Oligonucleotide probes for genosensor chips

Information

  • Patent Application
  • 20040111221
  • Publication Number
    20040111221
  • Date Filed
    April 09, 2003
    21 years ago
  • Date Published
    June 10, 2004
    20 years ago
Abstract
Software for designing optimized sets of oligonucleotide probes for use in genosensors (oligonucleotide microarrays) is disclosed. The selection of probe sequences is based on multiple criteria including thermal stability of the probe-target pairs, similarity degree of the probes with respect to other DNA sequences, and evaluation of the secondary structure of target molecules. The programs were written in the programming language Borland Delphi by means of Object-Oriented Programming (OOP) techniques. The Genosensor Probe Design computer program disclosed herein facilitates the design of optimized arrays of probes which accurately represents the characteristics of the nucleic acid molecule under study, such as its identity or its differences in sequence or abundance with respect to other molecules.
Description


COMPUTER PROGRAM LISTING APPENDIX

[0002] Computer program listings are submitted on compact disc in compliance with 37 C.F.R. §1.96, and are incorporated by reference herein. A total of 2 compact discs (including duplicates) are submitted herein. The files on each compact disc are listed below:
1FolderFilesSize (KB)Date CreatedScript.doc484/9/03AdjustLenAbout.dfm42/3/00About.pas42/3/00Adjust.cfg42/3/00Adjust.dof42/3/00Adjust.dpr42/3/00Adjust.res48/9/99Adjust.tmp49/13/01Adjust.txt42/5/00Analysis.pas82/5/00Ddmodel.dat48/26/98MainForm.dfm42/5/00MainForm.pas82/5/00AlnClustAlnClust.dpr45/2/01AlnClust.res47/20/99Calculate.pas127/23/99Methods.dfm49/13/99Methods.pas47/23/99ParDlg.dfm47/23/99ParDlg.pas47/23/99BuiltDBBuild.cfg42/15/02Build.dof42/15/02Build.dpr42/15/02Build.res43/16/99BuiltDB.dfm43/31/02BuiltDB.pas83/31/02ListBx.dfm43/31/02ListBx.pas43/31/02DataCriteria.dat411/21/02Criteriaval.dat411/21/02Ddmodel.dat48/26/98DDOtraLucia.dat42/16/99DDSantaLucia.dat49/9/98DDunified.dat48/3/99Doktycz.dat46/30/98InitialVal.dat48/3/99Nndata.dat48/3/99Prueba.dat42/16/99Similar.dat42/22/03SSDoktycz.dat46/30/98VH.dat42/27/03Dialog boxesAboutBx.dfm282/27/03AboutBx.pas48/14/99Aboutold.dfm48/3/99Aboutold.pas48/14/99AddOligoDlg.dfm42/25/03AddOligoDlg.pas42/25/03AddProbe.dfm42/25/03AddProbe.pas82/25/03BegFrm.dfm48/3/99BegFrm.pas48/3/99CirProDlg.dfm42/12/03CirProDlg.pas42/12/03CritDlg.dfm42/12/03CritDlg.pas42/24/03DefCrit.dfm411/22/02DefCrit.pas42/12/03Destiny.dfm42/25/03Destiny.pas42/25/03DoubDlg.dfm48/3/99DoubDlg.pas208/3/99FieldsDlg.dfm42/9/03FieldsDlg.pas42/9/03GotoDl.dfm48/3/99GotoDl.pas48/3/99Import.dfm42/25/03Import.pas82/27/030InfoDNA.dfm88/3/99InfoDNA.pas168/3/99okcancl1.dfm48/3/99okcancl1.pas48/3/99okcancl2.dfm48/3/99okcancl2.pas48/3/99OlgDlg.dfm82/25/03OlgDlg.pas162/27/030Params.dfm42/25/03Params.pas82/25/03Paramsold.dfm411/23/02Paramsold.pas411/23/02PCRdlg.dfm48/3/99PCRdlg.pas42/6/03Projdlg.dfm42/27/03Projdlg.pas82/27/03SearchB.dfm48/3/99SearchB.pas48/3/99SelApp.dfm42/22/03SelApp.pas42/22/03SimilDlg.dfm42/21/03SimilDlg.pas82/21/03SimResDlg.dfm43/31/02SimResDlg.pas43/31/02SingDlg.dfm48/3/99SingDlg.pas88/3/99TablesDlg.dfm42/9/03TablesDlg.pas42/7/03VHCONF.dfm42/27/03VHCONF.pas42/27/03GenBankAbout.dfm208/13/99About.pas48/13/99Analysis.pas82/17/02Converter.cfg42/15/02Converter.dof42/15/02Converter.dpr42/15/02Converter.res48/12/99GBexample.txt42/16/02Genbank.bak48/11/99Genbank.cfg48/12/99Genbank.dof48/12/99Genbank.dpr88/12/99Genbank.pas48/11/99Genbank.zip364/7/03Main.dfm42/18/02Main.pas42/17/02Working.dfm42/15/02Working.pas42/15/02HelpGenosensor.cnt48/3/99Genosensor.gid1212/6/002Genosensor.hlp1288/3/99Genosensor.hpj48/3/99Genosensor.rtf88/3/99Genosensor.sh3123/16/00Genosensor.sn388/27/99GPDHelp.cnt48/16/99GPDHelp.gid1212/6/02GPDHelp.hhc48/14/99GPDHelp.hhk48/14/99GPDHelp.hhp48/14/99GPDHelp.hlp888/1699GPDHelp.hpj48/16/99GPDHelp.log48/16/99GPDHelp.rtf88/16/99GPDHelp.sh3128/14/99SH000000.shg688/14/99SH000001.shg528/14/99VHhelp.cnt48/18/99VHhelp.gid123/11/02VHhelp.hlp1608/18/99VHhelp.hpj48/18/99VHhelp.log48/18/99VHhelp.rtf328/18/99VHhelp.sh31328/16/99LibrariesCombin.pas84/3/02Combinar.dpr49/29/01DNAclass.pas202/27/03DNAoligo.pas242/27/03Formats.pas122/20/03Graph2D.pas129/18/01Indicar.dfm411/17/02Indicar.pas411/17/02OlgClass.old368/3/99OlgClass.pas4411/21/02OlgClass2.old408/3/99OOPlist.pas88/3/99RichFmt.pas123/10/02Rtfcolor.pas48/3/99Tools.pas84/2/02VH.pas42/17/02VHclass.pas82/17/02MapSeqAnsiStr.pas169/30/98Codific.dof410/9/98Codific.dpr410/9/98Codific.res49/28/98Codigo.pas810/2/98Combina2.pas89/30/98MapSeq.cfg43/29/02MapSeq.dof43/29/02MapSeq.dpr83/29/02MapSeq.res49/29/98MiniStr.pas1611/30/98Prueba.seq209/30/98Test.seq411/30/98TestDNA.seq49/29/98Tools.pas49/29/98ProbesHom16s.aln206/21/00Analysis.˜pa49/5/99Analysis.pas44/5/01Calculate.˜df49/5/99Calculate.˜pa49/5/99Calculte.dfm45/1/01Calculate.pas84/5/01DDmodel.dat48/26/98OligosHepC.˜tx49/5/99OligosHepC2.˜tx49/6/99Options.dfm44/5/01Options.pas44/5/01ProbesHom.˜dp47/27/99ProbesHom.cfg44/5/01ProbesHom.dof44/5/01ProbesHom.dpr44/5/01ProbesHom.res47/23/99Results.tmp44/5/01SectionD.dpr45/19/99ProjectDBDNAWin.dfm42/21/03DBDNAWin.pas82/23/03EdDNA.dfm88/14/02EdDNA.pas162/21/03Genosensor.dpr42/27/03Genosensor.ini42/27/03Homol.pas88/3/99Main.dfm683/1/03Main.pas443/1/03OligoDB.pas88/3/99OligoFrm.dfm43/1/03OligoFrm.pas83/1/03Pjrclass.pas48/29/02Prjclass.pas202/27/03Simil.pas122/22/03SqModule.dfm42/25/03SqModule.pas42/25/03Tables.pas122/25/03VH_form.dfm42/27/03VH_form.pas122/27/03VHClass.pas82/27/03StabilityComplete.txt85/13/01DoubleMis.dat45/13/01DoubleMis.txt45/12/01ExtractN.cfg45/13/01InitialVal.dat48/13/00Main.dfm45/20/01Main.pas86/28/01NNdata.dat48/12/00NNdata2.dat85/13/01OlgClass.pas449/29/02Oligodes.rtf410/16/02RichFmt.pas86/27/01Rtfcolor.pas46/11/01Stability.˜dpr46/11/01Stability.cfg410/16/02Stability.dof410/16/02Stability.dpr410/16/02Stability.dsk44/6/03Stability.res45/11/01TotalNN.txt45/12/01Watch.txt45/13/01Structure WindowsCombin.pas88/5/99ConfigDlg.dfm46/25/01ConfigDlg.pas46/25/01DNAClass.pas206/20/01DNAoligo.pas246/20/01DrawDNA.pas86/26/01Estructure.cfg46/26/01Estructure.dpr46/26/01Estructure.res46/25/01Formats.pas129/29/99InitialVal.dat48/13/00Main.dfm366/26/01Main.pas86/26/01NNdata.dat48/12/00Olgclass.pas405/12/01OOPlist.pas86/20/01PAEI.DNA46/17/01Results.tmp46/26/01Results.txt46/21/01SecEstructure.dfm46/25/01SecEstructure.pas46/25/01Secondary.pas86/20/01Secundaria.cfg46/21/01Secundaria.dpr46/25/01Secundary.pas46/20/01Tools.pas84/3/00Watch.txt46/26/01TandemAnalysis.dfm47/30/99Analysis.pas84/3/00DDmodel.dat48/26/98DDOtraLucia.dat42/16/99DDSantaLucia.dat49/9/98Gabriel.txt44/3/00Homol.pas48/4/99Main.dfm44/3/00Main.pas84/3/00Minimal.txt47/30/99Params.dfm48/3/99Params.pas48/4/99Prt1old.cap44/2/01Prt1old.DNA43/23/99Prt1old.stk84/2/01Results.tmp44/6/01Results2.txt44/3/00SC18S.txt48/3/99SC25S.txt88/3/99Tandem.cfg44/3/00Tandem.dof44/3/00Tandem.dpr48/4/99Tandem.res47/30/99Working.dfm43/11/99Working.pas43/11/99VirtualHybAboutbx.dfm1168/3/99Aboutbx.pas48/3/99Analysis.pas43/31/02Avel.dna48/3/99Bpul.dna48/3/99Bsp2.dna48/3/99Config.dfm43/31/02Config.pas43/31/02DrawDNA.pas48/3/99InitialVal.dat48/3/99Main.dfm163/10/02Main.pas243/31/02Mismatch.dat48/3/99NNdata2.dat88/3/99okcancl2.dfm46/17/98okcancl2.pas46/17/98OlgClass.pas449/3/01OOPlist.pas88/3/99PosStbl.dat48/3/99Probes.txt47/26/99Registry.dpr48/3/99Test.pas48/12/99VirtualHyb.cfg43/31/02VirtualHyb.dof43/31/02VirtualHyb.dpr43/31/02VirtualHyb.res48/15/99



BACKGROUND OF THE INVENTION

[0003] 1. Field of the Invention


[0004] The present invention relates generally to the fields of molecular biology and nucleic acid analysis. More specifically, the present invention provides a novel computer program for the design of optimized sets of oligonucleotide probes for microarrays.


[0005] 2. Description of the Related Art


[0006] Genosensors, also called oligonucleotide microarrays or “DNA chips,” are miniature devices containing arrays of oligonucleotide probes tethered to a surface (Beattie, 1997a). By hybridizing target nucleic acid molecules to these arrays and analyzing the resultant hybridization patterns, comparative analysis of sequences can be conducted (Beattie, 1997b), such as detection of specific mutations, identification of microorganisms (Beattie, 1997a), profiling of gene expression (Duggan et al., 1999), and verification of sequencing data (Hacia, 1999). For any given DNA or RNA sequence a large number of potential probes could be derived; however, only a small subset is needed to manifest the desired characteristics of the analyte nucleic acid molecules. In order to design probes for successful use in genosensors it is necessary to minimize the probability of unspecific (mismatched) hybridization between the probe and any nucleic acid sequence other than the intended target site (Doktycz and Beattie, 1997). A computer program is needed to design probes that are useful in microarray analysis.


[0007] A computer program called Genosensor Probe Designer (GPD) is disclosed herein, which can be used for selecting the most suitable probes for a genosensor chip based upon several factors that could affect the hybridization process. These factors include thermal stability, secondary structure, and alternative binding sites within the nucleic acid analyte.


[0008] It is well known that thermal stability of duplex nucleic acids depends on nucleotide sequence, chain length and nucleic acid concentration, as well as the identity and concentration of counterions. It is possible to find optimal hybridization conditions for specific binding of any given probe with its target molecule, but when the hybridization reaction is carried out with numerous probes and target molecules (as with genosensor chips), a loss in specificity can occur. The loss occurs particularly if the thermal stabilities of arrayed probes paired with their target sequences vary widely, or if the complexity of the analyte nucleic acid is sufficiently high to present alternative, mismatch-containing hybrids. Thus, the hybridization of multiple probes with a nucleic acid analyte can produce signals that are partially or completely due to imperfectly matched hybrids (Doktycz and Beattie, 1997). This kind of ambiguous hybridization signal depends on the sequence and the identity of the non-paired bases. A complete understanding of the thermal stability of hybrids formed between probes and nucleic acid molecules requires information about the energetic contributions for all the possible interactions that can take part in the hybridization process (Doktycz and Beattie, 1997).


