

TGGCC

TTT

minus43

(((())))((()))

4569

minus5504

PSAG

26TT

TTTA

CGCA

CCGGGTA

CGTT

TTTG

TCTC

TTT

minus26

(((((())))))

4463

minus6012

PSAG

27TT

TTTA

CTTA

ATGCG

TACG

TTTT

TTGGCC

TTT

0

NA

NA

PSAG

28TT

TTTA

ATAT

CCGGGGAAC

GTC

GTG

GCC

CTTT

minus41

((((()))))

4264

minus86

1PS

AG29

TTTT

TACT

TAAT

GCG

GAC

GTT

TTTT

GGCC

TTT

minus17

(((())))

4451

minus6583

PSAG

210

TTTT

TACG

CACC

GGGTA

ACGTC

GTG

GCC

CTTT

minus54

(((((())))))

4261

minus685

PSAG

211

TTTT

TACT

TAAT

GAAT

ACGTT

TTTT

GGCC

TTT

0

NA

NA

PSAG

212

TTTT

TACG

CACC

GGGTA

CGTT

TTTT

GGCC

TTT

minus24

(((((())))))

4041

minus7784

2



Name of aptamer


(Hz)






119860





05

1

15

2

25

3

Adso

rptio

n pe

r are

a (ng

cm

2)

ΔPSap45


00

2

4

6

8

10

12

14

Adso

rptio

n pe

r are

a (ng

cm

2)

3020 40 5010Time (min)

PSAG24PSAG28

PSAG212ΔPSap45



4 Conclusions






Acknowledgments


References






































of






Disease Markers



OncologyJournal of





PPAR Research



Journal of

ObesityJournal of

























Table1First-g

enerationsequ

encesa

ndresults

from

thea

nalysis

andcalculations

Nam

eSequ

ence

(51015840-31015840)

Minim

umfre

eenergy(kcalm

ol)

Dot-bracketform

atZD

OCK

score

ZRANKscore

Rank

PSAG

11TT

TTTC

TGTT

GCC

CGGAAC

GTC

GTG

GCC

CTTT

minus56

(((((())))))

1734

minus8144

1PS

AG12

TTTT

TGTG

GTG

TTTA

TTGTT

TACT

GTC

CCTT

Tminus02

(((())))

4927

minus6091

PSAG

13TT

TTTC

TGGTG

TTTA

TTCC

ATCA

AAT

ATCT

TTminus31

((((((()))))))

4003

minus6033

PSAG

14TT

TTTA

ATAT

CAAC

TTGGTT

TACT

GTC

CCTT

Tminus02

(())

4393

minus7008

PSAG

15TT

TTTC

TGGAAT

GAT

TTCC

CGGTT

GTC

TCTT

Tminus33

(((((())))))

3878

minus724

7PS

AG16

TTTT

TACT

TTGTT

TATT

GTT

TACT

GTC

CCTT

T0

NA

NA

PSAG

17TT

TTTC

TGGTC

CGGGTA

CGTT

TTTT

GGCC

TTT

minus48

((((((()))))))

3973

minus7149

8PS

AG18

TTTT

TCCG

CAGTT

TATT

GTT

TACT

GTC

CCTT

Tminus35

((((()))))

4005

minus6755

PSAG

19TT

TTTG

TGTT

GCC

CGGAAC

GTC

GTA

TATC

TTT

minus08

(((((())))))

4816

minus6889

PSAG

110TT

TTTA

ATAT

CAAC

TTGCC

ATCA

AGGCC

CTTT

minus33

((()))

4501

minus6888

PSAG

111

TTTT

TGTG

TTGCC

ATTT

CCCG

GTT

GTC

TCTT

Tminus13

((()))

486

minus6473

PSAG

112

TTTT

TACT

TAAT

GCG

GAAC

GTC

GTG

GCC

CTTT

minus14

(((())))

4258

minus7689

5PS

AG113

TTTT

TGTG

TTGCC

CGGAAC

TTTT

TTGGCC

TTT

minus3

(((((())))))

3977

minus6701

PSAG

114TT

TTTC

CGCA

CCGGGTA

CGGTC

GTG

GCC

CTTT

minus108

(((((((())))))))

4345

minus7891

2PS

A G115

TTTT

TAAT

ATCA

ACTT

GCC

AGGTT

GTC

TCTT

Tminus27

(((((())))))

4436

minus6184

PSAG

116TT

TTTA

CTTA

ATGAT

TTCC

CTCA

AAT

ATCT

TT0

NA

NA

PSAG

117TT

TTTA

ATAT

CCGGGTA

CGTT

TTTT

GGCC

TTT

minus24

(((((())))))

4752

minus7457

6PS

AG118

TTTT

TCCG

CACA

ACTT

GCC

ATCA

AAT

ATCT

TTminus11

((()))

