Bio tics

8/6/2019 Bio tics

1/43

BioinformaticsBioinformatics ABE 2007 ABE 2007

Kent Koster Kent Koster Group 3Group 3

8/6/2019 Bio tics

2/43

Why bioinformatics?Why bioinformatics?

Other tec hniques raise more questionsOther tec hniques raise more questionsthan t hey answer. Bioinformatics is w hatthan t hey answer. Bioinformatics is w hat

answers t he questions t hose tec hniquesanswers t he questions t hose tec hniquesgenerate.generate.

8/6/2019 Bio tics

3/43

OutlineOutline

Bioinformatics DefinedBioinformatics DefinedEvolution of BioinformaticsEvolution of Bioinformatics

Bioinformatics Histor yBioinformatics Histor yCommon Uses of BioinformaticsCommon Uses of BioinformaticsProcedures and Tools of BioinformaticsProcedures and Tools of BioinformaticsOur ProcedureOur ProcedureOur ResultsOur ResultsResourcesResources

8/6/2019 Bio tics

4/43

Bioinformatics DefinedBioinformatics Defined

Bioinformatics is broad term covering t he use of Bioinformatics is broad term covering t he use of computer algorit hms to anal yze biological data.computer algorit hms to anal yze biological data.Differs from computational biolog y in that w h ileDiffers from computational biolog y in that w h ilecomputational biolog y is the use of computer computational biolog y is the use of computer tec hnolog y to solve a single, hypothesistec hnolog y to solve a single, hypothesis--basedbasedquestion, bioinformatics is t he omnibus use of question, bioinformatics is t he omnibus use of computeri zed statistical anal ysis to makecomputeri zed statistical anal ysis to makestatistical or comparative inferences.statistical or comparative inferences.i.e. converting data to information.i.e. converting data to information.

8/6/2019 Bio tics

5/43

The nebulous genesis of The nebulous genesis of

bioinformaticsbioinformatics19771977 - -X174 P hage Genome sequencedX174 P hage Genome sequenced19901990 Paper publis hed in t hePaper publis hed in t he Journal of Journal of Molecular Biology Molecular Biology describes sequencedescribes sequencealignment searc h algorit hmalignment searc h algorit hm1990s1990s Software used to find fragment overlapSoftware used to find fragment overlapfor the Human Genome Projectfor the Human Genome Project

19921992 NCBI takes over GenBank DNANCBI takes over GenBank DNAsequence database in response to t he growingsequence database in response to t he growingnumber of gene patentsnumber of gene patents

8/6/2019 Bio tics

6/43

The nebulous genesis of The nebulous genesis of

bioinformaticsbioinformatics19941994 Entre z Global Quer y CrossEntre z Global Quer y Cross--DatabaseDatabaseSearc h System allows users to searc h GenBankSearc h System allows users to searc h GenBankdatabasedatabase

19951995 Dr. Owen Wh ite writes software to helpDr. Owen Wh ite writes software to helpfind gene elements (promoters, start and stopfind gene elements (promoters, start and stopcodons, etc.) in t he sequenced Haemop h iluscodons, etc.) in t he sequenced Haemop h ilusinfluen zae genomeinfluen zae genome

19961996 NCBINCBI--BLAST created to provide powerfulBLAST created to provide powerfulheuristic searc hes against t he GenBankheuristic searc hes against t he GenBankdatabasedatabase

8/6/2019 Bio tics

7/43

Genomics to Proteomics t hroug h Genomics to Proteomics t hroug h

BioinformaticsBioinformaticsBecause proteins are ultimatel y the tool of all* geneBecause proteins are ultimatel y the tool of all* geneexpression, proteomics is, in effect, t he product scienceexpression, proteomics is, in effect, t he product sciencemade possible b y bioinformaticsmade possible b y bioinformatics

A proteome is t he collection of all proteins expressed in A proteome is t he collection of all proteins expressed ina cell at a given timea cell at a given timeEver y organism has 1 genome, but man y proteomesEver y organism has 1 genome, but man y proteomesIn addition to h igh throug hput protein anal ysis,In addition to h igh throug hput protein anal ysis,proteomics is researc hed t hroug h cDNA anal ysis (RTproteomics is researc hed t hroug h cDNA anal ysis (RT--