[0009] Furthermore, the target DNA or RNA molecules are capable of forming stable secondary structures that can make some target sequences inaccessible to hybridization with the complementary oligonucleotide probes. Moreover, large targets are also likely inhibited sterically from approaching the surface of the array (Southern et al., 1999). In order to avoid these problems, several approaches can be followed. If a reasonable prediction of the secondary structure of the target could be made, probes could be selected from regions that are not tied up in secondary structure. Effects of secondary structure could be reduced by fragmenting the nucleic acid preferably to a size close to that of the oligonucleotides on the array (Southern et al., 1999). Also, strategies of annealing with auxiliary oligonucleotides (tandem hybridization) have been proposed to eliminate interfering secondary or higher-order structures or to cover up unwanted (redundant) hybridization sites within the target DNA (Maldonado-Rodriguez et al., 1999a, 1999b; Maldonado-Rodriguez and Beattie, 2001). Finally, when genosensor chips are used to reveal differences between closely related nucleic acid sequences, the probes must be selected to specifically identify a particular sequence. In this case probes must be selected from regions with sufficient sequence variability to minimize nonspecific hybridization with related molecules. On the other hand, when probes are required for identifying a group of similar sequences, probes must be selected from conserved regions.


[0010] Several works dealing with nucleic acid sequence analysis and oligonucleotide probe design have been published previously (Bushnell et al., 1999; Galper et al., 1993; Shütz and von Ahsen, 1999; Vahrson et al., 1996; Li and Stormo, 2001; Pozhitkov and Tautz, 2002). One interesting work is Vahrson's library, called SCL-a, which is a C++ Object-Oriented library similar in some respects to that disclosed herein. Vahrson's library is specialized in the management of dynamical memory for manipulating long DNA sequences, whereas the library disclosed herein is specialized in the calculation of thermodynamic stability and the search for potential hybridization sites.


[0011] Object-Oriented support included in the Object Pascal library of Delphi is similar to that provided in C++. Classes are similar between Delphi and C++ programming languages; however, Object-Pascal language has a clearer syntax than that used by C++, and Delphi code can be easily translated to C++ if required with minimal complexity. Moreover, Delphi compiled native programs can run faster than those produced using C++ compilers.


[0012] A spreadsheet software program for thermodynamic melting point prediction of oligonucleotide hybridization based in the NN model has recently been developed (Shütz and von Ahsen, 1999). However, this program does not predict the specific hybridization patterns that could be expected with a given set of probes, and does not design sets of probes for the variety of Genosensor applications that are described herein.


[0013] Also a program for selection of optimal DNA probes for gene expression arrays has been published recently (Li and Stormo, 2001). Although this program uses criteria for selection of probes similar to those implemented in the software that is described in the present work, it is intended to select relatively long probes (more than 20 bases long) which are less convenient for single mutation discrimination. Also, an algorithm and program for selecting specific probes for species identification with microarrays has been published recently (Pozhitkov and Tautz, 2002). This algorithm considers position of mismatches which influences the selection; however, information is lacking about the experimental performance of the probes selected with this program.


[0014] Thus, in order to identify conserved or variable regions the complete alignment of the sequences under study must be conducted prior to selection of the most appropriate target regions for the subsequently designed probes. Consequently, in the design of optimized sets of oligonucleotide probes for nucleic acid analysis on genosensor arrays, careful consideration of numerous factors must be done, including the characteristics of the nucleic acid analyte, the type of analysis being performed, thermal stability of probe-target duplexes, secondary structure within the target sequence, and alternative probe binding sites within the target nucleic acid. The Genosensor Probe Designer software disclosed herein takes all of these factors into consideration.



SUMMARY OF THE INVENTION

[0015] A program for selecting optimized sets of oligonucleotide probes for use in genosensor chips (also known as oligonucleotide microarrays) is developed and disclosed herein. Selection of probes is based on thermodynamic stability, similarity degree properties and suitable parameter values for virtual hybridization. A more comprehensive analysis yielding even more confident predictions of hybridization patterns can be provided by inclusion of additional thermodynamic parameters such as the influence of dangling ends and contributions of non-standard nucleic acid interactions.


[0016] The Genosensor Probe Designer software disclosed herein was written in the programming language Borland Delphi (Borland International) Version 5.0. It can run on Windows 95, 98, Millennium, NT and XP operating systems. The computer program was developed based on the Object-Oriented Programming (OOP) methodology, and classes for representing the user interface, molecules, criteria and interactions are provided. The object-oriented feature of the Genosensor Probe Designer program facilitates the development of improved versions, accommodates among other information upgrades of the thermodynamic models.


[0017] The software developed herein was successfully tested in the laboratory, wherein a set of primers and probes directed against the 5′ end region of the 16S rRNA sequences of several strains of Pseudomonas and other bacterial species were designed. Results from these experiments show specific hybridization patterns for each sequence tested, and criteria for selecting probes and predicting hybridization patterns were derived.


[0018] Thus, the present invention is drawn to a software system for designing and selecting oligonucleotide probes for use in DNA microarrays as well as methods of using such software to select optimal sets of oligonucleotide probes for use in DNA microarrays.







BRIEF DESCRIPTION OF THE DRAWINGS

[0019]
FIG. 1 shows a screen image of the Genosensor Probe Designer program running in the Windows Millenium environment. The program uses an advanced Multiple Document Interface (MDI) for its user friendliness. DNA and oligonucleotide sequences are conveniently manipulated and stored in Databases that are controlled by the Borland Database Engine (BDE).


[0020]
FIG. 2 shows the heredity relations between developed classes for the Genosensor Probe Designer Program. In the Delphi programming language of the Genosensor Probe Designer Program, every class is a descendant from the TObject class. One group of classes represents the molecules involved in the hybridization process. TDNAbstracclass is an abstract class, which defines all the common properties of the DNA molecules. TDNAclass and TOligoclass are abstract classes also, which define specific characteristics for long and short DNA sequences, respectively. Abstract classes are not directly used for object representation but they are used as intermediate steps for the design of other objects. TOligoclass has three descendent classes, which comprise different implementations for calculating Tm and other thermodynamic properties. A second group of classes represents the interactions between molecules, and these classes are derived from a combination of TDNAclass and TOligoclass classes. TDNAoligomismatch, like its relative, Toligomismatch, uses the doublet model to calculate the Tm and free energy. TvirtualHybridization, the most specialized class of this second group, is used to predict hybridization patterns. TCriteriaclass and TSymilarityclass are other types of classes, directly derived from the TObject class, and comprise methods and thermodynamic criteria to select probes and to evaluate similarity between target and probes, respectively. The character “T” is normally used for class referencing, while the instance of the class (the “object”) is referenced using the same name as the class, omitting the “T.”


[0021]
FIG. 3 summarizes the applications and functions of the Genosensor Probe Designer program. The program can use single or multiple target sequences, oligonucleotide sequences, or oligonucleotide databases.


[0022]
FIG. 4 illustrates the importance of the stability of the probes at the experimental hybridization temperature. In the example two probes have similar sequence and the target site for probe A has a high degree of similarity with that of probe B. Melting temperature of probe A is designated as TmA. Probe B can hybridize ambiguously at lower temperature at the same site as probe A with a melting temperature TmB. If the experiment is carried out at a temperature close to TmA only probe A will give a hybridization signal. However, if the experimental temperature is close to TmB both probes can give hybridization signals.


[0023]
FIG. 5 illustrates the importance of the degree of similarity between oligonucleotides and target sequences for the design of probes for Genosensors. Probes used to identify a specific DNA sequence among a collection of similar targets must be directed against the sites found to have the highest variation between the sequences. However, probes required to identify a group of similar sequences must be directed against the most conserved regions of the sequences.


[0024]
FIG. 6 illustrates the importance of the accessibility of hybridization sites in target DNA molecules. In this example, two probes are directed against different sites in the target DNA molecule. However, the secondary structure of the target can prevent probe B from accessing its target.


[0025]
FIG. 7 shows an example of mapping all the probes of a given size in a particular sequence.


[0026]
FIG. 8 shows thermodynamic properties of oligonucleotides and energetic contribution of several substructures.


[0027]
FIG. 9 shows the first step for the virtual hybridization algorithm. Potential hybridization sites are found by evaluating the number of complementary bases and the number of contiguously paired bases between sites in the target sequence and the probe (placing both in an anti-parallel form). Sites where the number of complementary bases or the size of a block with contiguously paired bases is greater than MinBasesCom and MinBlockSize respectively are stored as potential hybridization sites.


[0028]
FIG. 10 shows the algorithm used to estimate secondary structure between probes and potential hybridization sites. M and N are used to design each sequence. Pairing between bases is checked by placing both sequences in an anti-parallel form. m and n are used to design positions in sequences M and N respectively. ParM and ParN designate nearest-neighbor doublets for sequences M and N. Match1 is the result for the comparison of the first base pair in a nearest-neighbor doublet, which can be true (match) or false (mismatch). Similarly Match2 is the result for the comparison of the second base pair. Matching patterns and their positions are used as flags to identify a particular substructure by means of a decision table, which is used to assign the correct free energy value associated to it. Here TableNN is the table of nearest-neighbor interactions values, and Len is the length of the duplex.


[0029]
FIG. 11 shows the effect of penultimate mismatches in oligonucleotide duplexes. The first structure where bases in both ends are paired is more unstable than the second where the bases in the ends are untied. (From Allawi and SantaLucia, 1997).


[0030]
FIG. 12 shows the sequence alignments and positions of 9mer probes along the 16S rRNA gene sequences.


[0031]
FIG. 13 displays the virtual hybridization results for all 90 9mer probes in regions “A” for the seven bacteria tested.


[0032]
FIG. 14 displays the virtual hybridization results for all 90 9mer probes in regions “B” for the seven bacteria tested.


[0033]
FIG. 15 displays experimental and virtual hybridization patterns for Region A of Pseudomonas aeruginosa using oligonucleotide probes designed by the Genosensor Probe Designer program. The probes were spotted in duplicate. Color codes: Gray, weak experimental signal; Black, strong experimental signal; Green, probes with perfect match; Red, sites with ΔG range between −12 and −10 Kcal/mol; Blue, sites with ΔG range between −10 and −8 Kcal/mol; Yellow, sites with ΔG range between −8 and −6 Kcal/mol.


[0034]
FIG. 16 displays the results of Kolmogorov-Smirnov normality test for distribution of calculated free energy values for all sites of potential hybridization and sites exhibiting strong experimental hybridization signals, sites of weak signal, and combined sites of strong+weak hybridization signals.


[0035]
FIG. 17 shows secondary structure of region A of 16s ribosomal RNA gene from P. aerugionosa predicted with RNA draw v1.1. All GU pairs were forbidden. Regions correspond to important targeted sites for probes have bases marked with squares.


[0036]
FIG. 18 shows the screen image of a program that calculates the thermal stability for perfect and mismatched hybrids. Free energy calculations use the same algorithm implemented in the VH algorithm. This function can be conveniently used for fast visualization of energetic contributions in the hybridization between any given pair of strands.


[0037]
FIG. 19 shows Virtual Hybridization program running in the Windows Millenium Desktop. The program has an advanced user interface which is menu driven. Here is shown a part of an analysis of potential hybridization between several probes and a Bacillus target sequence. Potential Hybridization sites are shown with their associated free energy values. Different colors are used to show sites with free energy lower than specific cut-off values. In this example, sites with free energy values in red correspond to the most stable values. Sites with medium stability are shown in green while blue sites correspond to the most unstable sites.


[0038]
FIG. 20 shows the Help system of the Genosensor Probe Designer Program. The GPD program has a complex help system which describes each of the functions implemented in the program.


[0039]
FIG. 21 shows the Sequence editor in the Genosensor Probe Designer Program. DNA sequences can be edited directly from the GPD program in order to obtain the complementary sequence, cut, or copy sections of the sequence, search PCR products, etc.


[0040]
FIG. 22 shows the Databases in the Genosensor Probe Designer Program. The GPD program has database capabilities to organize DNA sequence data and selected probes.


[0041]
FIG. 23 shows an example of mapping of probes in the genome of an organism using the Genosensor Probe Designer Program. The program uses a very fast algorithm in order to find all the probes of a given size which are present in the genome sequence of a specific organism or database. The program can then produce a graph listing the number of probes that appear at different frequencies in that sequence or database. The program can show all the possible probes found and their frequencies.


[0042]
FIG. 24 displays search of probes from aligned sequences in the Genosensor Probe Designer Program. Sequences can be aligned with the help of any alignment tool such as Clustal W. The program scans the alignment column by column in order to find the most variable (or conserved) regions which will then be used to design specific (or universal) probes.


[0043]
FIG. 25 shows the selection of probes for tandem hybridization (also named contiguous stacking hybridization) in the Genosensor Probe Designer Program. The GPD program has capabilities to select stacking and capture probes required for the tandem hybridization approach.







DETAILED DESCRIPTION OF THE INVENTION

[0044] In the Object-Oriented Programming (OOP) terminology, classes are abstract data structures which encapsulate data and operations or methods used to manipulate them. Objects are instances of the classes, used for representing any item such as an oligonucleotide molecule or a window of the user interface (Vahrson et al., 1996). Classes or objects used in the program can be divided into interface and application classes or objects, depending on whether they are involved in the design of the user interface, or in the process to select the probes.


[0045] Interface objects were developed using the Visual Component Library (VCL) provided within Delphi and the Orpheus 3.0 library (TurboPower Software Company). Both libraries have an extensive collection of Native Visual Components that permit the design of highly complex user interfaces. The user interface developed for the Genosensor Probe Designer program is based on the Multiple Document Interface (MDI) standard. FIG. 1 shows a general view of the Genosensor Probe Designer user interface.


[0046] The set of application objects is the core of the Genosensor Probe Designer program. They were developed using the Object Pascal Language provided for Delphi and these objects include all the classes for representing molecules, selection criteria, interactions between molecules and similarity degree. These objects are involved in the selection of probes. Since these objects are independent of the interface objects and the VCL, they can be used to develop other applications. Moreover, application objects can be divided into four groups. The first group of objects represents the DNA molecules involved in the hybridization process (“targets” and “probes”). The second set of objects represent interactions or processes involving the DNA molecules (e.g. the hybridization reaction). The third group includes objects which use methods based on thermodynamic criteria for rejecting or accepting probes and the fourth group contains objects with methods for evaluating similarity between targets and probes.


