View
247
Download
0
Category
Tags:
Preview:
Citation preview
The Immunological Bioinformatics group
•Immunological Bioinformatics group, CBS, Technical University of Denmark (www.cbs.dtu.dk)
•Ole Lund, Group Leader
• Morten Nielsen, Associate Professor• Claus Lundegaard , Associate Professor• Jean Vennestrøm, post doc.• Thomas Blicher (50%), post doc.• Mette Voldby Larsen, PhD student• Pernille Haste Andersen, PhD student• Sune Frankild, PhD student• Sheila Tang, PhD student• Thomas Rask (50%), PhD student• Nicolas Rapin , PhD student• Ilka Hoff , PhD student•Jorid Sørli, PhD student• Hao Zhang, PhD student
•MSc students
•Collaborators•IMMI, University of Copenhagen• Søren Buus MHC binding• Mogens H Claesson Elispot Assay•La Jolla Institute of Allergy and Infectious Diseases• Allesandro Sette Epitope database• Bjoern Peters•Leiden University Medical Center• Tom Ottenhoff Tuberculosis• Michel Klein•Ganymed• Ugur Sahin Genetic library•University of Tubingen• Stefan Stevanovic MHC ligands•INSERM• Peter van Endert Tap binding•University of Mainz• Hansjörg Schild Proteasome•Schafer-Nielsen• Claus Schafer-Nielsen Peptide synthesis•ImmunoGrid• Elda Rossi Simulation of the• Vladimir Brusic Immune system•University of Utrectht• Can Kesmir Ideas
The innate immune system
• Unspecific• Antigen independent• Immediate response• No training/selection hence no memory
• Pathogen independent (but response might be pathogen type dependent)
The adaptive immune system
• Pathogen specific
– Humoral
– Cellular
http://www.uni-heidelberg.de/zentral/ztl/grafiken_bilder/bilder/e-coli.jpg
Bacteria
http://en.wikipedia.org/wiki/Image:Aids_virus.jpg
Virus
http://tpeeaupotable.ifrance.com/ma%20photo/bilharzoze.jpg
Parasite
Antibody - Antigen interaction
Fab
Antigen
Epitope
Paratope
Antibody
The antibody recognizes structural properties of the surface of the antigen
Prediction of HLA binding specificityHistorical overview• Simple Motifs
– Allowed/non allowed amino acids
• Extended motifs– Amino acid preferences (SYFPEITHI)– Anchor/Preferred/other amino acids
• Hidden Markov models– Peptide statistics from sequence alignment
• SVMs and neural networks– Can take sequence correlations into account
SLLPAIVEL YLLPAIVHI TLWVDPYEV GLVPFLVSV KLLEPVLLL LLDVPTAAV LLDVPTAAV LLDVPTAAVLLDVPTAAV VLFRGGPRG MVDGTLLLL YMNGTMSQV MLLSVPLLL SLLGLLVEV ALLPPINIL TLIKIQHTLHLIDYLVTS ILAPPVVKL ALFPQLVIL GILGFVFTL STNRQSGRQ GLDVLTAKV RILGAVAKV QVCERIPTIILFGHENRV ILMEHIHKL ILDQKINEV SLAGGIIGV LLIENVASL FLLWATAEA SLPDFGISY KKREEAPSLLERPGGNEI ALSNLEVKL ALNELLQHV DLERKVESL FLGENISNF ALSDHHIYL GLSEFTEYL STAPPAHGVPLDGEYFTL GVLVGVALI RTLDKVLEV HLSTAFARV RLDSYVRSL YMNGTMSQV GILGFVFTL ILKEPVHGVILGFVFTLT LLFGYPVYV GLSPTVWLS WLSLLVPFV FLPSDFFPS