PCR)PCR)Proteomics represents a met hodical addition of largeProteomics represents a met hodical addition of largescale biolog y to traditional molecular biolog y, madescale biolog y to traditional molecular biolog y, madepossible b y bioinformaticspossible b y bioinformatics

8/6/2019 Bio tics

8/43

Common Uses of BioinformaticsCommon Uses of Bioinformatics

Homolog y and Comparative ModelingHomolog y and Comparative ModelingProtein or gene homolog y is s haredProtein or gene homolog y is s harednucleotide or amino acid sequences or nucleotide or amino acid sequences or domains s hared between different proteinsdomains s hared between different proteinsregardless of w hether from same or differentregardless of w hether from same or differentorganismorganism

Gene or Protein IdentificationGene or Protein IdentificationSearc h ing databases for nucleotide or aminoSearc h ing databases for nucleotide or aminoacid sequences t hat matc h sequences inacid sequences t hat matc h sequences inunknown samplesunknown samples

8/6/2019 Bio tics

9/43

So, how do ya do it?So, how do ya do it?

DNA SequencingDNA SequencingSequence FormatsSequence Formats

Sequence Homolog y Software ToolsSequence Homolog y Software Tools Aligning Tools Aligning Tools Annotated Information Annotated Information

Protein FoldingProtein Folding

8/6/2019 Bio tics

10/43

DNA SequencingDNA Sequencing

Sanger Met hodSanger Met hodNew nucleotide c hains of DNA beingNew nucleotide c hains of DNA beingreplicated b y DNA Polymerase are stoppedreplicated b y DNA Polymerase are stoppedwhen diwhen di--deox y nucleotides (added in t hedeox y nucleotides (added in t hereaction mixture in ~1/100 ratio) arereaction mixture in ~1/100 ratio) areincorperated into t he c hainincorperated into t he c hain

8/6/2019 Bio tics

11/43

DNA SequencingDNA Sequencing

Fluorescent d yes are bound to t heFluorescent d yes are bound to t heddNTPs, allowing t he molecule to detectedddNTPs, allowing t he molecule to detectedwhen it is excited b y a laser when it is excited b y a laser Terminated DNA c hains are run on a gel,Terminated DNA c hains are run on a gel,and fragments are resolved b y sizeand fragments are resolved b y sizeBy combining t he fluorescence readingsBy combining t he fluorescence readingsfrom eac h size nucleotide c hain, t he DNAfrom eac h size nucleotide c hain, t he DNAsequence is computedsequence is computed

8/6/2019 Bio tics

12/43

8/6/2019 Bio tics

13/43

Example SequenceExample Sequence

Ch romatograp hCh romatograp h

8/6/2019 Bio tics

14/43

Sequence Anal ysisSequence Anal ysisFirst T h ings FirstFirst T h ings First Sequence File Formats:Sequence File Formats:Most common for nucleotides: FASTA / MultiMost common for nucleotides: FASTA / Multi- -FASTAFASTA> followed b y an y unicode text, entire line read as sequence title> followed b y an y unicode text, entire line read as sequence titleCarriage return followed b y continuous 5Carriage return followed b y continuous 5-- 3 nucleotide sequence or 3 nucleotide sequence or protein sequence using 1protein sequence using 1- -letter codesletter codesExample:Example:>E. coli Globin>E. coli Globin--coupled c hemotaxis sensor y transducer (TMcoupled c hemotaxis sensor y transducer (TMdomain)domain)

ATGGA CC TGAT C AC AAATGCGATTTAGAGA CC TGAT C AC AAATG ATGGA CC TGAT C AC AAATGCGATTTAGAGA CC TGAT C AC AAATGCGATGA CC TGATC AC AAATGCGATGACC TGAT C AC AAATGCGACGATGA CC TGATC AC AAATGCGATGACC TGAT C AC AAATGCGATGTAAACC TGAT C AC AAATGCGATGACC TGAT C AC AAATGCGATTGTAAACC TGAT C AC AAATGCGATGACC TGAT C AC AAATGCGAT

CTAAACC TGAT C AC AAATGCGATGACC TGATC AC AAATGCGATTCTAAACC TGAT C AC AAATGCGATGACC TGATC AC AAATGCGATT AA AA

8/6/2019 Bio tics

15/43

Sequence Homolog y SoftwareSequence Homolog y SoftwareNCBINCBI--BLASTBLAST

Run b y the National Center for Biotec hnolog y Run b y the National Center for Biotec hnolog y InformationInformationBLAST uses a heuristic algorit hm based on t heBLAST uses a heuristic algorit hm based on t heSmit hSmit h--Waterman algorit hmWaterman algorit hm