[0047]
FIG. 2 shows the hierarchical relationships between the objects that were developed. TDNAclass objects are computer representations of DNA molecules longer than 50 bp. These objects include characteristic DNA properties such as length, molecular weight and sequence composition, as well as methods for reading DNA sequences from data files. Sequences must be in FASTA or Genbank formats. Toligoclass objects are virtual representations of sequences with length up to 50 bp. Separate DNA objects for long sequences and oligonucleotides were developed because models to predict thermal stability for long and short nucleic acid molecules are different. For example, nucleotide sequence effects cannot be ignored in sequences shorter than 50 bp, while empirical models based on base composition alone can describe thermodynamic values such as Tm for long DNA sequences with reasonable precision. Nearest-neighbor (NN) models are better considered for describing properties of oligonucleotides (Cantor and Schimmel, 1980). Several data sets for calculating thermodynamic properties based in the nearest-neighbor model have been proposed and two main formats have been used for representing NN parameters, termed the singlet and the doublet formats (Owczarzy et al., 1997).


[0048] Singlet format considers contributions from H-bonding and stacking interactions separately and uses linearly independent equations for calculating thermodynamic properties. In the doublet format the entire NN interaction is considered in a single parameter and thermodynamic values for each of the NN base pairs are used. TOligoclass objects have descendent classes specialized for using NN parameters of both data formats. As can be seen in FIG. 2, a TOligoempirical class is also included to calculate oligonucleotide properties by using the empirical models. SantaLucia has shown a unified set of parameters (SantaLucia, 1998) which adequately describes thermodynamic properties for oligonucleotides using the doublet format and the complete set of NN parameters for energy contributions due to single mismatches and dangling ends has been published recently (Allawi and SantaLucia, 1997, 1998a, 1998b, 1998c; Peyret et al., 1999; Bommarito et al., 2000). These data sets have been implemented in the TOligomismatch class.


[0049] TDNAoligoclasses are combined classes derived from TDNAclass and TOligoclass, and were developed for describing interactions between oligonucleotides and target DNA molecules. These objects include methods for binary mapping of the DNA molecules, which are required for identifying potential sequences of probes generated from the target molecule and for evaluating similarity between oligonucleotides and target DNA sequences.


[0050] TCriteriaclass includes several criteria for rejecting or accepting oligonucleotide probes based upon their thermodynamic properties. The selection criteria are based upon user-specified [A+T] composition, [G+C] composition, melting temperature (Tm) range, enthalpy, entropy, free energy, internal repeated sequences, sequence symmetry of probes, and frequency of occurrence. The user can conveniently disable any of the evaluated properties in order to customize the selection procedure.


[0051] TSimilarityclass class includes methods for evaluating similarity between probes and targets. By using this class, specific probes can be selected by establishing convenient cut-off values for similarity between probes and sites within the target DNA sequences. To accomplish this, a search of similarity can be performed against sequences contained within Internal or External DNA databases. The naming conventions for databases are based upon considerations of whether the DNA sequences are defined within the internal DNA database that is generated when the program starts (see below). External databases can contain several DNA sequences within a single text file and they must be in FASTA format, while internal databases are composed of several text files, each containing just a single DNA sequence in FASTA or Genbank formats. Methods defined in TSimilarityclass can show very detailed information about all the sites within the DNA sequences where probes have significantly high similarity degree. This information can be useful to predict potential hybridization of the probes as well as to know whether the probes are specific for one particular DNA sequence or a set of sequences.


[0052]
FIG. 3 shows in detail the applications performed by the Genosensor Probe Designer program. The program can be used for automatic or manual selection of probes targeted to particular DNA sequences. For any application, input data can include target DNA sequences (single or multiple), DNA sequence databases, oligonucleotide sequences or oligonucleotide databases. Target DNA sequences must be in Genbank or FASTA formats (only FASTA format for files containing multiple sequences). For the aligned sequences analysis, sequences must be aligned first with any alignment tool such as Clustal X (Thompson et al., 1997). As soon as the initial data are read, a genosensor application can be specified.


[0053] From the sequences specified in the initial data section, an internal DNA database table is created including the entire target DNA sequences from which probes will be selected. Then, a primary table of oligonucleotides is generated containing all possible probes of a defined length that can be derived from the DNA target molecules. When the user-defined selection criteria within Tcriteriaclass are applied, a secondary table is created containing just the probes with appropriate composition and thermodynamic properties. An additional search can then be performed with the probes contained in the secondary table in order to make a further selection of specific probes. For this analysis, the program uses the methods defined in an instance of TSimilarityclass. Similarity degree is evaluated between probes and selected target DNA molecules contained in the internal DNA database (defined before) or in external DNA databases which are text files containing sequences saved in FASTA format.


[0054] An additional analysis can be made with the table of specific probes in order to obtain oligonucleotides of variable length but similar stability. This analysis adds or deletes nucleotides from one or both ends of the probe based on the flanking DNA sequence in the target site. As the length of the probes changes an additional similarity analysis can be performed in order to delete sequences that no longer have the desired specificity.


[0055] Using the final table of specific oligonucleotides, an analysis of the possible hybridization of the probes against the DNA sequences in the internal database can be performed. This analysis shows target sites for perfect hybridization, as well as sites that could be potential sites for ambiguous hybridization due to complementarity degree and stability values. This approach is called Virtual Hybridization (VH).


[0056] In order to perform the VH analysis the following rules are defined in the program:


[0057] (i) For a given probe length (L) a minimal value (minbasescom) for the number of complementary bases between probes and potential hybridization sites is defined such that 2≦minbasescom≦L.


[0058] (ii) For a given probe length (L) a minimal value (minblocksize) for the length of contiguously paired bases within potential hybridization sites is specified such that 2'<minblocksize≦L.


[0059] (iii) The complementarity degree between probes and potential hybridization sites is evaluated along the target sequences (representing all nucleic acid sequences present in the analyte). Potential Hybridization Sites are selected as sites where the number of complementary bases between the probe and the evaluated site is equal to or greater than minbasescom or sites where a block of contiguously paired bases is equal to or greater than minblocksize.


[0060] (iv) Potential hybridization sites with 100% complementarity with the probe are shown as sites with only Watson-Crick base-pairing or sites of perfect hybridization.


[0061] (v) Potential hybridization sites with less than 100% complementarity with the probe are shown as sites with some non-Watson-Crick paired bases or sites of ambiguous hybridization.


[0062] (vi) An additional analysis of the thermal stability of probes paired with their respective potential hybridization sites is performed using the NN model (including mismatch data). Free energy values calculated by this model are useful to estimate the hybridization probability. In general, at more negative ΔG° values the probability of stable pairing is higher.


[0063] (vii) ΔG° cutoff values are defined (Gcutoff). Potential hybridization sites with ΔG° values equal to or greater than Gcutoff are shown as sites of high hybridization probability.


[0064] Virtual Hybridization (VH) is an important new capability that can yield a predicted hybridization pattern from any given combination of oligonucleotide probes and target sequences. By incorporating VH analysis during the selection of probes, ambiguous hybridization can be minimized or avoided entirely, thus optimizing the effectiveness of genosensor chips. A more comprehensive implementation of the VH strategy can be provided by expansion of the predicted vs. experimental hybridization data set, utilizing a larger collection of sequence-verified targets. Currently, the program parameters are adjusted in order to show only the sites of high hybridization probability such that a virtual hybridization pattern is obtained.


[0065] The Genosensor Probe Designer software disclosed herein can use any or all of the program applications described below for selecting the probes.


[0066] In the simple selection program application a set of probes for a particular DNA sequence will be generated. Selection criteria are based only on thermodynamic and composition properties.


[0067] The mapping application is a very fast procedure for selecting all the possible probes of any size (up to 15-mers) that can be obtained from any sequence or sequences, usually a complete genomic sequence.


[0068] The non-aligned sequences application is used in order to select probes that are specific for particular sequences (which could be similar or not). However, the search procedure is time-consuming and could be slow for the analysis of long sequences.


[0069] In the aligned sequences application, DNA sequences aligned with any alignment tool such as Clustal-X are read. Then, the alignment is scanned column by column. If oligonucleotide probes are required for hybridization with the conserved regions of the alignment, contiguous columns with identical bases in all the rows are selected until the probe size is reached. Such probes are sent to the primary table of oligonucleotides. Alternatively, probes could be required for revealing the differences between the DNA sequences included in the alignment. In the latter case, columns containing several base variations are selected and probes are derived from those columns in order to obtain a set that can be used to represent the maximum differences between the aligned sequences.


[0070] In the tandem hybridization application (Maldonado-Rodriguez and Beattie, 2001; Maldonado-Rodriguez et al., 1999a,b), short probes (typically 5-mer to 9-mer) called the “capture” probes can be selected as in the single selection application. After the secondary table of capture probes is built, longer “stacking” probes are selected contiguous to the 5′ or 3′ end of the capture probes. Stacking probes are typically 20-30 nucleotides or longer without restriction regarding their sequences. Alternatively, capture and stacking probes can be selected manually from specific sites within the target.


[0071] The following examples are given for the purpose of illustrating various embodiments of the invention and are not meant to limit the present invention in any fashion. One skilled in the art will appreciate readily that the present invention is well adapted to carry out the objects and obtain the ends and advantages mentioned, as well as those objects, ends and advantages inherent herein. Changes therein and other uses which are encompassed within the spirit of the invention as defined by the scope of the claims will occur to those skilled in the art.



EXAMPLE 1

[0072] Program Details and Algorithms


[0073] There are several aspects of oligonucleotide interactions which are taken into consideration in the algorithms of the Genosensor Probe Designer program. Some of these design considerations are illustrated and discussed in FIGS. 4-6.


[0074] Frequently it is important to know the occurrence of particular n-mers in a particular DNA sequence in order to be able to design probes for DNA finger printing or SBH. The occurrence of particular n-mers can be determined by searching all possible combinations of n-mers or mapping of the DNA target sequence.


[0075] The number of possible combinations of n-mers is equal to 4n. Therefore, there are 16 combinations for dimers. As an example, the occurrence of each dimer on the sequence: 5′-TATAGTAGAAACCACAA-3′ (SEQ ID NO.1) can be searched by comparing target/pattern (using Brute search, Boyer-Moore, or Quicksearch) with the pattern sliding along the target. Searching each dimer will obtain the following results: AA=3, TA=3, AC=2, TC=0, AT=1, TT=0, AG=2, TG=0,CA=2, GA=1, CC=1, GC=0, CT=0, GT=1, CG=0, GG=0.


[0076] For mapping all the probes of a given size in a particular sequence, each probe sequence can be represented with a number in base 4: Probe=N1N2. Nm with Ni={A, C, G, T} with m<16. If A=0, C=1, G=2, and T=3 then, in the case of dimers:
2Seq base4base10Seq base4base10AA = 000GA = 208AC = 011GC = 219AG = 022GG = 2210AT = 033GT = 2311CA = 104TA = 3012CC = 115TC = 3113CG = 126TG = 3214CG = 137TT = 3315


[0077] And the occurrence of each dimer on the sequence: 5′-TATAGTAGAAACCACAA-3′ (SEQ ID NO.1) can be mapped as shown in FIG. 7.


[0078] For ambiguous sequences, the characters R, Y, M, K, S, W, H, B, V, D, N and X are used to represent ambiguous bases in DNA sequences: R={A,C}, Y={C,T}, M={A,C}, K={G,T}, S={G,C}, W={A,T}, H={A,C,T}, B={G,T,C}, V={G,C,A}, D={G,A,T}, N or X={A,C,T,G}. If target=‘ACNAC’ and pattern=‘ACTAC’, then the probability of perfect match between target and pattern can be defined as: P=1/# combinations=1/2r 3h 4n, where r=number of Rs, Ys, Ms, Ks, Ss and Ws in target; h=number of Hs, Bs, Vs, and Ds; n=number of Ns or Xs.


[0079] The searching and mapping methods described above were used for finding specific n-mer in the E. coli genome sequence (4,639,221 bp) using a Pentium 260 Mhz processor. The time of search for Quicksearch was 4.2 hours, whereas the time of search for mapping was 0.52 hour.


[0080] Also, these methods were used for finding specific n-mers in the Small Subunit rRNA Database (22,324 sequences, average size=1093 bp). This database has ambiguous bases in several sequences (9271 sequences). Using a Pentium 260 Mhz processor, the time of search for Quicksearch was 1.5 days, whereas the time of search for mapping was 0.45 hour.


[0081] Virtual Hybridization Algorithm


[0082] Accurate prediction of the thermal stability for hybridization reactions has become a very important issue in several molecular biology techniques. Reliable prediction of the free energy associated with the formation of the DNA duplex is necessary in order to select the optimal experimental conditions for specific detection and/or identification of target molecules. This aspect is dramatically important in DNA microarrays where several hybridization reactions are carried out simultaneously. If the thermal stability of the different probes used in the array is too different, a loss in the specificity of the hybridization can result, whereby “hybridization signals” are due to both perfect match between the probes and target but also some imperfectly matched hybrids, which is known as ambiguous hybridization.


[0083] Thermodynamic models have been developed in order to predict the free energy values and other thermodynamic parameters associated with the thermal stability of DNA duplexes. The most accurate model for oligonucleotides is based on stacking interactions between neighboring bases in the duplex, and it is named the nearest-neighbor interactions model.


[0084] The nearest-neighbor model can predict the thermal stability for both perfectly and imperfectly matched duplexes, and can thus be used for predicting whether oligonucleotides can hybridize under a given set of experimental conditions. We call such a theoretical approach Virtual Hybridization (VH). If accurate parameters for the nearest-neighbor model are used, then VH can simulate the results of hybridization experiments under defined experimental conditions with a high degree of accuracy. Parameters have been derived and disclosed herein to predict patterns of hybridization of PCR products derived from 16S rRNA genes of several bacterial species with arrays of 9-mer probes. In this section a more detailed description of the algorithm used for virtual hybridization is given.