4613

minus769

4PS

AG119

TTTT

TACG

CACC

GGGTA

CGTT

TTTT

GGCC

TTT

minus24

(((((())))))

4041

minus7784

3PS

AG120

TTTT

TCCT

TAAT

GAT

TTCC

CGGTT

GTC

TCTT

T0

NA

NA



Name

Average andstandard








PSAG24PSAG28PSAG212

ΔPSap45

minus200

minus150

minus100

minus50

0

50

Freq

uenc

y ch

ange

s (H

z)

2 4 6 8 10 12 14 16 18 200Time (minute)






Table3Second

-generationsequ

encesa


Nam

eSequ

ence

(51015840-31015840)

Minim

umfre

eenergy(kcalm

ol)

Dot-bracketform

atZD

OCK

score

ZRANKscore

Rank

PSAG

21TT

TTTC

TGGAAT

GAT

TAAC

GTC

GTG

GCC

CTTT

minus33

((((((()))))))

4344

minus6868

PSAG

22TT

TTTA

CTTA

ATGCG

GTC

CCGGGGGGCT

CTTT

minus49

((((()))))

4577

minus6916

PSAG

23TT

TTTC

TGGAAT

GAT

TACG

TTTT

TTGGCC

TTT

minus1

(((((())))))

4161

minus5612

PSAG

24TT

TTTA

ATAT

CCGGGTT

CCCG

GTT

GTC

TCTT

Tminus67

((((((()))))))

1754

minus7091

3PS

AG25

TTTT

TCTG

GAAT

GAT

TTCC

CGGTT

TGGCC

TTT

minus43

(((())))((()))

4569

minus5504

PSAG

26TT

TTTA

CGCA

CCGGGTA

CGTT

TTTG

TCTC

TTT

minus26

(((((())))))

4463

minus6012

PSAG

27TT

TTTA

CTTA

ATGCG

TACG

TTTT

TTGGCC

TTT

0

NA

NA

PSAG

28TT

TTTA

ATAT

CCGGGGAAC

GTC

GTG

GCC

CTTT

minus41

((((()))))

4264

minus86

1PS

AG29

TTTT

TACT

TAAT

GCG

GAC

GTT

TTTT

GGCC

TTT

minus17

(((())))

4451

minus6583

PSAG

210

TTTT

TACG

CACC

GGGTA

ACGTC

GTG

GCC

CTTT

minus54

(((((())))))

4261

minus685

PSAG

211

TTTT

TACT

TAAT

GAAT

ACGTT

TTTT

GGCC

TTT

0

NA

NA

PSAG

212

TTTT

TACG

CACC

GGGTA

CGTT

TTTT

GGCC

TTT

minus24

(((((())))))

4041

minus7784

2



Name of aptamer


(Hz)






119860





05

1

15

2

25

3

Adso

rptio

n pe

r are

a (ng

cm

2)

ΔPSap45


00

2

4

6

8

10

12

14

Adso

rptio

n pe

r are

a (ng

cm

2)

3020 40 5010Time (min)

PSAG24PSAG28

PSAG212ΔPSap45



4 Conclusions






Acknowledgments


References






































of






Disease Markers



OncologyJournal of





PPAR Research



Journal of

ObesityJournal of


















Name

Average andstandard








PSAG24PSAG28PSAG212

ΔPSap45

minus200

minus150

minus100

minus50

0

50

Freq

uenc

y ch

ange

s (H

z)

2 4 6 8 10 12 14 16 18 200Time (minute)






Table3Second

-generationsequ

encesa


Nam

eSequ

ence

(51015840-31015840)

Minim

umfre

eenergy(kcalm

ol)

Dot-bracketform

atZD

OCK

score

ZRANKscore

Rank

PSAG

21TT

TTTC

TGGAAT

GAT

TAAC

GTC

GTG

GCC

CTTT

minus33

((((((()))))))

4344

minus6868

PSAG

22TT

TTTA

CTTA

ATGCG

GTC

CCGGGGGGCT

CTTT

minus49

((((()))))

4577

minus6916

PSAG

23TT

TTTC

TGGAAT

GAT

TACG

TTTT

TTGGCC

TTT

minus1

(((((())))))

4161

minus5612

PSAG

24TT

TTTA

ATAT

CCGGGTT

CCCG

GTT

GTC

TCTT

Tminus67

((((((()))))))

1754

minus7091

3PS

AG25

TTTT

TCTG

GAAT

GAT

TTCC

CGGTT

TGGCC

TTT

minus43

(((())))((()))

4569

minus5504

PSAG

26TT

TTTA

CGCA

CCGGGTA

CGTT

TTTG

TCTC

TTT

minus26

(((((())))))

4463

minus6012

PSAG

27TT

TTTA

CTTA

ATGCG

TACG

TTTT

TTGGCC

TTT

0

NA

NA

PSAG

28TT

TTTA

ATAT

CCGGGGAAC

GTC

GTG

GCC

CTTT

minus41

((((()))))