Algorithm searc hes database for a small string wit h in Algorithm searc hes database for a small string wit h inthe quer y (default 11 for nucleotide searc hes), t henthe quer y (default 11 for nucleotide searc hes), t henwhen it detects a matc h , searc hes for s haredwhen it detects a matc h , searc hes for s harednucleotides at eac h end of t he seed to extend t henucleotides at eac h end of t he seed to extend t he

matc hmatc hGaps are taken into account, t hen t he matc hes areGaps are taken into account, t hen t he matc hes arepresented in order of statistical significancepresented in order of statistical significancehttp://www.ncbi.nlm.ni h .gov/BLAST/http://www.ncbi.nlm.ni h .gov/BLAST/

8/6/2019 Bio tics

16/43

Different T ypes of BLASTDifferent T ypes of BLAST

NucleotideNucleotide--nucleotide BLAST (BLASTN):nucleotide BLAST (BLASTN):Basic nucleutide sequence searc hesBasic nucleutide sequence searc hesThe BLAST t hat you used for your sequencesThe BLAST t hat you used for your sequences

ProteinProtein--protein BLAST (BLASTP):protein BLAST (BLASTP):Similar tec hnolog y used to searc h amino acidSimilar tec hnolog y used to searc h amino acidsequencessequences

PositionPosition--Specific Iterative BLAST (PSISpecific Iterative BLAST (PSI- -BLAST):BLAST): A more advance protein BLAST useful for anal yz ing A more advance protein BLAST useful for anal yz ingrelations hips between divergentl y evolved proteins.relations hips between divergentl y evolved proteins.

8/6/2019 Bio tics

17/43

Different T ypes of BLASTDifferent T ypes of BLAST

BLASTX and BLASTN variants:BLASTX and BLASTN variants:Use sixUse six--frame translation for proteins andframe translation for proteins andnucleotides, respectivel y, in the searc hnucleotides, respectivel y, in the searc h

MegaBLAST:MegaBLAST:Used for BLASTing several sequences atUsed for BLASTing several sequences atonce to cut down on processing load andonce to cut down on processing load andserver reportingserver reporting- -timetime

8/6/2019 Bio tics

18/43

Interpreting BLAST ResultsInterpreting BLAST ResultsMax/Total ScoreMax/Total Score

Calculated from t he number of matc hes and gaps.Calculated from t he number of matc hes and gaps.Higher relative to your quer y lengt h is better Higher relative to your quer y lengt h is better

E Value: E=KmnE Value: E=Kmn(e(e-- S)S)

Translation: E Value gives you t he number of entriesTranslation: E Value gives you t he number of entriesrequired in t he database for a matc h to happen b y required in t he database for a matc h to happen b y random c hance. e.g. E=erandom c hance. e.g. E=e --66 means t hat one matc h means t hat one matc h would be expected for ever y 1,000,000 entries in t hewould be expected for ever y 1,000,000 entries in t he

databasedatabaseSmaller E Values are better Smaller E Values are better Values larger t han E=eValues larger t han E=e --55 too likely to be due totoo likely to be due tochancechance

8/6/2019 Bio tics

19/43

Interpreting BLAST ResultsInterpreting BLAST Results

Quer y CoverageQuer y CoverageThe percent of t he quer y sequence matc hedThe percent of t he quer y sequence matc hedby the database entr yby the database entr y

Max IdentMax IdentThe percent identit y, i.e. t he percent t hat t heThe percent identit y, i.e. t he percent t hat t hegenes matc h up with in the limits of t he fullgenes matc h up with in the limits of t he full

matc h (e.g. deletions or additions reduce t h ismatc h (e.g. deletions or additions reduce t h isvalue)value)

8/6/2019 Bio tics

20/43

Sequence Aligning SoftwareSequence Aligning Software

C lustal (free)C lustal (free)C lustalXC lustalX SoftwareSoftware

C lustal W C lustal W WebWebDNAStar ($$$)DNAStar ($$$)Functionalit y is similar, but difference is inFunctionalit y is similar, but difference is ininterface, tools, and speed of algorit hmsinterface, tools, and speed of algorit hmsh ttp://www.ebi.ac.uk/clustalw/h ttp://www.ebi.ac.uk/clustalw/

