GenomicPeptideFinder:

Mass spectrometric data mining in genomic data bases

Institute for Plant Biochemistry and Biotechnology University of Münster, Germany

Peng, J. and Gygi, S.P. (2001) Proteomics: the move to mixtures. J. Mass Spectrom., 36,

1083-1091.

Schematic depiction of an ion trap mass spectrometer

Mass Spectra Recording (Triple Play)

35

Rel

ativ

e A

bund

ance

15

60

85

1157.5703.2 885.0578.8 765.9

+ c Full ms [ 400.00-2000.00]

400 600 800 1000 1200 1400 1600 1800 2000m/z

0

5

10

20

25

30

40

45

50

55

65

70

75

80

90

95

100626.3

835.5

982.4

610.21054.4

1156.2852.2

503.9 1217.7445.1 1469.71259.8

1

Scan 4501 2

+ d Z ms [ 622.30-632.30]

622 623 624 625 626 627 628 629 630 631 632m/z

05

1015

20

25

30

35

40

45

50

55

60

65

70

75

8085

90

95

100 626.1

626.6

627.1

627.7

Scan 4502

+ c d Full ms2 626.28@35.00 [ 160.00-1890.00]

200 400 600 800 1000 1200 1400 1600 1800m/z

0

5

10

15

20

25

30

35

40

45

50

55

60

65

70

75

80

85

90

95

100

Rel

ativ

e A

bund

ance

479.4535.8

828.2

957.3

715.2958.2

1070.3406.2 602.2

3

Scan 4503

4500 4505

Fragmentation Spectrum

M /Z

2 0 0 4 0 0 6 0 0 8 0 0 1 0 0 0 1 2 0 0 1 4 0 0 1 6 0 0

Inte

nsity

0

2 0 0 x 1 0 3

4 0 0 x 1 0 3

6 0 0 x 1 0 3

8 0 0 x 1 0 3

1 x 1 0 6

SLSPAQVSFAVEAK G V

KVAEVAFSVQAPS L S

b 1 4b 1 3b 1 2b 1 1b 1 0b 9b 8b 7b 6b 5b 4b 3

y 1 5y 1 4y 1 3y 1 2y 1 1y 1 0y 9y 8y 7y 6y 5y 4y 3

G

b 1 5

H2N—CH—C—NH—CH—C—NH—CH—C—NH—CH—C—OH

R1 R3R2 R4

O OO O

a

x

b

y

c

zsslifi

Correlation analysis

# deltCn XCorr Sp Reference Peptide1. 0.0000 4.2902 382.4 1152RZ (-)GSGDAAYPGGPFFNLFNLGK2. 0.5359 1.3963 70.0 582RZ (V)VCLMSAIALPYNGGLRPPGV3. 0.6028 1.1949 43.1 258RZ (-)GGGVVCFDASFRSPKFGIGK5. 0.6160 1.1553 70.1 2360RZ (-)VDSGWGGVVVVALAPYNLGR6. 0.6666 1.0029 37824 37RZ (K)CVLWAYPGVALGRVQVLGK11. 0.7333 0.8022 37647 304RZ (V)GGWLGRLVGGATVLGKAYPGV12. 0.7351 0.7970 29.2 1080RZ (-)IAAYPGVSPGLMIHYNIGR13. 0.7774 0.6695 22.0 304RZ (K)AYPGVGELVRGGCMMCSIGK14. 0.7842 0.6492 31.4 37RZ (W)AYPGVALGRGIRLGKCVLW

Database selection

>1080ZRIAAYPGVSPGLMIHYNIGR>1137RZAAYPGATQPGATELARRLGK>1152RZGSGDAAYPGGPFFNLFNLGK>1152ZRGSGDAAYPGGPFFNLFNLGK>2360RZVDSGWGGVVVVALAPYNLGR>240RZHPGVVCRPGRGGGCSRHIGKHPGVVCCSRHRRSHTIGK

Problem

• Well resolved mass spectra without identification after database search

• De novo amino acid sequencing with subsequent manual search lead to peptide identification

Sequence availability• Every fourth peptide escapes detection by MS due to an intron which

splits the peptide on the “genomic level”– Choudhary, J. S., Blackstock, W. P., Creasy, D. M. and Cottrell, J. S. (2001) Interrogating the human genome using

uninterpreted mass spectrometry data.Trends Biotechnol, 19, S17-22.