[0085] The Nearest-Neighbor Model


[0086] The nearest-neighbor model considers thermal stability as sequence-dependent and in terms of base pair doublets (nearest-neighbors). In duplex DNA there are ten possible combinations of such nearest-neighbors:
3AA AG AC GA GG TG CG GC AT TA|| || || || || || || || || || TT TC TG CT CC AC GC CC TA AT


[0087] The NN model assumes that the stability of a given base pair depends on the identity and the orientation of the neighboring base pairs. In the most simple format of the NN model, the free energies for oligonucleotide duplexes result from the sum of its nearest-neighbor interactions. It has been proven that the NN model also serves to estimate the contribution of other secondary structure components. Currently, the NN model has been extended to predict the contribution of several components of the secondary structure of DNA oligonucleotides such as: Stacking regions, Internal single mismatches, Dangling ends.


[0088] The Algorithm.


[0089] Virtual Hybridization analysis has been conceived as a two-stage approach. The first stage aims to find the most probable sites within a target sequence where a particular probe can hybridize. This search includes two parameters called minbasescom and minblocksize. The values of these parameters are dependent on the probe size and, for a given probe length: 2≦minblocksize≦minbasescom≦L, where L is the length of the probe, minblocksize is the minimal accepted length of contiguously paired bases between a probe and a potential hybridization site, and minbasescom is the minimal accepted number of complementary bases between a probe and a potential hybridization site. A Virtual Hybridization search begins by looking for sites in the target molecule where the number of contiguously paired bases or the number of complementary bases with the probe are equal to or greater than minblocksize or minbasescom, respectively. Sites found by means of this search are stored and considered as potential hybridization sites.


[0090] The second stage of the virtual hybridization approach is the calculation of the free energy associated with the formation of duplexes between the probe and the potential hybridization sites.


[0091] Bases between a particular site and the probe are compared in order to decide if bases can be paired (match) or not (mismatch). Matches and mismatches between bases are grouped by pairs. This approach does not presently consider gaps on that comparison. Different components of the secondary structure of the duplex can be found in this way. Table 1 summarizes several secondary structure components (substructures) and abbreviations used to represent them.
4TABLE 1Secondary Structure Components For DNA DuplexesStructureAbbreviation5′ Terminal mismatchEndMis53′ Terminal mismatchEndMis3Double internal mismatchDoubInt3′ Double external mismatchDoubEx35′ Double external mismatchDoubEx5Perfect paired dinucleotidePerfectPenultimate single mismatch close to 3′ endPenSing3Penultimate single mismatch close to 5′ endPenSing5Single internal mismatchSingMis


[0092] Matching patterns for each of the nearest-neighbors of the duplex can be combined with the positions of such patterns in the duplex in order to identify a secondary structure component as is illustrated in FIG. 9. Then, free energy values associated with each substructure can be calculated by means of a decision table as illustrated in FIG. 10. The free energy value for the duplex will be the sum of all the individual free energy values associated for each substructure.


[0093] The nearest-neighbor data set has not been completed to represent the contributions of all the secondary structure components of the DNA. For this reason some values need to be estimated from separate considerations.


[0094] Thermodynamic values for terminal mismatches are not published yet. However, some studies indicate that terminal mismatches have stabilizing values (negative free energy values). Free energy values for terminal mismatches in this prediction have been assumed to be zero, however, a more confident prediction of the thermal stability will be possible when precise values for these contributions are available.


[0095] Free energy values for double and multiple contiguous internal mismatches are also not published. Estimated values have been derived for these interactions assuming that internal mismatches have destabilizing values (positive free energies).


[0096] Penultimate mismatches deserve special attention. Free energy values for structures with penultimate mismatches are greater (more unstable) than values calculated by considering that the paired bases in the end adjacent to the penultimate mismatch are untied as shown in FIG. 11. Theses calculations are in agreement with experimental observations. When free energy values for such duplexes are calculated with the algorithm described before, ends are not untied. An additional analysis is performed when penultimate mismatches are found in order to untie the paired bases in the ends adjacent to the mismatches.


[0097] Once the free energy value for the duplex is calculated, it is compared with specified cut-off values. The cut-off values can be assigned for particular experimental conditions. If free energy of the duplex is less than or equal to the cut-off value (i.e., if it has a greater negative free energy value than the cut-off value), then site is marked as a high probability of hybridization site or probable signal.



EXAMPLE 2

[0098] Oligonucleotide Arrays Targeted to Microbial 16S rRNA Genes


[0099] Successful performance of algorithms implemented in the Genosensor Probe Designer program has been demonstrated through the discrimination of Pseudomonas aeruginosa and closely related bacterial strains via hybridization fingerprinting using genosensor chips. The seven test strains and the accession numbers for the 16S rRNA gene sequences are listed in Table 2.
5TABLE 2Accession NumbersLengthAccessionDefinition(nt)AB002661Bacillus sp. gene for 16S rRNA, complete sequence.1553AB020208Bacillus pumilus DNA for 16S ribosomal RNA.1471AF064460Pseudomonas veronii 16S ribosomal RNA gene,1521complete sequence.D84006Pseudomonas alcaligenes 16S rRNA gene,1530complete sequence.Z76662P. fluorescens 16S rRNA gene.1507AF017749Stenotrophomonas maltophilia 16S ribosomal RNA1507gene, complete sequence.D84020Pseudomonas putida 16S rRNA gene, complete1527sequence.X06684Pseudomonas aeruginosa DNA for 16S rRNA.1537AB001447Pseudomonas syringae pv. Myricae DNA for 16S1538ribosomal RNA.


[0100] DNA sequences encoding 16S rRNA genes of the seven microbial test strains (including closely related pseudomonads) were aligned with the aid of the Clustal X program (Thompson et al., 1997). From the analysis of the produced alignment, three conserved regions which delimit two highly variable regions (named “A” and “B”) near the 5′ end of the gene were found.


[0101] The conserved regions were analyzed to find convenient PCR primers for the amplification of the regions A and B based on established rules for PCR primer selection (Rychlik, 1993). PCR forward primers were derivatized with a 5′-Biotin group to facilitate single-stranded target DNA purification. PCR reverse primers were derivatized with a 5′-fluorescent label (CY3 or CY5) to enable visualization of the hybridization patterns.


[0102] The primer sequences (5′->3′) were as follows:
6Region A forward primer5′-CTCCTAQCGGGAGGCAGCAG-3′(SEQ ID NO.2)Region A reverse primer5′-GTATTACCGCGGCTGCTGG-3′(SEQ ID NO.3)Region B forward primer5′-CCAGCAGCCGCGGTAATAC-3′(SEQ ID NO.4)Region B reverse primer5′-GGCGTGGACTACCAGGGTATC-3′(SEQ ID NO.5)


[0103] For PCR amplification of the combined region AB, the 5′-biotin forward primer for Region A was used together with the 5′-fluorescent reverse primer for Region B. The PCR product from region “A” is located between nucleotides 356 and 554, yielding a 197 bp fragment for P. aeruginosa and a 198 bp fragment for the six other species. The PCR product from region “B” is positioned between nucleotides 533 and 830, forming a 293 bp fragment for P. aeruginosa and a 294 bp fragment for the remaining species. The distal pair of primers yields a larger “AB” PCR product encompassing regions “A” and “B” that has a length of 471 bp for P. aeruginosa and 473 bp for the other species.


[0104] The sequences of the seven PCR products for each region were realigned using Clustal X and the resultant alignment was analyzed using the aligned sequences module of the Genosensor Probe Designer program to select 9-mer probes. In this module an alignment of DNA sequences is read and then scanned column by column. Columns containing several base variations are selected and probes are derived from those columns in order to obtain a set that can be used to represent the maximum differences between the aligned sequences. In order to verify that probe sequences were present only within the intended target sites, all selected probes were tested for 100% degree of similarity against all sequences in the alignment.


[0105] As a result of this process a set of 87 nonamer probes (28 for region “A” and 59 for region “B”) directed against the most highly variable regions of the alignment was selected (Table 3). FIG. 12 displays the sequence alignments and positions of 9mer probes along the 16S rRNA gene sequences. These 9mer probes were not designed for similar duplex stability (Tm value) and were not analyzed for similarity degree below 100% within the target sequences, thus it was anticipated that some of them could yield ambiguous hybridization with other sites within the 16S rRNA gene targets.


[0106] PCR products derived from each of the seven bacterial test species were hybridized to genosensor chips bearing the selected probes for each region. Results of these experiments showed unique hybridization patterns for each species. The experimental details of the PCR, oligonucleotide array formation, and hybridization experiments have been reported separately (Reyes-Lopez et al., 2003).
7TABLE 3Listing of Oligonucleotide Probe Sequences. CalculatedTm And ΔG0Values. And Species To Which EachProbe Is Specific (Single Perfect Match)ΔG0SequenceTm(Kcal/Targeted SpeciesProbe(5′->3′)(° C.)mol)(perfect sequence match with test strains)1AAGTCTGAC32.03−10.49B. pumilus2AGTCTGACG37.60−11.66B. pumilus3GGAGCAACG42.53−12.72B. pumilus4GTGAGTGAT32.18−10.54B. pumilus5TGAGTGATG32.22−10.55B. pumilus6GGAGAAGCC39.68−12.08- - - 7GCAGTTACC35.95−11.27P. syringae, P. veronii8CAGTTACCT30.93−10.31P. syringae, P. veronii9CGAGAGTAA31.38−10.35B. pumilus10GAGAGTAAC28.35−9.62B. pumilus11GTAACTGCT33.02−10.71B. pumilus12ATACCTTGC32.77−10.71P. aeruginosa, P. alcaligenes13AACTGCTCG39.85−12.16B. pumilus14TACCTTGCT34.42−11.11P. aeruginosa, P. alcaligenes15TACGTTAGT28.85−9.93P. putida, P. fluorescens16CCTTGCTGT39.02−11.98P. aeruginosa, P. alcaligenes17CGTTAGTGT33.56−10.80P. putida, P. fluorescens18CGTGATTGT35.05−11.13P. syringae, P. veronii19GTTAGTGTT27.90−9.63P. putida, P. fluorescens20GCTGTITGA35.24−11.16P. aeruginosa, P. alcaligenes21GCTGTTTTG34.18−10.86P. syringae22AGTGTITCG34.90−11.08P. fluorescens23AGTGTTTTC29.30−9.91P. putida24GTTGGGATG35.50−11.20- - - 25GTGTYFCGA35.04−11.10P. fluorescens26CCTTGACGG40.76−12.32B. pumilus27TTGACGTTA31.17−10.38P. syringae, P. aeruginosa, P. alcaligenes, P. veronii28GTACCTAAC28.08−9.60B. pumilus29GTGGTTCAG35.70−11.20P. aeruginosa, P. alcaligenes30TGGTTCAGC39.14−12.00P. aeruginosa, P. alcaligenes31GGTTCAGCA39.14−12.00P. aeruginosa, P. alcaligenes32GTTCAGCAA35.24−11.16P. aeruginosa, P. alcaligenes33TTCAGCAAG34.43−11.00P. aeruginosa34AGCAAGCTT37.74−11.77P. alcaligenes35AGCAAGTFG35.09−11.14P. aeruginosa36GCAAGCTTG39.04−11.94P. alcaligenes37GCAAGTTGG37.95−11.70P. aeruginosa38CAAGCTTGA34.43−11.00P. aeruginosa39AAGCTTGAT31.06−10.43- - -40AAGTCCGTT36.33−11.47B. pumilus41AAGTCTGAT28.56−9.93P. aeruginosa42AGCTTGATG33.72−10.88- - -43AGTCCGTTG38.76−11.92B. pumilus44AGTCTGATG31.34−10.38P. aeruginosa45GCTTGATGT34.54−11.04- - -46GTCCGTTGT39.53−12.08B. pumilus47GTCTGATGT32.18−10.54P. aeruginosa48CTTGATGTG31.25−10.25- - -49TCCGTTGTG39.56−12.09B. pumilus50TCTGATGTG32.22−10.55- - -51GTGGATACT30.35−10.21P. aeruginosa52AAGCTACTG32.18−10.55P. alcaligenes53AACTACTGA27.89−9.77P. fluorescens, P. veronii54AACTGACTG32.75−10.64P. aeruginosa55AGCTACTGA33.22−10.85P. alcaligenes56ACTACTGAG29.78−10.05P. fluorescens, P. veronii57ACTGACTGA33.81−10.94P. syringae, P. putida58ACTGGCAAG39.02−11.98P. aeruginosa59GCTACTGAG34.03−10.85P. aeruginosa, P. alcaligenes60CTACTGAGC34.03−10.85P. fluorescens, P. veronji61CTGACTGAC34.57−10.94P. syringae, P. putida62CTGGCAAGC42.84−12.78P. aeruginosa, P. alcaligenes63CTACTGAGC34.03−10.85P. fluorescens, P. veronji64CTGACTGAC34.57−10.94- - -65CTGGATGAC34.36−10.94P. fluorescens, P. veronii66GACTGACTA29.96−10.07P. syringae, P. putida67GGCAAGCTA38.60−11.91P. fluorescens, P. veronii68ACTGACTAG29.78−10.05P. syringae, P. putida69GCAAGCTAG36.30−11.35P. aeruginosa, P. alcaligenes70TGAGCTAGA32.52−10.71P. syringae, P. putida71CAAGCTAGA31.51−10.41- - -72CTAGAATGT25.55−9.21P. aeruginosa73GTAGAGGTG32.85−10.61P. aeruginosa74TAGAGGTGG34.23−11.01P. aeruginosa75AGAGGTGGT38.26−11.87P. aeruginosa76GAGGTGGTA35.07−11.17P. aeruginosa77AGGTGGTAG34.92−11.15P. aeruginosa78GGTGGTAGA35.07−11.17P. aeruginosa79TGGTAGAAT27.55−9.77P. fluorescens80GTGGTGAAG35.70−11.20B. pumilus81CTCTCTGGT35.15−11.17B. pumilus82TCTCTGGTC35.30−11.19B. pumilus83CTCTGGTCT35.15−11.17- - -84TGGACCAAC38.01−11.76P.aeruginosa85ACTGTACTG31.36−10.36- - -86ACCAACATT31.39−10.50P.aeruginosa87CTGTACTGA30.68−10.22- - -88CATTGACAC32.10−10.41- - -Probes 1-28 were designed for region A and probes 29-88 were designed for region B. Probes designed for species other than the seven test strains are listed as “- - -” in column 5.