8/6/2019 Bio tics

21/43

SMARTSMART

SimpleSimple Modular Modular Arch itecture Arch itecture Researc h Researc h ToolTool

Run b y EMBL (European Molecular Run b y EMBL (European Molecular Biology Laborator y)Biology Laborator y)Wh ile BLAST compares nucleotideWh ile BLAST compares nucleotidesequences and t hen informs you of an y sequences and t hen informs you of an y domains t hat ma y have been annotated todomains t hat ma y have been annotated tothem, SMART compares b y domainsthem, SMART compares b y domains

8/6/2019 Bio tics

22/43

PFAMPFAM

Protein domain databaseProtein domain databaseManuall y curated, trading volume for qualit yManuall y curated, trading volume for qualit yUses h idden Markov models for domainUses h idden Markov models for domainpattern recognitionpattern recognitionRun b y Sanger Institute in t he UKRun b y Sanger Institute in t he UKHeuristic server Heuristic server- -load anal ysis predicts w hen ke y load anal ysis predicts w hen ke y

protein anal ysis report is due and cras hes server protein anal ysis report is due and cras hes server h ttp://www.sanger.ac.uk/Software/Pfam/h ttp://www.sanger.ac.uk/Software/Pfam/

8/6/2019 Bio tics

23/43

InterproInterpro

Database of protein domains andDatabase of protein domains andfunctional sitesfunctional sites

Best source of annotationBest source of annotationOther tools sometimes draw annotationOther tools sometimes draw annotationfrom Interprofrom Interpro

Run b y the European BioinformaticsRun b y the European BioinformaticsInstituteInstituteh ttp://www.ebi.ac.uk/interpro/h ttp://www.ebi.ac.uk/interpro/

8/6/2019 Bio tics

24/43


Lowest energ y state foldingLowest energ y state foldingAb initio: Ab initio: tremendousl y resource heav y, cantremendousl y resource heav y, canonly be done for tin y proteinsonly be done for tin y proteinsDistributed computing is used for midDistributed computing is used for mid- -sizedsizedproteinsproteins

Folding@HomeFolding@Home

Human Proteome Folding ProjectHuman Proteome Folding ProjectRosetta@HomeRosetta@HomePredictor@HomePredictor@Home

8/6/2019 Bio tics

25/43


SoftwareSoftware--assisted manual foldingassisted manual foldingUse knowledge of bioc hemistr y to fold proteinUse knowledge of bioc hemistr y to fold proteininto predicted structure, t hen software to findinto predicted structure, t hen software to findlowest energ y statelowest energ y state

Commercial Programs:Commercial Programs:Protein S hopProtein S hop

ProfoldProfold

8/6/2019 Bio tics

26/43

Manual Motif VerificationManual Motif Verification

Ramac handran PlotRamac handran Plot ratio of to ratio of to angles on N and C terminals of subunitangles on N and C terminals of subunit

8/6/2019 Bio tics

27/43

Our ProcedureOur Procedure

Colonies were selected from nutrient platesColonies were selected from nutrient platesEac h group selected two colonies to sequenceEac h group selected two colonies to sequenceColonies w h ich survived ampicillin treatment wereColonies w h ich survived ampicillin treatment were

possibl y transformed b y the vector, w hich containedpossibl y transformed b y the vector, w hich containedan ampicillin resistance genean ampicillin resistance genePresence of PDI insert was expected to disrupt ccdBPresence of PDI insert was expected to disrupt ccdB(lethal protein) and LacZ gene expression in vector (lethal protein) and LacZ gene expression in vector

plasmidplasmidLacZ expression resulted in some blue colonies, asLacZ expression resulted in some blue colonies, asthe colonies were able to cleave Xthe colonies were able to cleave X- -Gal substrate intoGal substrate intoblue productblue product

8/6/2019 Bio tics

28/43

Initial Questions Guiding Colony Initial Questions Guiding Colony SelectionSelection

How did some blue colonies survive?How did some blue colonies survive?Did all blue colonies come from t he P CR product?Did all blue colonies come from t he P CR product?Did the w h ite colonies contain t he PDI inserts?Did the w h ite colonies contain t he PDI inserts?