4264

minus86

1PS

AG29

TTTT

TACT

TAAT

GCG

GAC

GTT

TTTT

GGCC

TTT

minus17

(((())))

4451

minus6583

PSAG

210

TTTT

TACG

CACC

GGGTA

ACGTC

GTG

GCC

CTTT

minus54

(((((())))))

4261

minus685

PSAG

211

TTTT

TACT

TAAT

GAAT

ACGTT

TTTT

GGCC

TTT

0

NA

NA

PSAG

212

TTTT

TACG

CACC

GGGTA

CGTT

TTTT

GGCC

TTT

minus24

(((((())))))

4041

minus7784

2



Name of aptamer


(Hz)






119860





05

1

15

2

25

3

Adso

rptio

n pe

r are

a (ng

cm

2)

ΔPSap45


00

2

4

6

8

10

12

14

Adso

rptio

n pe

r are

a (ng

cm

2)

3020 40 5010Time (min)

PSAG24PSAG28

PSAG212ΔPSap45



4 Conclusions






Acknowledgments


References






































of






Disease Markers



OncologyJournal of





PPAR Research



Journal of

ObesityJournal of

















Table3Second

-generationsequ

encesa


Nam

eSequ

ence

(51015840-31015840)

Minim

umfre

eenergy(kcalm

ol)

Dot-bracketform

atZD

OCK

score

ZRANKscore

Rank

PSAG

21TT

TTTC

TGGAAT

GAT

TAAC

GTC

GTG

GCC

CTTT

minus33

((((((()))))))

4344

minus6868

PSAG

22TT

TTTA

CTTA

ATGCG

GTC

CCGGGGGGCT

CTTT

minus49

((((()))))

4577

minus6916

PSAG

23TT

TTTC

TGGAAT

GAT

TACG

TTTT

TTGGCC

TTT

minus1

(((((())))))

4161

minus5612

PSAG

24TT

TTTA

ATAT

CCGGGTT

CCCG

GTT

GTC

TCTT

Tminus67

((((((()))))))

1754

minus7091

3PS

AG25

TTTT

TCTG

GAAT

GAT

TTCC

CGGTT

TGGCC

TTT

minus43

(((())))((()))

4569

minus5504

PSAG

26TT

TTTA

CGCA

CCGGGTA

CGTT

TTTG

TCTC

TTT

minus26

(((((())))))

4463

minus6012

PSAG

27TT

TTTA

CTTA

ATGCG

TACG

TTTT

TTGGCC

TTT

0

NA

NA

PSAG

28TT

TTTA

ATAT

CCGGGGAAC

GTC

GTG

GCC

CTTT

minus41

((((()))))

4264

minus86

1PS

AG29

TTTT

TACT

TAAT

GCG

GAC

GTT

TTTT

GGCC

TTT

minus17

(((())))

4451

minus6583

PSAG

210

TTTT

TACG

CACC

GGGTA

ACGTC

GTG

GCC

CTTT

minus54

(((((())))))

4261

minus685

PSAG

211

TTTT

TACT

TAAT

GAAT

ACGTT

TTTT

GGCC

TTT

0

NA

NA

PSAG

212

TTTT

TACG

CACC

GGGTA

CGTT

TTTT

GGCC

TTT

minus24

(((((())))))

4041

minus7784

2



Name of aptamer


(Hz)






119860





05

1

15

2

25

3

Adso

rptio

n pe

r are

a (ng

cm

2)

ΔPSap45


00

2

4

6

8

10

12

14

Adso

rptio

n pe

r are

a (ng

cm

2)

3020 40 5010Time (min)

PSAG24PSAG28

PSAG212ΔPSap45



4 Conclusions






Acknowledgments


References






































of






Disease Markers



OncologyJournal of





PPAR Research



Journal of

ObesityJournal of


















Name of aptamer


(Hz)






119860





05

1

15

2

25

3

Adso

rptio

n pe

r are

a (ng

cm

2)

ΔPSap45


00

2

4

6

8

10

12

14

Adso

rptio

n pe

r are

a (ng

cm

2)

3020 40 5010Time (min)

PSAG24PSAG28

PSAG212ΔPSap45



4 Conclusions






Acknowledgments


References






































of






Disease Markers



OncologyJournal of





PPAR Research



Journal of

ObesityJournal of




















Acknowledgments


References






































of






Disease Markers



OncologyJournal of





PPAR Research



Journal of

ObesityJournal of




























of






Disease Markers



OncologyJournal of





PPAR Research



Journal of

ObesityJournal of
















The Combination of Computational and Biosensing ...downloads.hindawi.com › journals › bmri ›...

Documents

Transcript of The Combination of Computational and Biosensing ...downloads.hindawi.com › journals › bmri ›...