[0107] To help explain discrepancies between anticipated and observed hybridization results obtained with the 16S rRNA gene system, the virtual hybridization module of the Genosensor Probe Designer (GPD) software (described further in Example 3) was applied to the data. The virtual hybridization program calculates the Gibbs free energy (ΔG°) of mismatched or matched probes when aligned to target DNA (Table 4). The GPD software runs each probe along the full target DNA sequence (including all test strains) and calculates binding energy as ΔG° for each position. The selection can be adjusted to search for pairings with high, medium or low probability to produce stable hybridization under a given hybridization condition. FIGS. 13 and 14 display the virtual hybridization results for all 90 9mer probes in regions “A” and “B” for the seven bacteria tested. The results were divided into eight stability groups of increasingly negative ΔG° values (+2 to 0, 0 to −2, −2 to −4, −4 to −6, −6 to −8, −8 to −10, −10 to −12 and −12 to −14 kcal/mol). The green bars in the back row represent the total number of predicted hybridization signals for each stability group, including both perfectly matched and mismatched hybrids. The bars in the remaining rows represent total (blue), strong (red) and weak (yellow) experimentally observed signals. On top of each bar is indicated the total number of signals for all bacterial species that were predicted (green bars) or experimentally observed (other colors) in each stability group. In theory, a Gaussian distribution of predicted ΔG° values (green bars) should be produced for a random collection of probes of a given length. However, in the design of probes, oligonucleotides having single mismatches at any site within the target sequences were generally excluded, and this restriction accounts for the reduced number of hybrids in the stability group with a ΔG° of −8 to −10 kcal/mol. For experimentally detected hybrids it can be seen that the number of signals detected generally decreased as the stability of the hybrid decreased. When the experimentally detected signals were divided into strong and weak signals (red and yellow bars, respectively), it could be seen that for region “A” (FIG. 13), in the group with the highest ΔG° value (−10 to −12 kcal/mol) all signals were strong (100%). In the next stability group (−8 to −10 kcal/mol) most hybridization signals (75%) were strong and the remaining (25%) were weak. In the next group with even lower stability (−6 to −8 kcal/mol) all the signals (100%) were weak. A similar result was seen with the −4 to −6 kcal/mol group, where 85% of the signals were weak and 15% were strong. The same analysis was conducted for region “B” with very similar results (FIG. 14). In general, these results suggest that, with few exceptions, the virtual hybridization analysis explains the perfectly matched and mismatched hybridization signals obtained. Assuming that hybridization signals arise primarily from the most stable duplex structure found for each probe within the target (Table 4), it can be seen that the experimentally detected mismatched hybrids generally contained mismatches positioned at their ends, where destabilization is minimal (Doktycz et al., 1995), and in many of these cases the mismatches were of a type known to have low destabilizing effect (Maldonado-Rodriguez and Beattie, 2001; Peyret et al., 1999; Doktycz et al., 1995; Yang et al., 1997). The good correlation between the experimentally observed hybridization signal intensity and the predicted free energy of binding raises our confidence in the data interpretation. In the case of the relatively small number of pairings with high negative ΔG value that were not experimentally detected, these may be largely attributed to sequence differences between the samples and reference strains, or to the formation of secondary structure within the target (Matzura and Wennborg, 1996).


[0108] Hybridization signals that are observed but not predicted (i.e., involving mismatched hybrids) belong nevertheless to a specific target and consequently, can contribute to the identification of bacterial species via hybridization fingerprinting. As further detailed in Example 3, the hybridization data obtained in the 16S rRNA gene study provided the “working materials” for development of the Virtual Hybridization strategy which is useful in the interpretation of hybridization results, and can guide the selection of optimal sets of oligonucleotide probes for a given genosensor application.
8TABLE 4Comparison of Predicted Versus Experimentally ObservedHybridization In Region B of P. Aeruginosa.SequenceΔ G°PositionTargetedΔ G°No(5′->3′)Tm (° C.)[2]Kcal/m olM.S.S.site (3′->5′)M.S.S.28GTACCTAAC28.1−9.60257CCTCGTTTG−2.5429GTGGTTCAG35.7−11.2059CACCAAGTC−11.2030TGGTTCAGC39.1−12.0069ACCAAGTCG−12.0031GGTTCAGCA39.1−12.0070CCAAGTCGT−12.0032GTTCAGCAA35.2−11.1671CAAGTCGTT−11.1633TTCAGCAAG34.4−11.0072AAGTCGTTC−11.0034AGCAAGCTT37.7−11.7775TCGTTCGAA−11.7735AGCAAGTTG35.1−11.1475TCGTTGGAA−7.1536GCAAGCTTG39.0−11.9476CGTTCGAAC−11.9437GCAAGTTGG38.0−11.7027CGTTCGCAA−5.9738CAAGCTTGA34.4−11.0077GTTCGAACT−11.0039AAGCTTGAT31.1−10.4378TTCGAACTA−10.4340AAGTCCGTT36.3−11.47283ATCAGGTCC−5.8641AAGTCTGAT28.6−9.9378TTCGAACTA−4.6642AGCTTGATG33.7−10.8879TCGAACTAC−10.8843AGTCCGTTG38.8−11.92284TCAGGTGCG−5.8644AGTCTGATG31.3−10.3879TCGAACTAC−5.1145GCTTGATGT34.5−11.0480CGAACTACA−11.0446GTCCGTTGT39.5−12.08161AAGGACACA−4.3847GTCTGATGT32.2−10.5480CGAACTACA−7.1248CTTGATGTG31.3−10.2581GAACTACAC−10.2549TCCGTTGTG39.6−12.09162AGGACACAT−4.3850TCTGATGTG32.2−10.5581GAACTACAC−7.9751GTGGATACT30.4−10.21103CATCTATAT−4.1052AAGCTACTG32.2−10.55124TTCGATGAC−10.5553AACTACTGA27.9−9.77125TCGATGACT−7.3354AACTGACTG32.8−10.64230ATGACTGTG−6.9155AGCTACTGA33.2−10.85125TCGATGACT−10.8556ACTACTGAG29.8−10.05126CGATCACTC−6.9157ACTGACTGA33.8−10.94231TGACTGTGA−6.9158ACTGGCAAG39.0−11.98231TGACTGTGA−6.1259GCTACTGAG34.0−10.85126CGATGACTC−10.8560CTACTGAGC34.0−10.85127GATGACTCG−10.8561TGACTGAGC34.6−10.94220ACATGACTG−6.9162CTGGCAAGC42.8−12.7824CTACGTTCG−8.2163CTACTGAGC34.0−10.85122GATGACTCG−10.8564CTGACTGAC34.6−10.94228ACATGACTG−4.9165CTGGATGAC34.4−10.94233GACCTGACA−5.8766GACTGACTA30.0−10.07230ATGACTGAG−6.9167GGCAAGCTA38.6−13.9175CGTTCGAAC−9.4968ACTGACTAG29.8−10.05231ACTCGATCT−6.9169GCAAGCTAG36.3−11.3576CGTTCGAAC−9.4970TGAGCTAGA32.5−10.71131ACTCGATCT−10.7171CAAGCTAGA31.5−10.41131ACTCGATCT−7.9672CTAGAATGT25.6−9.21154CATCTTAAA−5.0473GTAGAGGTG32.9−10.61145CATCTCCAC−10.4174TAGAGGTGG34.2−12.03144ATCTCCACC−11.0175AGAGGTGGT38.3−11.97147TCTCCACCA−11.4776GAGGTGGTA35.1−11.17148CTCCACCAT−11.1777AGGTGGTAG34.9−11.15149TCCACCATC−11.1578GGTGGTAGA35.1−11.17150CCACCATCT−11.1779TGGTAGAAT27.6−9.77152ACCATCTTA−9.7780GTGGTGAAG35.7−11.20205CACCCCTTC−9.4181CTCTCTGGT35.2−11.27275ATGGGACCA−6.3182TCTCTGGTC35.3−11.10274TGTGACCAT−6.2183CTCTGGTCT35.2−11.17277GCGACCATC−6.3184TGGACCAAC38.0−11.76214GACTGGTCG−4.5385ACTGTACTG31.4−10.36226TGACATGAC−10.3686ACCAACATT31.4−10.30198TTGTGGTCA−3.3087CTGTACTGA30.7−10.22227GACATGACT−10.2288CATTGACAC32.1−10.41290ATGACTGTG−6.26


[0109] The first column shows the 9mer probe number as given in Table 3. The second column displays the 9mer probe sequence (5′->3′ direction). The third column lists the calculated Tm values for each 9mer probe, paired with its complementary target sequence. The fourth column lists the binding energy (AG® in Kcal/mol) calculated for each probe, paired with its complementary target sequence. The fifth column lists the first nucleotide position of the most stable site (MSS) for binding of each probe within region B of P. aeruginosa.


[0110] The sixth column displays the target sequence (3′->5′ direction) for each MSS within region B of P. aeruginosa, with paired positions written in black and mispaired positions indicated in orange. The last (seventh) column lists the calculated binding energy (ΔG° in Kcal/mol) for hybridization of each probe at its MSS within region B of P. aeruginosa.



EXAMPLE 3

[0111] Virtual Hybridization


[0112] Using virtual hybridization (VH) analysis rules as described earlier, the set of selected 16S rRNA gene probes and target sequences were analyzed to predict virtual hybridization patterns. Table 5 represents the predicted and experimentally observed hybridization signals obtained in all of these experiments, grouped according to predicted free energy of probe-target hybrids. FIG. 15 shows a comparison of predicted and experimental hybridization patterns in a representative experiment.
9TABLE 5Occurrence of Potential Hybridization Sites, Experimental Hybridization Signals, AndProbability of Hybridization As A Function of Predicted Free Energy Range.StrongWeakTotal# of sites# of sites withexperimentalexperimentalexperimentalProbability ofΔG rangepredictedperfect matchsignals (S)signals (W)signals (S + W)hybridization  0 to +2500000.00−2 to 0  700000.00−4 to −22300110.04−6 to −41560110110.07−8 to −61790041410.23−10 to −8 678627330.49−12 to −108482366420.50−14 to −12110000.00Total =522914385128


[0113] Predicted free energy values for potential hybridization sites were compared with experimentally obtained hybridization signals in order to select idealized parameters for reliable virtual hybridization. These comparisons revealed that predictions of the hybridization of 9mer probes could be made with the highest degree of accuracy using the set of parameters listed in Table 6. Moreover, it appears that the prediction of hybridization based on the evaluation of the number of contiguously paired bases and thermal stability of the bases involved in the hybridization could be sufficient for designing genosensor chips with the greatest capacity to manifest the desired characteristics of the analyzed sequences.
10TABLE 6Improved Parameters For Selecting Probes (9-Mer Length)By Virtual HybridizationParametervalueMinimal number of complementary bases (minbasescom) (nt)6Minimal size of block with contiguously paired bases5(minblocksize) (nt)First ΔG cut-off value (Gcutoff1) (Kcal/mol)−10Second ΔG cut-off value (Gcutoff2) (Kcal/mol)−8


[0114] Similar experiments can be readily performed to derive suitable parameters for Virtual Hybridization analysis of probes shorter and longer than 9mer. An important feature of Virtual Hybridization analysis provided within the Genosensor Probe Designer software is that the program can accommodate future expansion of knowledge concerning the thermodynamics of unusual nucleic acid interactions. For example, as additional information becomes available on the free energy contributions of mismatched base pairs and secondary or tertiary structures, the new thermodynamic data sets can be directly used in the existing Virtual Hybridization software to provide a still more comprehensive and reliable prediction of hybridization patterns.


[0115] From the analysis of the VH results it can be seen that about 50% of the potential hybridization sites (perfect or ambiguous) with free energy range between −12 and −10 Kcal/mol yielded detectable experimental signal (strong signals for the most part). A similar situation existed for potential sites with free energy range between −10 and −8 Kcal/mol, where 49% of the sites gave experimental signal, but this time the proportion of weak signals was higher than the strong ones. This situation changes considerably for the free energy range of −8 to −6 Kcal/mol, where only 23% of the sites gave experimental signal (all weak signals). For weaker free energy ranges (above −6 Kcal/mol) the relative number of experimental signals with respect to the number of potential hybridization sites was even lower.


[0116] A Kolmogorov-Smirnov normality test showed that the distributions of free energy values for the potential hybridization sites and for the sum of weak and strong signals, differ considerably from the normality. However, strong and weak signals assume a normal distribution. In order to demonstrate if calculated free energy values for “strong” and “weak” experimental signals were significantly lower (more stable) than the free energy distribution for all potential hybridizations sites, we perform both parametric and nonparametric tests for all cases.


[0117] Statistical tests showed that free energies for strong signals differ considerably from the free energies for all potential hybridization sites for both PCR products. Free energies for weak signals at the PCR products for region A are not significantly different to the free energies for all potential hybridization sites, but the difference is significantly for PCR products for region B and also when we combine the results for both regions. In conclusion, this analysis showed that strong hybridization signals were significantly obtained at more negative free energies (more stable) while weak signals are usually found at negative free energies too, but the discrimination power between them and the energies for all potential sites is considerably lower than that for strong signals. Moreover, it can be seen that both signals as a whole are significantly present at more negative free energies.


[0118] The results of statistical analyses for Regions A+B are shown in Table 7 and FIG. 16.
11TABLE 7Statistical Analyses of Virtual And Experimental Hybridization In RegionsA + B.Kolmogorov-Smirnov normality test for distributionof predicted free energiesSitesnMeanMedianStDv.p-valueAll M.S.S.517−7.002−6.5302.335<0.01Strong signals41−10.832−10.9400.771>0.15Weak signals87−7.589−7.5401.6640.031Strong + weak128−8.628−8.2202.090<0.01n = number of sites; StDv = Standard deviation; M.S.S. = moststable sites.Calculated p-values for Mann-Whitney and 2-Sample t-test forcomparison of predicted free energiesComparisonMan-Whitney2-Sample t-testAll M.S.S. vs strong signals0.00000.000All M.S.S. vs weak signals0.00110.005All M.S.S. vs strong + weak0.00000.000M.S.S. = most stable sites.