Were some colonies able to survive wit hout t heWere some colonies able to survive wit hout t heampicillin resistance plasmid?ampicillin resistance plasmid?Wh at was t he actual sequence of t he commercialWh at was t he actual sequence of t he commercialpositive control insert?positive control insert?Some samples were transformed wit h inserts collectedSome samples were transformed wit h inserts collected

from P CR instead of gel electrop horesis. Could havefrom P CR instead of gel electrop horesis. Could havenonnon--PDI sequences have ligated to t he vector and beenPDI sequences have ligated to t he vector and beeninserted into bacteria?inserted into bacteria?

8/6/2019 Bio tics

29/43

ProcedureProcedure

Samples were prepared wit h T3 and T7Samples were prepared wit h T3 and T7(forward and backward) primers in solution(forward and backward) primers in solutionfor sequencingfor sequencingSamples were sent to UH Manoa lab for Samples were sent to UH Manoa lab for sequencingsequencingCh romatogram results were viewed wit h Ch romatogram results were viewed wit h Finc h TV to determine qualit yFinc h TV to determine qualit y

8/6/2019 Bio tics

30/43

ProcedureProcedureSequences were trimmed at 5 and 3Sequences were trimmed at 5 and 3ends, t hen restriction en zyme sites on t heends, t hen restriction en zyme sites on t hevector were attempted to be located wit h vector were attempted to be located wit h Finc h TVFinc h TV

8/6/2019 Bio tics

31/43

ProcedureProcedure

Sequences were exported in FASTA formatSequences were exported in FASTA formatProcedure was repeated for t he ot her strandsProcedure was repeated for t he ot her strandsPair Pair--wise alignment was performed for bot h wise alignment was performed for bot h strands of eac h sample wit h EBIs toolsstrands of eac h sample wit h EBIs toolsConsensus sequence from pair Consensus sequence from pair- -wise alignmentwise alignmentwas searc hed for in BLASTwas searc hed for in BLAST

Gene information was located from BLASTGene information was located from BLASTannotation and TAIR websiteannotation and TAIR website

8/6/2019 Bio tics

32/43

8/6/2019 Bio tics

33/43

Group 3 ResultsGroup 3 Results

Sequenced 1 blue and 1 w h ite colon y fromSequenced 1 blue and 1 w h ite colon y fromsame platesame plateColonies were transformed wit h PCRColonies were transformed wit h PCRproduct, not gelproduct, not gel- -recovered DNArecovered DNAWh ite colonies had PDI insertWh ite colonies had PDI insertBlue colonies had 154Bp partial insert,Blue colonies had 154Bp partial insert,

disrupting ccdB gene, but remaining indisrupting ccdB gene, but remaining in- -frame and allowing for a partiall y functionframe and allowing for a partiall y functionLacZ alp ha gene to be expressedLacZ alp ha gene to be expressed

8/6/2019 Bio tics

34/43

Group 3 Wh ite Colon yGroup 3 Wh ite Colon y

T7 strand definitivel y showed t he presenceT7 strand definitivel y showed t he presenceof a PDI insertof a PDI insert

8/6/2019 Bio tics

35/43

8/6/2019 Bio tics

36/43

8/6/2019 Bio tics

37/43

Group 3 Blue Colon yGroup 3 Blue Colon yHowever, T7 strand was salvageableHowever, T7 strand was salvageable

A 154 nucleotide sequence was found A 154 nucleotide sequence was foundbetween t he restriction sitesbetween t he restriction sites

8/6/2019 Bio tics

38/43


Wh ite Colon y from P CR product s howedWh ite Colon y from P CR product s howedPDI gene in bot h T3 and T7 strandsPDI gene in bot h T3 and T7 strands

Wh ite colon y from gel purification:Wh ite colon y from gel purification:T7 strand sequenced as multiple signalsT7 strand sequenced as multiple signalsT3 strand sequenced excellentl y T3 strand sequenced excellentl y

8/6/2019 Bio tics

39/43

Group 1 Gel Wh ite Colon yGroup 1 Gel Wh ite Colon y

T3 sequence s howed onl y nucleotidesT3 sequence s howed onl y nucleotides15401540--2320 of t he vector 2320 of t he vector