CLGGLLTMV FIAGNSAYE KLGEFYNQMKLVALGINA DLMGYIPLV RLVTLKDIV MLLAVLYCL AAGIGILTV YLEPGPVTA LLDGTATLR ITDQVPFSVKTWGQYWQV TITDQVPFS AFHHVAREL YLNKIQNSL MMRKLAILS AIMDKNIIL IMDKNIILK SMVGNWAKVSLLAPGAKQ KIFGSLAFL ELVSEFSRM KLTPLCVTL VLYRYGSFS YIGEVLVSV CINGVCWTV VMNILLQYVILTVILGVL KVLEYVIKV FLWGPRALV GLSRYVARL FLLTRILTI HLGNVKYLV GIAGGLALL GLQDCTMLVTGAPVTYST VIYQYMDDL VLPDVFIRC VLPDVFIRC AVGIGIAVV LVVLGLLAV ALGLGLLPV GIGIGVLAAGAGIGVAVL IAGIGILAI LIVIGILIL LAGIGLIAA VDGIGILTI GAGIGVLTA AAGIGIIQI QAGIGILLAKARDPHSGH KACDPHSGH ACDPHSGHF SLYNTVATL RGPGRAFVT NLVPMVATV GLHCYEQLV PLKQHFQIVAVFDRKSDA LLDFVRFMG VLVKSPNHV GLAPPQHLI LLGRNSFEV PLTFGWCYK VLEWRFDSR TLNAWVKVVGLCTLVAML FIDSYICQV IISAVVGIL VMAGVGSPY LLWTLVVLL SVRDRLARL LLMDCSGSI CLTSTVQLVVLHDDLLEA LMWITQCFL SLLMWITQC QLSLLMWIT LLGATCMFV RLTRFLSRV YMDGTMSQV FLTPKKLQCISNDVCAQV VKTDGNPPE SVYDFFVWL FLYGALLLA VLFSSDFRI LMWAKIGPV SLLLELEEV SLSRFSWGAYTAFTIPSI RLMKQDFSV RLPRIFCSC FLWGPRAYA RLLQETELV SLFEGIDFY SLDQSVVEL RLNMFTPYINMFTPYIGV LMIIPLINV TLFIGSHVV SLVIVTTFV VLQWASLAV ILAKFLHWL STAPPHVNV LLLLTVLTVVVLGVVFGI ILHNGAYSL MIMVKCWMI MLGTHTMEV MLGTHTMEV SLADTNSLA LLWAARPRL GVALQTMKQGLYDGMEHL KMVELVHFL YLQLVFGIE MLMAQEALA LMAQEALAF VYDGREHTV YLSGANLNL RMFPNAPYLEAAGIGILT TLDSQVMSL STPPPGTRV KVAELVHFL IMIGVLVGV ALCRWGLLL LLFAGVQCQ VLLCESTAVYLSTAFARV YLLEMLWRL SLDDYNHLV RTLDKVLEV GLPVEYLQV KLIANNTRV FIYAGSLSA KLVANNTRLFLDEFMEGV ALQPGTALL VLDGLDVLL SLYSFPEPE ALYVDSLFF SLLQHLIGL ELTLGEFLK MINAYLDKLAAGIGILTV FLPSDFFPS SVRDRLARL SLREWLLRI LLSAWILTA AAGIGILTV AVPDEIPPL FAYDGKDYIAAGIGILTV FLPSDFFPS AAGIGILTV FLPSDFFPS AAGIGILTV FLWGPRALV ETVSEQSNV ITLWQRPLV
Sequence information
• Score sequences to weight matrix by looking up and adding L values from the matrix
A R N D C Q E G H I L K M F P S T W Y V 1 0.6 0.4 -3.5 -2.4 -0.4 -1.9 -2.7 0.3 -1.1 1.0 0.3 0.0 1.4 1.2 -2.7 1.4 -1.2 -2.0 1.1 0.7 2 -1.6 -6.6 -6.5 -5.4 -2.5 -4.0 -4.7 -3.7 -6.3 1.0 5.1 -3.7 3.1 -4.2 -4.3 -4.2 -0.2 -5.9 -3.8 0.4 3 0.2 -1.3 0.1 1.5 0.0 -1.8 -3.3 0.4 0.5 -1.0 0.3 -2.5 1.2 1.0 -0.1 -0.3 -0.5 3.4 1.6 0.0 4 -0.1 -0.1 -2.0 2.0 -1.6 0.5 0.8 2.0 -3.3 0.1 -1.7 -1.0 -2.2 -1.6 1.7 -0.6 -0.2 1.3 -6.8 -0.7 5 -1.6 -0.1 0.1 -2.2 -1.2 0.4 -0.5 1.9 1.2 -2.2 -0.5 -1.3 -2.2 1.7 1.2 -2.5 -0.1 1.7 1.5 1.0 6 -0.7 -1.4 -1.0 -2.3 1.1 -1.3 -1.4 -0.2 -1.0 1.8 0.8 -1.9 0.2 1.0 -0.4 -0.6 0.4 -0.5 -0.0 2.1 7 1.1 -3.8 -0.2 -1.3 1.3 -0.3 -1.3 -1.4 2.1 0.6 0.7 -5.0 1.1 0.9 1.3 -0.5 -0.9 2.9 -0.4 0.5 8 -2.2 1.0 -0.8 -2.9 -1.4 0.4 0.1 -0.4 0.2 -0.0 1.1 -0.5 -0.5 0.7 -0.3 0.8 0.8 -0.7 1.3 -1.1 9 -0.2 -3.5 -6.1 -4.5 0.7 -0.8 -2.5 -4.0 -2.6 0.9 2.8 -3.0 -1.8 -1.4 -6.2 -1.9 -1.6 -4.9 -1.6 4.5
Scoring a sequence to a weight matrix
RLLDDTPEVGLLGNVSTVALAKAAAAL
Which peptide is most likely to bind?Which peptide second?