[0119] It can be seen from the above Virtual Hybridization results that some probes that were predicted to hybridize perfectly or ambiguously by the VH analysis at stable ΔG ranges failed to yield experimental hybridization signals. For example, for region A of P. aeruginosa probes 12 and 14 were predicted to yield hybridization signals, however these signals were not observed. Some possible explanations of this phenomenon include:


[0120] (i) UNCERTAINTY OF TARGET SEQUENCES. Genbank reference sequences used to design the probes (cited in Example 2) may be different from the actual PCR product sequences derived from the environmentally isolated test strains. Thus, differences between experimental and virtual hybridization patterns (including both unexpected and unobserved signals) could be due to differences between reference and actual target sequences.


[0121] (ii) SECONDARY STRUCTURE OF THE PRODUCT. Hybridization reactions were carried out at low temperature which favors the formation of secondary structure within the target molecule. A simplified evaluation of the secondary structure of the target is provided within the Genosensor Probe Designer program. Self-complementary sections of the target molecules longer than a given value are searched for and listed. If target sites of the probes overlap with sections with potential secondary structure, these probes could have lower probability of yielding hybridization signals. Results in this case are somewhat equivocal, but they suggest than the absence of some hybridization signals could be due to the secondary structure of the target molecule.


[0122] Secondary structure within the 16S rRNA gene sequences was further evaluated with the help of the program RNA draw v1.1. Since this program was designed to predict secondary structure of RNA sequences, the prediction will be just an approximation for DNA targets. In order to provide a closer prediction of the secondary structure of the DNA target, G-U pairs (common in RNA molecules) were forbidden (G-T pair are uncommon in DNA). For PCR product A of P. aerugionosa, an interesting region which includes bases 50 through 77 contains the targeted sites for probes 12, 14, 16, 20 and 27. As seen in FIG. 17, the complete sequence of probes 12, 14 and part of the sequence of probes 16 and 20 are positioned within a region capable of forming stable secondary structure. Probe 27, which yielded a hybridization signal, is positioned outside of this region. Although predicted secondary structures do not satisfactorily explain all the cases of missing hybridization signals, they may explain the absence of some predicted hybridization signals, especially since hybridization reactions were carried out at low temperatures.



EXAMPLE 4

[0123] Source Code of Genosensor Probe Designer Program and Associated Materials


[0124] The source code of the Genosensor Probe Designer Program, together with associated data and documents are provided in the computer program listings appendix.


[0125] The source code include: (1) the source code for the main Genosensor Probe Designer program; (2) the source code for objects and other tools used for representing DNA molecules, interactions and selection criteria; (3) the source code for dialog boxes that art part of the GPD program interface; (4) data files containing thermodynamic properties used by the GPD program; and (5) documents and Microsoft help projects for building the Help system of the GPD program.


[0126] The Genosensor Probe Designer Program is further described in the following Table 8, which includes a description of program units contained in the main GPD program.
12TABLE 8Description of Units for the Main GPD ProgramPascal UnitDelphi Form(*.pas)(*.DFM)TypeDescriptionAboutBxAboutBoxDialogShows a Dialog box with thename of the program, versionand credits.AddProbeAddProbeDlgDialogShows a dialog box where anoligonucletide sequence andproperties can be introducedand added to a database.Catch.pasMainUnit to collect importantinterfaceinformation about data fromthe Genosensor Software asmaximum and minimumthermodynamic values.CombineLibraryThis unit contains proceduresand functions for calculatingall the possible combinationsof sequences that can beobtained from a sequence withambiguous bases.CritProDlgCriteriaPropDlgDialogShows a dialog box where thevalues of the Criteriaparameters used for probeselection can be visualized andedited.CritDlgCriteriaDlgDialogDialog box which permits theuser to enable or disable thecriteria parameters used forprobe selection.CriteriaSelLibraryContains the source code forthe TCriteria class whichcontains all the informationrequired to select or reject apotential probe exceptsimilarity data. The TCriteriaobject has specializedproperties describingwhether a particular propertyis active and its minimal andmaximal values. Methods wereimplemented to verifywhether a particularoligonucleotide property isbetween the minimum andmaximum values.DBDNAWinDBDNAFormMainDisplays a Form in whichinterfaceinformation about the DNAinternal database is shown.DestinyProbeDestDlgDialogDisplays a dialog box in whichthe destination of a probe(primary or secondary table)can be specified.DNAclassLibraryContains the source code forthe TDnaclass class anddescendents for the objectrepresentating andmanipulating long DNAsequences.DNAoligoLibraryContains the source code forthe TDNAoligo class and itsdescendents. These arecombined classes obtainedfrom TDNAclass andToligoclass and are used tosimulate the hybridizationbetween long and short DNAsequences, and implementsalso several tools to compareboth kinds of sequencesEdDNAEditDNAMainShows a form in which a DNAinterfacesequence from the internalDNA database can be edited.The editor implementstraditional edit capabilitiessuch as Cut, Paste, Delete andSelect as well as severalsearch capabilities specializedfor DNA sequences (theypermit, for example, the use ofcharacters for representingambiguous bases). The editoralso contains some useful toolsfor obtaining thecomplementary sequence, orfor modifying names.FieldsDlgVisibleFieldsDlgDialogShows a dialog box where thevisible fields of theoligonucleotide database canbe selected. This selection doesnot affect the structure of thedatabase itself.FormatsLibraryThis unit contains proceduresand functions formanipulation of DNAsequence, especially forFormat conversions (Genbankto FASTA).GotoDlGotoDlgDialogShows a dialog box whichpermits the user to jump to adesired position of the DNAsequence that is edited in theEditDNA window.Graph2DLibraryContains the source code forthe TgraphicXY class which isused to create a XY coordinatesystem.ImportImportDlgDialogShows a dialog box in whichthe user specifies the datacontained in a text file for anoligonucleotide sequence thatwill be imported to the GPDprogram (Data can include thename of the sequence, thesequence, and concentrationsof probe and salt.MainMainFormMainThis is the main unit of theinterfaceGPD program. It administratesthe main form of theapplication and main menus.From this unit all Dialog boxesand Libraries are called andcontrolled throughout theentire design procedure.OlgClassLibraryImplements source code forthe TOligoclass and itsdescendents used to representand manipulateoligonucleotide sequences andtheir properties.OlgDlgOligoDlgDialogDisplays a Dialog box where anoligonucleotide sequence,derived from a specificposition within a DNAsequence placed in the EditorWindow, can be shown. ThisDialog is used to see thethermidynamic properties of asingle probe and also formanual selection of probes(probes can be sent to theprimary or secondary tablesfrom this dialog box).OligoFrmOligoDBFormMainShows information aboutinterfaceprobes contained in theprimary and secondary tables.It shows the final sequencesof probes resulting from thedesign procedures.OOPListLibraryObject-orientedimplementation of a singlylinked list. Methods areincluded for creating the list,inserting data, deletingelements, counting thenumber of elements anddisposing of the list. Code isbased on example 5.3 from S. Sengputaand C. P. Korobkin,C++ Object-Oriented DataStructures (1994), pp: 145-156,but the code impemented hereis in Delphi and avoidsrecursive calls.ParamsSimilParamsDlgDialogDisplays a dialog box wherethe parameter values forsimilarity searches can beshown and modified.PCRdlgPCRsearchDlgDialogDisplays a dialog box used forthe sequence editor where apair of short sequences can beintroduced in order to testwhether a subsequence existswithin the edited DNAsequence which is delimitedby such pair of sequences.This unit is useful for findingPCR products.PrjclassLibraryImplements the source code ofthe Projects created for theGPD program. A project is afile that tracks all tables,properties and criteria usedfor the design procedure ofprobes.ProjDlgProjectDlgDialogShows a dialog box where theproperties of the designproject can be specified ormodified.RichFmtLibraryThis unit has routines forwriting documents in the RichText Format (RTF). Routinesare based on the RTFSpecification Version 1.3(1994). All the RTF routinesare encapsulated in the classTRichroutines. When aninstance of TRichroutines iscreated an RTF file isproduced. The RTF file isclosed and saved into the diskwhen the instance isdestroyed.RtfcolorLibraryContains values of theconstants defined for RGBcolors used in Rich TextFormat Documents.SearchBSearchDlgDialogDialog box called for the EditDNA windows to search for aparticular short sequencewithin a DNA sequence. Theshort sequence can includecharacters for representingambiguous bases.SelAppSelAppDlgDialogThis dialog box permits theuser to specify the applicationto which the designed probeswill be destined.SimilLibraryThis unit contains the sourcecode for Tsimilarityclasswhich holds functions andprocedures for the similaritysearch in. It also implementsfunctions for showing adetailed description ofsimilarity betweenoligonucleotides and DNAsequences.SimilDlgSearchSimilarityDlgDialogDialog box called within theEdit DNA windows to searchfor a particular shortsequence within a DNAsequence. This search revealsdetailed information about thesimilarity degree between theprobe and the specific DNAsequence.SimResDlgSimilResDlgDialogHidden Dialog box used onlyfor testing some programmingissues. It shows probesselected at any stage of thedesign and is not available forthe user.SqModuleDataModule1MainDatamodule for the GPDinterfaceprogram. This moduleadministrates the use of all thetables in different points ofthe program (DNA internaldatabases, primary andsecondary tables).TablesLibraryContains definition of alltables used in the GPDprogram and auxiliaryprocedures to create the tablesand to insert data into them.TablesDlgTablesDialogDialogThis dialog box shows a list ofall the available tables whichhave been created in a designproject. The user can combinecreated tables to construct aquery.ToolsLibraryThis unit contains proceduresand functions for routines ofmiscellaneous and generaluse.VHclassLibraryContains the source code forthe Tvirtualhybridiza-tionclass, which is used to predicthybridization patterns.VHConfVHDLGDialogThis dialog box showsparameters for configurationof the VH analysis.VH_formVHFormMainForm for showing the resultsinterfaceof the VH analysis.



EXAMPLE 5

[0127] Source Code of Accessory Programs


[0128] A number of accessory programs that extend the application of the Genosensor Probe Designer (GPD) software are also provided in the computer program listings appendix. The listed source codes include:


[0129] (1) The source code of the program Virtual Hyb. This program performs the Virtual Hybridization analysis independent of the GPD program (within which this function is also incorporated) and has some limited capabilites for manipulating files.


[0130] (2) The source code of a program called AlnClustal. This program is used to select “specific” or “universal” probes from a Clustal alignment of sequences.


[0131] (3) The source code of the program BuiltDB. This program is used to build a FASTA library of sequences (One file containing multiple sequences in FASTA format).


[0132] (4) The source code of a program called Genbank, which converts sequences from the Genbank format into the FASTA format.


[0133] (5) The source code of the program ProbesHom. This is an auxiliary application useful for evaluation of the similarity between multiple probes or primers and single or multiple DNA sequences. It implements a search algorithm which is similar to that used in the VH program, looking for sites with a number of total similar bases or with a block of contiguous similar bases.


[0134] (6) The source code of a program called Structure Windows. This is a program to a simplify evaluation of secondary structure of targets. This program simply locates zones inside of a target molecule that can be self-paired and calculates their free energy value. The program uses an algorithm similar to “build a dot graph,” wherein the DNA sequence is placed on X and Y axes. Coordinates X,Y represent the potential pairing of the Base number X with the base number Y. If these two bases can be paired, a dot is placed in such coordinate. In a graph such as this, diagonals represent zones of the DNA molecules that can be self-paired. This program does not currently show the dot graph but it shows all the sequences in the diagonals.


[0135] (7) The source code of the program Tandem. Tandem is a program that selects sets of probes for the tandem hybridization approach including a capture probe and a stacking probe. This program may be readily modified to accommodate updated information on the thermodynamic properties of tandem hybridization.


[0136] (8) The source code of a program named AdjustLen. This program modifies the length of a set of probes in order to obtain a set probes with decreased Tm variation. The basic GPD program generates probes with the same length and a similar stability. However, stability criteria can be disabled in a GPD session, then the probes can be modified on their length using AdjustLen until their stability approaches a defined Tm value. The algorithm used in AdjustLen adds nucleotides to 3′ or 5′ end, then calculates Tm. Deletion of bases is not currently implemented, so after the AdjustLen program runs the probes will have the same or longer length.


[0137] (9) The source code of a program called MapSeq. This program uses a very fast algorithm to build a “map” of all possible sequences of a given size found within a given DNA sequence.


[0138] (10) The source code of a program named Stability. This program shows graphically the calculation of thermal stability for perfect and mismatched hybrids. Free energy calculations use the same algorithm implemented in the VH algorithm. This function can be conveniently used for fast visualization of energetic contributions in the hybridization between any given pair of strands, as illustrated in the screen image of FIG. 18.