• Each human and Chlamy gene statistically contains 8 exons– Deutsch, M. and Long, M. (1999) Intron-exon structures of eukaryotic model organisms. Nucleic Acids Res, 27, 3219-3228.

Merchant, S. et al., (2007) Science; in press

• 40-60% of human genes are alternatively spliced– Modrek, B. and Lee, C. (2002) A genomic view of alternative splicing. Nat Genet, 30, 13-19.

• 50% of a genome may not be complementable by EST data– Seki, M., Narusaka, M., Kamiya, A., Ishida, et. al. (2002) Functional annotation of a full-length Arabidopsis cDNA collection.

Science, 296, 141-145.

• Gene prediction is far from being correct (Only 15-20% of known gene structures are predicted correctly.)

– Brent, M.R. and Guigo, R. (2004) Recent advances in gene structure prediction. Curr Opin Struct Biol, 14, 264-272.

• Post translational modifications– Perkins, D. N., Pappin, D. J., Creasy, D. M. and Cottrell, J. S. (1999) Probability-based protein identification by searching

sequence databases using mass spectrometric data. Electrophoresis, 20, 3551-67.

intron

Right Peptide FragmentLeft Peptide Fragment

Intron

GGPFFNLFNLGK......(K)GSGDAAYP

GTGAGCAGATAGAGGAGAGAGCGCGCGAGAGAGGGCCGCGGCTGCAGCTGGTGTGGGCCAAGGCTGGCGGAGACAGCCAGCAGGGGTCAGGGGGAGGGGCACAGGGCAGAGGCAGACGCCGGCGCTGTCGGGTGCGGGATGCGGGTTGCTGATGGATGGAAACTCCGATCCGCGGCGGACGGTTTCGTGCTATGTAGCTGCTCAACAGGGTTTGCATGCTCCTGGCTGACAGCGCATATGACGGTCCCTTCCCCCGCTCCGCAG

exonintron

Found result

Display resultDeNovo-Sequencingof MS/MS data

RunGenomicPeptideFinder

Database search Found result

Select genomicdatabase

Set optionsRun query

Yes

Yes

MS/MSData

No

Selectsquencedatabase

Verification bydatabase search

0

1

2

3

4

5

6

180 380 580 780 980 1180

M/Z

Inte

nsity

* 1

000

000

KA I A Q L E E D Y L

LY D E E L Q A I A K

De Novo Sequencing

1016.4901.6114.8 ~ D (115.02)D

DE

901.6772.4129.2 ~ E (129.1)E

nsi78_11.1803.1806.2.dtaSequence Absolute Relative

Probability Probability[LW]QYSEVLH[AR] 6.4% 41%[PD]SQYSQVLH[AR] 11.5% 31.6%WLQYSEVLH[AR] 4.3% 7.2%

… … …

De Novo Amino Acid Sequence Predictions

De novo amino acidsequencing

Foundresult

Select databaseEnter querySet optionsRun GPF

MS/MSdata

RunGenomicPeptideFinder

Search for sequences for which the mass of the

precursor ion matches the mass of thetryptic peptide.

Store significantcandidates in theresult database

Foundresult

As process 2 butallow for sequenceerrors within the

database.

Foundresult

Combine sequences andcut out putative introns.compare the mass of

the resulting tryptic peptideto the mass of the

precursor ion.

Foundresult

As process 2 butallow for errors in thequery correct these

with the sequences fromthe database.

Process 1 Process 2 Process 3 Process 4

Yes Yes Yes Yes

First (stringent)sequence comparison

Second (less stringent)sequence comparison

RF1: PDPRCHQHQVVRERRNPPRRLVQRLLG-PLHHLLRGDRGYAVR-AAPPSGAGREGQALVGRF2: SIFRHPKQDFGFDPLGLLDPVNSGGFIEPKWLQYSEVIHARWAMLGAAGCIAPEVLGAAGRF3: -SPMPPTSSGSRAA-SPPPARTTATGLTPTPSSSWRSWLCSSLSCAAFRCGEGGAGARRARF4: PLLAVSASLGPHQLQPRPSLARSLLYLLTGVGSVARALEDRLQAQEVLLAHGAGVAEVLQRF5: PCWLSPPALAHTSCSRGPLSRALSSICSPG-AASPEPLKIASRPRKYCWPMEPG-RKSCRRF6: THARRGSVIGCNCNQPQLQRCTRGFK-NRFLRRFFFFARRAPAPPSPHLKAAQLSELHSH