8/6/2019 Bio tics

40/43


Wh ite Colon y from gel purificationWh ite Colon y from gel purificationWh ite colonies sequenced wit h PDI geneWh ite colonies sequenced wit h PDI gene

Blue w/ Wh ite Ring Colon y from P CRBlue w/ Wh ite Ring Colon y from P CRBoth T3 and T7 strand sequencing s howedBoth T3 and T7 strand sequencing s howedconsistent multiple signalsconsistent multiple signals

8/6/2019 Bio tics

41/43


1 wh ite colon y from P CR and 1 w h ite1 wh ite colon y from P CR and 1 w h itecolon y from gel purification werecolon y from gel purification weresequencedsequencedBoth showed PDI geneBoth showed PDI gene

8/6/2019 Bio tics

42/43

Final RemarksFinal Remarks All white colonies had t he PDI gene, except one wit h a modified vector All white colonies had t he PDI gene, except one wit h a modified vector All blue colonies were transformed wit h the direct P CR product (not gel All blue colonies were transformed wit h the direct P CR product (not gelpurified)purified)Group 3 s howed t hat a small (154Bp) insert t hat sta ys inGroup 3 s howed t hat a small (154Bp) insert t hat sta ys in--frame wit h theframe wit h theLacZ gene can knockLacZ gene can knock- -out t he ccdB, w hile still allowing t he expression of anout t he ccdB, w hile still allowing t he expression of anat least partiall y functioning LacZ geneat least partiall y functioning LacZ geneSome blue colonies wit h wh ite rings could be 2 separate lines livingSome blue colonies wit h wh ite rings could be 2 separate lines livingtoget her toget her

Bacteria transformed wit h ampicillin resistance gene could deplete area of Bacteria transformed wit h ampicillin resistance gene could deplete area of ampicillin, allowing bacteria wit hout t he gene to crowd t he w h ite bacteria out of ampicillin, allowing bacteria wit hout t he gene to crowd t he w h ite bacteria out of the area of depleted ampicillinthe area of depleted ampicillinHow could bacteria wit hout t he insert survive bot h ccdB expression and ampicillinHow could bacteria wit hout t he insert survive bot h ccdB expression and ampicillinselection in brot h?selection in brot h?

ccdB gene could be lost due to mutationccdB gene could be lost due to mutationBactaria could have cut plasmid, deleting t he ccdB, but retaining LacZ possibl y andBactaria could have cut plasmid, deleting t he ccdB, but retaining LacZ possibl y andampicillin resistance genesampicillin resistance genes

No group sequenced t he positive control insertNo group sequenced t he positive control insert sequence still a m yster y!sequence still a m yster y!

8/6/2019 Bio tics

43/43

ResourcesResourcesh ttp://www.bioinformatics.orgh ttp://www.bioinformatics.orgh ttp://http://s yntheticbiolog y.org/Tools. h tmlh ttp://http://s yntheticbiolog y.org/Tools. h tmlNCBI BLAST:NCBI BLAST: h ttp://www.ncbi.nlm.ni h .gov/BLAST/h ttp://www.ncbi.nlm.ni h .gov/BLAST/SMART:SMART: h ttp://smart.emblh ttp://smart.embl- -heidelberg.de/heidelberg.de/PFAM:PFAM: h ttp://www.sanger.ac.uk/Software/Pfam/h ttp://www.sanger.ac.uk/Software/Pfam/Interpro:Interpro: h ttp://www.ebi.ac.uk/interpro/h ttp://www.ebi.ac.uk/interpro/Canadian Bioinformatics Helpdesk Newsletter (Ramac handranCanadian Bioinformatics Helpdesk Newsletter (Ramac handranPlot):Plot):h ttp://gc helpdesk.ualberta.ca/news/22sep05/cb hd_news_22sep05.ph ttp://gc helpdesk.ualberta.ca/news/22sep05/cb hd_news_22sep05.phphpFinc h TV:Finc h TV: http://www.geospi za.com/finc htv/http://www.geospi za.com/finc htv/EBI Pair EBI Pair--wise alignment:wise alignment:h ttp://www.ebi.ac.uk/emboss/align/index. h tmlh ttp://www.ebi.ac.uk/emboss/align/index. h tmlTAIR:TAIR: h ttp://www.arabidopsis.orgh ttp://www.arabidopsis.org

Bio tics

Documents