[0139] The Genosensor Probe Design software system is further described in the following Table 9, which includes descriptions of program units contained in the GPD accessory programs.
13TABLE 9Description of Other Units Included in AuxiliaryApplications for the GPD ProgramDelphi FormPascal Unit (*.pas)(*.DFM)TypeDescriptionAdjustLenMainFormForm1MainMain Window of the Adjust LeninterfaceprogramAnalysisLibraryThis unit implementsprocedures to adjust the lengthof probes in order to obtain aset o probes with similarstability but different length.The unit uses the DNAsequences located on both sidesof the region within the DNAsequence from which theprobe was derived. Theprogram adds a nucleotide toone or other side, each timelocating first the nucleotidethat gives the minimal Tmcontribution. The nucleotideaddition is continued until theTm values reaches a defined Tmvalue. This Tm value can bedefined by the user or it can beautomatically selected from themaximal Tm value of theoriginal set of probes. Average,Minimal and Maximal Tmvalues are calculated for theoriginal set of probes and forthe adjusted set. Deletion ofnucleotides from the ends is notcurrently allowed.AlnClustMethodsMainFormMainMain Window of the AlnClustInterfaceprogramCalculateLibraryThis unit implementsprocedures and functions toselect specific (or universal)probes from a ClustalAlignment of DNA sequences.The number of DNA sequencesin the alignment is currentlylimited to 50. The algorithmchecks the alignment columnby column and from eachcolumn the potential probesare derived considering basesto both sides of the column.Selected probes from a samecolumn are checked forsimilarities or differences. Ifthe number of similarities (forfinding universal probes) ordifferences (for specificprobes) exceeds a previousdefined cut-off value, the probeis selected and saved in a textfile where all selected probesare contained. These probescan then be analyzed in theGPD program to refine theselection procedure usingthermodynamic criteria.Routines are also included towrite the alignment, calculateconsensus sequences andindicate the number of basesper column.ParDlgParamsDlgDialogDisplays a dialog box showingthe selection parameters.BuildBuildDBForm1MainMain Form of the BuildinterfaceprogramListBxSelLisDlgDialogShows a dialog box where theuser can select the DNA files tobe used to built a FASTA libraryof the sequences (A filecontaining multiple DNAsequences).StabilityMainMainFormMainMain window of the stabilityinterfaceprogram, which displays adetailed description of thethermodynamic interactionspresent in the hybridizationbetween two short sequences(less than 50 nt).StructureMainForm1MainMain Window of the StructureInterfaceprogramSecondaryLibraryStores segments of the DNAsequence that can be paired(secondary structure elements)in a linked list.DNAclassA variant of the TDNAclasswhich includes methods tocalculate a simplifieddescription of the secondarystructure of a DNA molecule.The algorithm considers thatthe DNA sequence is placed inthe X and Y axes of a coordinatesystem. A search is performedthrough diagonals on suchcoordinate system. CoordinatesX, Y represent bases at positionsX and Y respectively of the DNAsequence. If any two bases canbe paired a value of “1” isassigned to the point on thecoordinate systemcorresponding to their X and Ypositions, otherwise a “0” valueis assigned to such a point. Theexistence of consecutive valuesof “1” in diagonals indicatesstretches of contiguous bases inthe DNA sequence that can pairwith each other. Such baseshave a high probability to bepaired in the secondarystructure of the molecule. Freeenergy values are calculatedfor each section ofcontiguously paired basesfound. The GPD software can bereadily modified to checkwhether selected probes canhybridize with sites that havehigh probability of formingstable secondary structure.Such probes can have a lowprobability to give ahybridization signal.DrawDNALibraryDraws a colored representationof the DNA sequence and savesit as a RTF file.ConfigDlgConfDlgDisplays a dialog box where thesearch for sites with potentialstable secondary structure canbe configured.Genbank converterMainForm1MainMain windows of the genbankInterfaceconverter programAnalysisLibraryContains procedures andfunctions to read files in thegenbank format and to convertthem in files in FASTA format.This unit was the predecessorfor the Formats unit in the GPDprogram.ProbesHomCalculateMain FormMainMain window of the ProbesHominterfaceprogramanalysisLibraryImplementes functions andprocedures to evaluate thedegree of similarity betweenprobes and multiple DNAsequences.OptionsOptionsDlgDialogDisplays a dialog box where thesearch of similarity can beconfigured.MapSeqCodigoLibraryContains procedure to convertDNA sequences to numericrepresentations on base 10 orvice versa (a preliminary stepfor mapping). DNA sequencesare converted in a numberbetween 0 and 4N —1, where N isthe length of the sequence.Several of these routines havebeen already implemented inthe TDNAclass and Toligoclassobjects.Combina2LibraryCalculates all possiblecombinations of sequences thatcan be derived from a sequencewith ambiguous bases.MiniStrLibrayContains source code forbuilding binary maps of all theprobes of a given size that canbe derived from a particularDNA sequence. A binary map isa table where the rows have anumber between 0 and 4N —1.Each row thus represents aspecific sequence of length N.The table has additionalcolumns to show the frequencyof each sequence and it caninclude the positions wheresuch sequence was found. Suchtables are commonly referredto in computer science as LookUp tables. Tables are built inbinary format such that theaccess to each row is very fast.TandemMainMainFormMainMain window for the TandemInterfaceprogramAnalysisLibrarySelects capture and stackingprobes from a particular DNAsequence, for the tandemhybridization approach.Selection is base on the thermalstability of the capture probeand the free energy of thebases located at the junction ofboth probes. The algorithm canbe readily modified to considerfuture more completeknowledge of the stability oftandem probes.ParamsParamsDlgDialogDisplays a dialog box where theparameters of the selection oftandem probes can be specified.VirtualHybMainMainFormMainMain Window of the VirtualInterfaceHyb programAnalysisLibraryImplements all procedures topredict hybridization patternby means of the VH algorithm.DrawDNAMainDisplays a map of the DNAInterfacesequence of interest and thelocation of the hybridizationprobes.



EXAMPLE 6

[0140] Screen Images of the Genosensor Probe Designer Program


[0141] The operation of the Genosensor Probe Designer program is further illustrated in FIGS. 19-25, which display screen images produced by the program's user interface.



EXAMPLE 7

[0142] Installation of Programs


[0143] To install the GPD and VH programs on a PC, setup programs are created with the help of a tool called InstallShield Express for Delphi. The InstallShield Express program first creates files called scripts, which contain all the information required by InstallShield Express to build the setup programs. The scripts for GPD and VH are listed below. To build the setup program, the files called in the script must be saved in the directories specified in the script. The script for installation of Genosensor Probe Designer program [SCRIPT.DOC] is provided in the computer program listings appendix.



EXAMPLE 8

[0144] Tutorial on Use of GPD Program


[0145] A brief tutorial is provided to exemplify the use of the GPD/VH software disclosed herein.


[0146] Case 1: Completely Automatic Design


[0147] 1. Run the GPD program from the “Genosensor Probe Designer” program group, which can be located in the “Programs” menu.


[0148] 2. As soon as the program starts a Welcome Box is presented. Click OK to continue.


[0149] 3. The GPD user interface has several menus, a two tool bars. The left toolbar is intended to be used as Quick Reference for Design.


[0150] 4. Click the “New Project” item in the tool bar or from the menu. A dialog box asking for parameters is presented. The default project's name is “untitled” you need to specify a new name y you want to conserve the results because every time that the program starts the “untitled” project is overridden. For this example we write “Case1” (spaces are not allowed in the name) as the name for this project. The Dialog box permits to specify the initial characteristics for the project. The probe source specifies the origin of the probes. In this example probes will be derived from a specific set of DNA sequences, which will be loaded into an Internal DNA database, as it will be seen soon. The application option is for now disabled, you can select any possible application and it will not have any effect in the design. A target strand must be defined. If the sequence of the probes will be derived from the direct strand (so, the probes will have the same sequence as the direct strand) target strand must be set as “Complementary”. In other words, target strand is the strand against the probes will hybridize. Probe length is the length that all probes selected for the program will have. This version of the program permits only the design of probes of the same length. Oligo concentration, salt concentration and thermodynamic model are required for calculation of Enthalpy, Entropy, Free energy and Tm. Although thermodynamic model permits to be set to Empirical, this option is disabled for now. The creation of an internal DNA database must be checked in this example. The parameters in this box can be changed at any time from the “Options menu” (Or in the left tool bar) in the “Project options” item. When the project is created, its name must be appearing in the text line located in the top tool bar.


[0151] 5. The next step will be to add the DNA sequences, from which probes will be derived, to the internal DNA database. By pressing the “Open DNA sequences” button in the left tool bar a open dialog box is showed and files can be loaded in this box by selecting the desired file(s) and then pressing the “open” button. This dialog permits multiple selection of probes. In order to do this task, files can be selected with the mouse while the <control> or the <shift> keys are pressed. <control> key permits the selection of non-consecutive files in the list while <shift> is used to select multiple consecutive files. The extension “dna” is suggested for DNA sequence files, however other extensions can be used. The only requirement is that files MUST be in Genbank or FASTA formats and only one sequence is permitted by file. If the format of the sequence is not recognized a message indicating that problem is showed and the sequence will not be included in the database. The files used in this example are located in the installation folder. If the path for the installation is the default, this path will be “c:\program files\GPD\examples” For this example select the files “Avel.dna”, “BpuI.dna”, “BspI.dna” and “PaeI.dna”. After pressing the “Open button” the sequences will be read and a table, where their properties are resumed, will be showed. The property “Checked” in this table can be turned to false by typing F and then <Enter>. This property permits include or exclude DNA from specific analysis. By default, all sequences are checked.


[0152] 6. After the criteria has been verified, a primary table of probes will be generated by pressing the button “generate primary table” from the left toolbar. A primary table contains all the probes that can be derived from the checked DNA sequences contained in the internal DNA database. After that the search concludes, a dialog box is showed, indicating the number of different probes that have been found.


[0153] 7. If the item “Selection Criteria” from the Options menu is selected, a Dialog Box is showed with the criteria used for the selection of probes. In the default case, all the options are selected, which means that probes will be selected according with their values of composition and thermodynamic properties resumed in this dialog box. Criteria can be conveniently enabled or disabled for the user. In this example the search of convenient probes will be performed with all the criteria. Specific values used for testing these criteria can be found in the “Criteria Properties” dialog box at the Option menu. Currently values of these properties can not be modified directly in the dialog box.


[0154] 8. After the criteria has been verified, a secondary table must be built. This can be accomplished by pressing the button “Generate Secondary Table” from the left tool bar. When the search is finished a window is showed containing all the selected probes and their properties. Important properties from this window are the sequence, Tm an free energy values, and the file name of the DNA sequence from which the probe was obtained. If this window is empty, this means that the selection criteria used for design were too strong. A new search can be performed, by enabling or disabling some of the criteria used.


[0155] 9. Probes selected in the previous step were selected by considering the frequency of the probe in the primary table if this option was enabled in the “Selection Criteria” dialog box. However probes still are not verified for similarity degree against the DNA sequences in primary table. This step could be necessary when probes, specific for particular targets, are required. Specificity can be checked against the DNA sequences contained in the Internal DNA database or with other DNA sequences (which can be placed in a external DNA database as is described below). Current users must be careful with the parameters of the Similarity search. Parameters can be accessed in the “Similarity Search Parameters” option from the “Options menu”. If probes where derived from the “direct” strand of the DNA sequences contained in the internal DNA database, (which means that they will hybridize with the complementary sequence), similarity must be checked against the direct strand of these sequences in order to obtain probes with low probability of hybridize against the “complementary” sequence of such DNA sequences. Selection is based in the minimal number of total identical bases between probes and sites inside the target DNA sequence required for rejection or a minimal length of identical and consecutive bases between probes and sites. Similarity can be checked against External DNA databases. A external DNA database is a file containing multiple DNA sequences in FASTA format.


[0156] 10. External DNA databases can be created from individual DNA sequences with the utility program “Build Sequence Lib” included in the Genosensor Probe Designer programs group or any other text editor. For this example the Build Sequence Lib program must be run. Then the directory were the individual DNA sequences are located must be specified in the left list. Select: “C:\program files\GPD\examples”. Use the same path for the destination of the library. Set the name of the library to “Example.txt” and the press the “Select files” button. A list of all the files with the extension “*.dna” is showed. Press the button marked with “>” to select all the files. And then “OK”. Then press the button “Build FASTA library”. Close the program and check that the sequence “Example.txt” was saved in the folder.


[0157] 11. In the Similarity Search Parameters dialog box in the GPD program, check the option “External Database”. A Dialog box asking for the name of the file is showed. Look for the “Example.txt” file in the “c:\program files\gpd\examples” path. In this example, the DNA sequences from which probes were selected, are included in the External database, and for this reason, this option must be checked as “YES” in the dialog box. Set the values for the number of similar bases and the block size to 8 for both, and press “OK”.


[0158] 12. Press the button “Search Specific Probes”. Unspecific probes are then removed from the secondary table and a list with 20 specific probes is showed.


[0159] 13. A virtual hybridization analysis (VH) can be performed in order to evaluate the possible hybridization patterns that can be obtained by hybridizing the probes against the sequences contained in the internal DNA database. Virtual Hybridization Parameters can be modified from the “Options” menu. VH analysis evaluates the possibility that such probes can be give a “experimental” hybridization signal, even if there is no perfect match with sites in the target sequences (cross-reaction). For the configuration, user must specify the strand against the probes will hybridize (Target strand), the selection criteria (which are similar to those of the “Similarity analysis, but here, complementarity more than similarity, is evaluated). Results will be formatted here to show only sites with free energy values more stable than the first cut-off value and showing the targeted sites from 3′ to 5′.


[0160] 14. Once the later parameters have been specified, press the “Virtual Hybridization” button in the left tool bar. A list of the probes and potential hybridization sites is shown for each target. This list includes the sequence of the probe and the free energy of its perfect-match hybrid with its specific target. If any probe has potential to hybridize against the tested target a site or sites are shown. Sites are referenced by their position, the sequence of the target at that position (3′ to 5′ in this example) and the free energy value for the hybridization at this site. Free energy values more stable than the first cut-off value, but lower than the second, are showed in red. Sites more stable than the second cut-off value are showed in green. Green sites are considered as the most probable hybridization sites. Usually the sites against the probes will hybridize perfectly, must be showed in this latter color. By combining the information of these list for each sequence a Virtual Hybridization pattern is generated. Patterns must be carefully reviewed in order to see if specific patterns were found for each sequence of the internal DNA database. For the example in case 1 Table 8 is built, representing the hybridization patterns that would be expected for each selected probe against all the DNA sequences tested. It can be seen that specific patterns were found for each sequence. Numbers in red correspond to potential hybridization signals that could be found if the hybridizations conditions are not too stringent.
14TABLE 8Virtual Hybridization Results Obtained For The Data In Case 1ΔG°SEQ ID(perfectBacillusNo.ProbeNo.match)P. aeruginosaB.pumilusspA. veronii1AAAGGTTGG−10.85−10.852AAGGFIGGT−11.29−11.293ACTTTCAGC−10.99−10.994AGGAAAGGT−10.98−10.985AGGTTGGTA−10.87−10.876ATAACTGCC−10.71−7.99−10.717CGAGAGTAA−10.35−10.358CGITCGAAA−11.38−11.389CTGTTTGAC−10.3610.3610GAAAGGTTG−10.31−10.3111GCTGTTTGA−11.16−11.1612GGAAAGGTT−10.70−10.7013GTAACTGCT−10.71−10 71−7.9914TAACTGCCA−11.28−7.99−11.2815TAACTGCTC−10.57−10.57−7.9916TGGTAGCTA−10.69−10.6917TGTTTGACG−11.25−11.25−7.36−7.3618TTCGAAAGG−10.89−10.8919TTGGTAGCT−11.11−11.1120TTTCAGCGA−11.74−11.74


[0161] Case 2: Manual selection of probes.