WLQYSEVIHAR 1402.4 Da nsi_63_1.1641.1644.2WLQYS

WLQYS RK EVIHA

Measured Mass: 1402.40 Da

Calculated Mass: 1401.58 Da

Mass Difference: 0.82 Da < 700 ppm (0.98 Da)

De novo amino acidsequencing

Foundresult

Select databaseEnter querySet optionsRun GPF

MS/MSdata

RunGenomicPeptideFinder

Search for sequences for which the mass of the

precursor ion matches the mass of thetryptic peptide.

Store significantcandidates in theresult database

Foundresult

As process 2 butallow for sequenceerrors within the

database.

Foundresult

Combine sequences andcut out putative introns.compare the mass of

the resulting tryptic peptideto the mass of the

precursor ion.

Foundresult

As process 2 butallow for errors in thequery correct these

with the sequences fromthe database.

Process 1 Process 2 Process 3 Process 4

Yes Yes Yes Yes

First (stringent)sequence comparison

Second (less stringent)sequence comparison

RF1: DFRYPGSMGQQYFLGLEAIFKGSGDAAYPGEQIEERARERGPRLQLVWAKAGGDSQQGSGGGAQGRGRRRRCRVRDAGC_RF2: TSATPAPWASSTSWAWRRSSRALATLPTPVSR_RRERAREGRGCSWCGPRLAETASRGQGEGHRAEADAGAVGCGMRVADRF3: LPLPRLHGPAVLPGPGGDLQGLWRRCLPR_ADRGESARERAAAAAGVGQGWRRQPAGVRGRGTGQRQTPALSGAGCGLLRF4: QAAVLDLLELQLLHGSLGLAQVEQVEEGAACGAGEGTVICAVSQEHANPVEQLHSTKPSAADRSFHPSATRIPHPTAPASASRF5: RRPFLISLSFSSFMAASVLPRLNRLKKGPPAERGKGPSYALSARSMQTLLSSYIARNRPPRIGVSIHQQPASRTRQRRRLPLPCARF6: GGRS_SP_ASAPSWQPRSCPG_TG_RRGRLRSGGRDRHMRCQPGACKPC_AAT_HETVRRGSEFPSISNPHPAPDSAGVCLCPV

RF1: WMETPIRGGRFRAM_LLNRVCMLLADSAYDGPFPRSAGGPFFNLFNLGKTEAAMKELKLKEIKNGRLRF2: GWKLRSAADGFVLCSCSTGFACSWLTAHMTVPSPAPQAAPSSTCSTWARPRLP_RS_SSRRSRTAARF3: MDGNSDPRRTVSCYVAAQQGLHAPG_QRI_RSLPPLRRRPLLQPVQPGQDRGCHEGAEAQGDQERPPRF4: ALCPSP_PLLAVSASLGPHQLQPRPSLARSLLYLLTGVGSVARALEDRLQAQEVLLAHGAGVAEVRF5: PLPLTPAGCLRQPWPTPAAAAALSRALSPLSAHRGRQRRQSP_RSPPGPGSTAGPWSRGSGSRF6: DLRVLQPLGLDEAAAVHGVQQAQGVEAEVLFGMTKNGQ-IALMLDRGGMSGHADFIAERL

[RC*] AAYPG[VV]CFNPYNLGK 2027.95 nsi78_03.2460.2463.2

Measured Mass: 2027.95 Da

Calculated Mass: 2027.98 Da

Mass Difference: 0.03 Da < 700 ppm (~ 1.4 Da)

AAYPG

AAYPG RKGSGD EQIEER

GSGDAAYPGGPFFNLFNLGKEQIEER

GSGD EQIEER

Measured Mass ≠ Calculated MassNLGK

NLGK

GPFFNLF

Measured Mass ~ Calculated Mass

Found result

Display resultDeNovo-Sequencingof MS/MS data

RunGenomicPeptideFinder

Database search Found result

Select genomicdatabase

Set optionsRun query

Yes

Yes

MS/MSData

No

Selectsquencedatabase

Verification bydatabase search

MS/MS Data

4 * 1D Gels200 bands

Spectra extraction400 000 spectra

Extract significantResults

Run Sequest(database search)

8 databases

Database

Analysis Workflow: Analysis of thylakoidmembranes after 1D-SDS-PAGE

Allmer et al. (2006) Proteomics 6, 6207-6220.