[0162] 1. Sometimes, the selection of probes cannot be made automatically because probes must be directed against specific sites of the target sequences as in the case of probes for detecting specific mutations. In such cases, the program permits the selection directly from the sequence. For this example we will create a new project after closing all the open windows and going to the “Files” menu to select “Close project”.


[0163] 2. Create a new project calling it “Case2” in the project options dialog box and set the probe source to “Manual input”.


[0164] 3. Open the DNA sequence “Pae2.dna” located in the “Examples” folder. The sequence data must then appear in the table of the Internal DNA database.


[0165] 4. Edit the sequence by using the “Edit Sequence” option from the “Edit menu” or with a double click with the mouse at the line of data for this sequence in the table. The sequence will be showed in the editor. This editor has several useful capabilities. The DNA sequence can be modified from here by adding or deleting bases. The editor has standard tools for Cut, Copy and Paste functions but also has specific functions for searching short sequences. The “Find” option in the “Functions” menu can search short sequences which can include letters for representing ambiguous bases. The function “Search for similarity” permits to evaluate the similarity degree of a short sequence along the DNA sequence. There is also a function to obtain the complementary strand of the entire DNA sequence. The position of the cursor on the DNA sequence in the editor is showed in the status bar at the bottom of program. In this example, with the function “Go to” from the “Functions” menu, go to the position 35. The cursor will be placed at this site. From the “Edit” menu, select “Edit oligonucleotide” and a dialog box will be showed. The dialog box shows information about the probe at that position, such as its length, sequence, complementary strand and thermodynamic properties. A graph of internal stability is included too. The internal stability graph shows the free energy values of groups of internal 5-mers of the probe and is useful for designing PCR primers and for the selection of probes for the tandem hybridization approach (documentation in development). The position of the probe in this dialog can be modified and the changes in the parameters are immediately showed in the dialog. Press the button “Add oligo to database” at the right bottom of this dialog box. A new dialog box asking for the destination of the probes is showed in the screen. If the user wants an automated evaluation of the probes that were manually selected from the sequence, it should be convenient to send the oligo to the primary table. If not evaluation is required, it should be better to send it to the secondary table. For the purposes of this example, the oligo sequence will be sent to the secondary table. From this dialog other probes can be included in the secondary table. Add the sequences corresponding to the positions 47 and 55 too. Close the dialog and the editor.


[0166] 5. Go to the “Results” menu and select “show available tables”. A table containing the selected oligos will be shown with all their thermodynamic properties.


[0167] 6. Go to the “Open DNA sequences” option to add more sequences to the internal DNA database for the Virtual Hybridization analysis. For this example select the sequences “Bpu2.dna” and “Ppu2.dna” and run the VH analysis. In this manual selection all the selected probes have high probability of cross-reaction with all the DNA sequences tested.


[0168] The following references were cited herein:


[0169] Allawi and SantaLucia, (1997) Biochemistry, 36:10581-10594.


[0170] Allawi and SantaLucia, (1998a) Biochemistry, 37:2170-2179.


[0171] Allawi and SantaLucia, (1998b) Nucl. Acids Res., 26:2694-2701.


[0172] Allawi and SantaLucia, (1998c) Biochemistry, 37:9435-9444.


[0173] Beattie, (1997a) Analytical microsystems: Emerging technologies for environmental biomonitoring. In Sayler, G. S., Sanseverino, J. and Davis, K. L. (eds), Biotechnology In The Sustainable Environment. Plenum Press, New York, pp. 249-260.


[0174] Beattie, (1997b) Genomic fingerprinting using oligonucleotide arrays. In Caetano-Anollés, G. and Gresshoff, P. M. (eds), DNA Markers. Protocols, Applications, and Overviews. Wiley-Liss, New York, pp. 213-224.


[0175] Bommarito et al., (2000) Nuc. Acids Res., 28:1929-1934.


[0176] Bushnell et al., (1999) Bioinformatics, 15:348-355.


[0177] Cantor and Schimmel, (1980) Biophysical Chemistry, Part III: The Behavior of Biological Macromolecules. W.H. Freeman and Company, San Francisco, pp. 1109-1265.


[0178] Doktycz et al., (1995) J. Biol. Chem, 270:8439-8445.


[0179] Doktycz and Beattie, (1997) Genosensors and model hybridization studies, In Beugelsdijk, T. J. (ed), Automation Technologies for Genome Characterization. John Wiley & Sons, New York, pp. 205-225.


[0180] Duggan et al., (1999) Nature Genetics Supplement, 21:10-14.


[0181] Galper et al., (1993) Knowledge-based simulation of DNA metabolism: Prediction of action and envisionment of pathways. In Hunter, L. (ed), Artificial Intelligence and Molecular Biology, AAAI Press, pp. 365-395.


[0182] Hacia, (1999) Nature Genetics Supplement, 21:42-47.


[0183] Li and Stormo, (2001) Bioinformatics 17:1067-1076.


[0184] Maldonado-Rodriguez and Beattie, (2001) Analysis of nucleic acids by tandem hybridization on oligonucleotide microarrays. In Rampal, J. B. (ed), DNA Arrays. Methods and Protocols. Humana Press, Totowa, N. J., pp.170:157-71.


[0185] Maldonado-Rodriguez et al., (1999a) Molec. Biotechnology, 11:1-12.


[0186] Maldonado-Rodriguez et al., (1999b) Molec. Biotechnology, 11:13-25.


[0187] Matzura and Wennborg, (1996) Comp. Appl. Biosci.12:247-249.


[0188] Owczarzy et al., (1997) Biopolymers, 44:217-239.


[0189] Peyret et al., (1999) Biochemistry, 38:3468-3477.


[0190] Pozhitkov and Tautz (2002) BMC Bioinformatics 3:9.


[0191] Reyes-Lopez et al., (2003) Nucl. Acids Res., 31:779-789.


[0192] Rychlik, (1993) Selection of primers for polymerase chain reaction. In White, B. A. (ed), PCR Protocols: Current Methods and Applications. Humana Press, Totowa, N. J., pp. 31-40.


[0193] SantaLucia, (1998) Proc. Natl. Acad. Sci., USA, 95:1460-1465.


[0194] Schütz and von Ahsen, (1999) BioTechniques, 27:1218-1224.


[0195] Southern et al., (1999) Nature Genetics Supplement, 21:5-9.


[0196] Thompson et al., (1997) Nucl. Acids Res., 24:4876-4882.


[0197] Vahrson et al., (1996) Comput. Appl. Biosci., 12:119-127.


[0198] Yang et al., (1997) Anal. Chim. Acta, 346:259-275.


[0199] Any patents or publications mentioned in this specification are indicative of the levels of those skilled in the art to which the invention pertains. Further, these patents and publications are incorporated by reference herein to the same extent as if each individual publication was specifically and individually indicated to be incorporated by reference.


Claims
  • 1. A software system for designing and selecting oligonucleotide probes for use in DNA microarray, said software system comprising program instructions for: (a) inputing target DNA sequences; (b) creating an internal DNA database table that includes the entire target DNA sequences from which oligonucleotide probes are selected; (c) generating a primary table of oligonucleotides that contains all possible probes of a defined length that can be derived from said DNA target sequences; (d) creating a secondary table of probes that only contains probes with appropriate composition and thermodynamic properties according to user-defined selection criteria; (e) selecting from said secondary table specific probes based on methods that evaluate similarity between the probes in said secondary table and selected targeted DNA molecules; and (f) performing virtual hybridization to analyze possible hybridization of said oligonucleotide probes against the DNA sequences in said internal DNA database, wherein said analysis shows target sites for perfect hybridization as well as sites that could be potential sites for ambiguous hybridization.
  • 2. The software system of claim 1, wherein said target DNA sequences are selected from the group consisting of single DNA sequence, multiple DNA sequences, DNA sequence database, oligonucleotide sequences and oligonucleotide database.
  • 3. The software system of claim 1, wherein said user-defined selection criteria comprise criteria for rejecting or accepting oligonucleotide probes based upon user-specified criteria selected from the group consisting of [A+T] composition, [G+C] composition, melting temperature (Tm) range, enthalpy, entropy, free energy, internal repeated sequences, sequence symmetry of probes, and frequency of occurrence.
  • 4. The software system of claim 1, wherein said software system comprises program instructions for selecting oligonucleotide probes according to a program application selected from the group consisting of simple selection application, mapping application, non-aligned sequences application, aligned sequences application and tandem hybridization application.
  • 5. The software system of claim 1, wherein said virtual hybridization is performed according the algorithm comprising the steps of: (a) selecting a minimal value (minbasescom) for the number of complementary bases between probes and potential hybridization sites, wherein said minbasescom is defined as 2≦minbasescom≦L (a given probe length); (b) selecting a minimal value (minblocksize) for the length of contiguously paired bases within potential hybridization sites, wherein said minblocksize is specified as 2≦minblocksize≦L (a given probe length); (c) selecting potential hybridization sites by evaluating the complementarity degree between said oligonucleotide probes and potential hybridization sites along the target sequences, wherein said potential hybridization sites are defined as sites where the number of complementary bases between the probe and the evaluated sites is equal to or greater than minbasescom or sites where a block of contiguously paired bases is equal to or greater than minblocksize; (d) calculating the free energy value (ΔG°) for the thermal stability of probes paired with their respective potential hybridization sites using the nearest-neighbor model; and (e) selecting sites of high hybridization probability, wherein said sites of high hybridization probability are potential hybridization sites with ΔG° values equal to or greater than pre-determined ΔG° cutoff values.
  • 6. The software system of claim 1, wherein said selection of oligonucleotide probes from said specific probes comprises adding or deleting nucleotides from one or both ends of said specific probes.
  • 7. A method for designing and selecting oligonucleotide probes for use in DNA microarray, comprising the steps of: (a) inputing target DNA sequences; (b) creating an internal DNA database table that includes the entire target DNA sequences from which oligonucleotide probes are selected; (c) generating a primary table of oligonucleotides that contains all possible probes of a defined length that can be derived from said DNA target sequences; (d) creating a secondary table of probes that only contains probes with appropriate composition and thermodynamic properties according to user-defined selection criteria; and (e) selecting from said secondary table specific probes based on methods that evaluate similarity between the probes in said secondary table and selected targeted DNA molecules; and (f) performing virtual hybridization to analyze possible hybridization of said oligonucleotide probes against the DNA sequences in said internal DNA database, wherein said analysis shows target sites for perfect hybridization as well as sites that could be potential sites for ambiguous hybridization.
  • 8. The method of claim 7, wherein said target DNA sequences are selected from the group consisting of single DNA sequence, multiple DNA sequences, DNA sequence database, oligonucleotide sequences and oligonucleotide database.
  • 9. The method of claim 7, wherein said user-defined selection criteria comprise criteria for rejecting or accepting oligonucleotide probes based upon user-specified criteria selected from the group consisting of [A+T] composition, [G+C] composition, melting temperature (Tm) range, enthalpy, entropy, free energy, internal repeated sequences, sequence symmetry of probes, and frequency of occurrence.
  • 10. The method of claim 7, wherein said software system comprises program instructions for selecting oligonucleotide probes according to a program application selected from the group consisting of simple selection application, mapping application, non-aligned sequences application, aligned sequences application and tandem hybridization application.
  • 11. The method of claim 7, wherein said virtual hybridization is performed according the algorithm comprising the steps of: (a) selecting a minimal value (minbasescom) for the number of complementary bases between probes and potential hybridization sites, wherein said minbasescom is defined as 2≦minbasescom≦L (a given probe length); (b) selecting a minimal value (minblocksize) for the length of contiguously paired bases within potential hybridization sites, wherein said minblocksize is specified as 2≦minblocksize≦L (a given probe length); (c) selecting potential hybridization sites by evaluating the complementarity degree between probes and potential hybridization sites along the target sequences, wherein said potential hybridization sites are defined as sites where the number of complementary bases between the probe and the evaluated sites is equal to or greater than minbasescom or sites where a block of contiguously paired bases is equal to or greater than minblocksize; (d) calculating the free energy value (ΔG°) for the thermal stability of probes paired with their respective potential hybridization sites using the nearest-neighbor model; and (e) selecting sites of high hybridization probability, wherein said sites of high hybridization probability are potential hybridization sites with ΔG° values equal to or greater than pre-determined ΔG° cutoff values.
  • 12. The software system of claim 7, wherein said selection of oligonucleotide probes from said specific probes comprises adding or deleting nucleotides from one or both ends of said specific probes.
  • 13. A method for designing and selecting optimized sets of oligonucleotide probes for use in DNA microarray, comprising the steps of: (a) inputing target DNA sequences; (b) creating an internal DNA database table that includes the entire target DNA sequences from which oligonucleotide probes are selected; (c) generating a primary table of oligonucleotides that contains all possible probes of a defined length that can be derived from said DNA target sequences; (d) creating a secondary table of probes that only contains probes with appropriate composition and thermodynamic properties according to user-defined selection criteria; (e) performing virtual hybridization to predict the hybridization pattern of probes contained in said secondary table; and (f) eliminating probes which have the potential for ambiguous hybridization, thereby yielding an optimized set of specific probes for use in DNA microarray.
CROSS-REFERENCE TO RELATED APPLICATION

[0001] This non-provisional patent application claims benefit of provisional patent application U.S. Serial No. 60/371,113, filed Apr. 9, 2002, now abandoned.

Provisional Applications (1)
Number Date Country
60371113 Apr 2002 US