Database

MS/MSData

Create dta filesfrom raw files2000 dta files

(3 min)

Extract significantResults(1 min)

Run Sequestseveral dbs

Variabel 1 min – 10 h(2h)

PEAKS15 dta files/ min

(2 h)

Create Queriesfor GPF~2000 (3 s)

Lutefisk

DeNovoX

GPFWindows local200 queries / h

(10 h)

GPFUNIX (LINIAC)

60 queries / P*h(3h @ 10 Processors)

Run Sequest on GPF-predictions21 dta files / min

(100 min)

Extract significantResults(1 min)

Total Processing about 14h per raw-file.

AutoMS

LC

MS/MS RawFile

MS Access

Database

PEAKS

Query Creation

GPF

Sequest

Result Extraction

Sequest

Dta-File Setup

Filters Result Extraction

Identified Peptides

47 (4)Complement

105 (17)60 (17)72 (5)PEAKS60 (17)448 (98)322 (19)GPF72 (5)322 (19)2496(22)Sequest

ComplementPEAKSGPFSequest

Confident peptide identifications and complementarities

Number of identified peptides (non redundant) is 2622

411292Comple-ment

1558332GPF

4711571317Sequest

%-missedmissedFound in gene modelsPeptides

Complement PEAKS-Sequest: 3%Improved by using GPF : 18%

Allmer et al. (2006) Proteomics 6, 6207-6220.

Peptides

ALE-GEIYDTFK, bnTY3_06302004.1322.1322.2.dta (Xcorr = 4.1)MEGEFTADNVAK, bnTY3_06302004.920.922.2.dta (Xcorr = 4.12)TGPPAVTVEDADK, bnTY3_06302004.742.746.2.dta (Xcorr = 3.26)TPIVFEGGDR, bnTY3_06302004.1003.1003.2.dta (Xcorr = 2.84)

Gene model: C_390061 [chlre2:164450]

MNRWNLLALTLGLLLVAAPFTKHQFAHASDEYEDDEEDDAPAAPKDDDVDVTVVTVKNWDETVKKSKFALVEFYAPWCGHCKTLKPEYAKAATALKAAAPDALIAKVDATQEESLAQKFGVQGYPTLKWFVDGELASDYNGPRD

ADGIVGWVKKKTGPPAVTVEDADKLKSLEADAEVVVVGYFKALEGEIYDTFKSYAAKTEDVVFVQTTSADVAKAAGLDAVDTVSVVKNFAGEDRATAVLATDIDTDSLTAFVKSEKMPPTIEFNQKNSDKIFNSGINKQLILWTTADDLKADAEIMTVFREASKKFKGQLVFVTVNNEGDGADPVTNFFGLKGATSPVLLGFFMEKNKKFRMEGEFTADNVAKFAESVVDGTAQAVLKSEAIPEDPYEDGVYKIVGKTVESVVLDETKDVLLEVYAPWCGHCKKLEPIYKKLAKRFKKVDSVIIAKMDGTENEHPEIEVKGFPTILFYPAGSDRTPIVFEGGDRSLKSLTKFIKTNAKIPYELPKKGSDGDEGTSDDKDKPASDKDEL* JGI annotation: Protein disulfide isomerase 1 (CrPDI1), encodes the chloroplastic Rb60 protein which is part of a complex involved in the control of psbA mRNA translation by light

NCBI: 0gi|4104541|gb|AAD02069.1|protein disulfide isomerase

SIB-BLAST: 0_CHLRE Protein disulfide isomerase RB60 (EC 5.3.4.1)

Kazusa: 0AF036939.1 Chlamydomonasreinhardtii protein disulfide isomerase

Other:

Scores

Example 1 – Gene model verification

Intron split and parts of intron split peptides are in red: ALE-GEIYDTFKSignificant non-intron split peptides are in blue: MEGEFTADNVAK

1

Scaff: 177403 SDGIVGWVKKKTGPPAVTVEDADKLKSLEADAEVVVVGYFKALE 177272+DGIVGWVKKKTGPPAVTVEDADKLKSLEADAEVVVVGYFKALE

sbjct: 145 ADGIVGWVKKKTGPPAVTVEDADKLKSLEADAEVVVVGYFKALE 188

Scaff: 176994 QGEIYDTFKSYAAKTEDVVFVQTTSADVAKAAGLDAVDTVSVVKNFAGE 176848+GEIYDTFKSYAAKTEDVVFVQTTSADVAKAAGLDAVDTVSVVKNFAGE

sbjct: 188 EGEIYDTFKSYAAKTEDVVFVQTTSADVAKAAGLDAVDTVSVVKNFAGE 236

JGI: tblastnblast protein

vs.translated nucleotide

Peptides

LEEGE-MPLNTYSNK, bnTY5a_062904.1079.1083.2.dta (Xcorr = 4.17)LYDEEL-QAIAK, bnTY5a_062904.1218.1222.2.dta (Xcorr = 4.14)

Gene model: C_1410005[chlre2:155574]

MQTVRAPAASGVATRVAGRRMCRPVAATKASTAVTTDMSKRTVPTKLEEGEVSACLGATTGRRKMGRGGERRSGSVGQELAGRILLICDFLFFILQMPLNTYSNKAPFKAKVRSVEKITGPKATGETCHIIIETEGKIPFWEGQSYGVIPPGTKINSKGKEVPHGTRLYSIASSRYGDDFDGQTASLCVRRAVYVDPETGKEDPAKKGLCSNFLCDATPGTEISMTGPTGKVLLLPADANAPLICVATGTGIAPFRSFWRRCFIENVPSYKFTGLFWLFMGVA

NSDAKLYDEELQAIAKAYPGQFRLDYALSREQNNRKGGKMYIQDKVEEYADEIFDLLDNGAHMYFCGLKGMMPGIQDMLERGLERNSGESSSEELLLSRTRQGGAGGGGTRTRRAAGRAGGMPRRKAKAPRAGRGTSMSGAV*

JGI annotation: Ferredoxin--NADP reductase, chloroplast precursor. The model doesn't completely fit the published cDNA sequence gi|1706781|sp|P53991|FENR_CHLRE[1706781],

NCBI: 9e-162gi|732532|gb|AAB40978.1|ferredoxin-NADP+ reductase

SIB-BLAST: 1e-162_VOLCA Ferredoxin-NADP+ reductase [fnr] [Volvox carteri]

Kazusa: 1e-170U10545.1 Chlamydomonasreinhardtii ferredoxin-NADP+ reductase

Other:

Scores

Example 2 - Gene model correction

cDNA sequence: FNR cDNA /FENR_CHLRE

MASLRKPSNHADRACSRRLRVATRVAGRRMCRPVAATKASTAVTTDMSKRTVPTKLEEGEMPLNTYSNKAPFKAKVRSVEKITGPKATGETCHIIIETEGKIPFWEGQSYGVIPPGTKINSKGKEVPTARLYSIASSRYGDDGDGQTASLCVRRAVYVDPETGKEDPAKKGLCSNFLCDATPGTEISMTGPTGKVLLLPADANAPLICVATG

TGIAPFRSFWRRCFIENVPSYKFTGLFWLFMGVGNSDAKLYDEELQAIAKAYPGQFRLDYALSREQNNRKGGKMYIQDKVEEYADEIFDLLDNGAHMYFCGLKGMMPGIQDMLERVAK EKGLNYEEWVEGLKHKNQWHVEVY

gi|1706781|sp|P53991|FENR_CHLRE[1706781]

NCBI: 0gi|1706781|sp|P53991|FENR_CHLRE Ferredoxin NADP

reductase

SIB-BLAST: 0FENR_CHLRE Ferredoxin--NADP reductase, chloroplast

Kazusa: 0U10545.1 Chlamydomonasreinhardtii ferredoxin-NADP+ reductase

Other:

Scores

1

2

Peptides

VNGGPAGEGLDPL-ADDPDTFAELK, bntyl27a_08182004.2232.2236.2.dta, Xcorr 3.64

Gene model: C_1460037 [chlre2:155818]

MAFALATSRKALQVTCKATGKKTAAKARIISYNQIEIAAAPKSSGVEFYGPNRAKWLGPYSENSTPAYLTGEFPGDYGWDTAGLSADPETFKRYRELELIHARWAMLGALGCLTPELLAKSGTKFGEAVWFKAGAQIFSEGGLD

YLGNPSLVHAQNIVATLAVQVILMGLIEGYRVNGGPAGEGLDPLYPGESFDPLGLADDPDTFAELKVKEIKNGRLAMFSMFGFFVQAIVTGKGPVQNLDDHLANPGVNNAFAFATKFTPSA

JGI annotation: Lhcbm4

Scores

NCBI: 2e-136gi|4139216|gb|AAD03731.1|light harvesting complex II

SIB-BLAST: 1e-136tr Q9ZSJ5 CHLRE [Lhcb2]

Kazusa: -

1

Gene model: C_1460037 [chlre2:155818]

MAFALATSRKALQVTCKATGKKTAAKARIISYNQIEIAAAPKSSGVEFYGPNRAKWLGPYSENSTPAYLTGEFPGDYGWDTAGLSADPETFKRYRELELIHARWAMLGALGCLTPELLAKSGTKFGEAVWFKAGAQIFSEGGLD

YLGNPSLVHAQNIVATLAVQVILMGLIEGYRVNGGPAGEGLDPLADDPDTFAELKVKEIKNGRLAMFSMFGFFVQAIVTGKGPVQNLDDHLANPGVNNAFAFATKFTPSA

Altered model is not supported by EST data

Scores

NCBI: 1e-126gi|4139216|gb|AAD03731.1|light harvesting complex II

SIB-BLAST: 1e-126tr Q9ZSJ5 _CHLRE [Lhcb2]

Kazusa: -

2

Example 3 – Alternative splicing?

GenomicPeptideFinder

• Detection of peptides in genomic databases– Increases yield 10 fold when compared to de novo sequencing alone

• Detection of peptides split by introns when deduced from genomic DNA– Enables identification of approximately 20% formerly undetectable

spectra• Gene Annotation

– Enables gene annotation from proteomic data• Additionally

– Detection of proteins by a single supporting peptide significantly enhances identification of

• Short proteins• Low abundant proteins• Modified proteins• Alternatively spliced proteins

– High-Throughput possibility

Allmer et al. (2004) FEBS Lett 562, 202-6.Allmer et al. (2006) Proteomics 6, 6207-6220.

Third generation GPF

• GPF 2: currently in development• improvements:

– also finds intron splits within nucleotide triplets– higher speed

• ~100 times faster than the old GPF• 3 queries per second on average• speedup due to the fact that an index is used and thus

peptides can be “looked up” instead of “searched for”

– includes visualization of hits• could be useful for genome annotation

• a web interface can be found atwww.uni-muenster.de/hippler/gpf

• use gpf as user name and hellofellow as password for the authentication

• GPF 2 is still work in progress, so the service may become temporarily unavailable or move to another location

Third generation GPF

2DB (relational database)

• Replaces PepProtDB– Web based– Multiuser– Fully searchable– Enhanced visualization

www.2db.de.ms

Experiment MS/MS Spectra-Analysis(e.g. SEQUEST)

Import, Filtering,Linkage

DataVisualization

Server

PHP

Local External

Internet

Workflow

Experiment IOverview

Experiment Details

List of all measured Spots / Bands

Spot/Band Labeling

Bild des Experimentsmit Imagemap

Experiment IIOverview

Experiment IIIOverview

MS - Details:• Identified Proteins • Peptide sequences• Link to other DBs• MS²-spectra • Software• Xcorr• ∆Cn• Tools (Plugins)

Spot/Band Details(pI, mass, rT)

Experiment Overview Experiment Details Spot/Band Details

Data export (*.csv)

Imange

ImageMap

Image

Application Structure

Conclusion

• 2DB speeds up data analysis– Automates protein identification– Provides results in Tables– Data organisation facilitates analysis

Acknowledgement

WESTFÄLISCHEWILHELMS-UNIVERSITÄT

MÜNSTER

Frederik Sommer

Christine Markert

Monica Zhang

Joshua Goldenberg

Jens AllmerAndreas BuschMarita Hermann

Susan HawatBernadeta KukuczkaSebastian KuhlgertBianca NaumannElisabeth Ostendorf

Michael SpechtMia Terashima

Michael Hippler

Einar J. Stauber

Blandine Rimbault

FRIEDRICH SCHILLERUNIVERSITÄT

JENA Institute of Plant Biochemistry and Biotechnology

Cooperations

David RoosUPENN, USA

FundingDeutsche ForschungsgemeinschaftResearch Grant UPENN