11.9 14.7 4.3
84nM 23nM 309nM
Example from real life
• 10 peptides from MHCpep database
• Bind to the MHC complex
• Relevant for immune system recognition
• Estimate sequence motif and weight matrix
• Evaluate motif “correctness” on 528 peptides
l ALAKAAAAMl ALAKAAAANl ALAKAAAARl ALAKAAAATl ALAKAAAAVl GMNERPILTl GILGFVFTMl TLNAWVKVVl KLNEPVLLLl AVVPFIVSV
Higher order sequence correlations• Neural networks can learn higher order
correlations!– What does this mean?
S S => 0L S => 1S L => 1L L => 0
No linear function can learn this (XOR) pattern
Say that the peptide needs one and only one large amino acid in the positions P3 and P4 to fill the binding cleft
How would you formulate this to test if a peptide can bind?
Sequence encoding (continued)
• Sparse encodingV:0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1
L:0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0
V.L=0 (unrelated)
• Blosum encodingV: 0 -3 -3 -3 -1 -2 -2 -3 -3 3 1 -2 1 -1 -2 -2 0 -3 -1 4
L:-1 -2 -3 -4 -1 -2 -3 -4 -3 2 4 -2 2 0 -3 -2 -1 -2 -1 1
R:-1 5 0 -2 -3 1 0 -2 0 -3 -2 2 -1 -3 -2 -1 -1 -3 -2 -3
V.L = 0.88 (highly related)V.R = -0.08 (close to unrelated)
Network ensembles
• No one single network with a particular architecture and sequence encoding scheme, will constantly perform the best
• Also for Neural network predictions will enlightened despotism fail – For some peptides, BLOSUM encoding with a four neuron hidden layer can best predict the peptide/MHC binding, for other peptides a sparse encoded network with zero hidden neurons performs the best
– Wisdom of the Crowd• Never use just one neural network• Use Network ensembles
Evaluation of prediction accuracy
ENS: Ensemble of neural networks trained using sparse, Blosum, and hidden Markov model sequence encoding
NetMHC-3.0 update
• IEDB + more proprietary data• Higher accuracy for existing ANNs• More Human alleles• Non human alleles (Mice + Primates)• Prediction of 8mer binding peptides for some alleles
• Prediction of 10- and 11mer peptides for all alleles
• Outputs to spread sheet
Prediction of 10- and 11mers using 9mer prediction tools
• Approach:• For each peptide of length L create 6 pseudo peptides deleting a sliding window of L- 9 always keeping pos. 1,2,3, and 9
• Example:• MLPQWESNTL = MLPWESNTL• MLPQESNTL• MLPQWSNTL• MLPQWENTL• MLPQWESTL• MLPQWESNL
Prediction of 10- and 11mers using 9mer prediction tools
• Final prediction = average of the 6 log scores:
• (0.477+0.405+0.564+0.505+0.559+0.521)/6• = 0.505
• Affinity:• Exp(log(50000)*(1 - 0.505)) = 211.5 nM
Proteasome specificity
• Low polymorphism– Constitutive & Immuno-proteasome
• Evolutionary conserved• Stochastic and low specificity
– Only 70-80% of the cleavage sites are reproduced in repeated experiments
Proteasome specificity
• NetChop is one of the best available cleavage method– www.cbs.dtu.dk/services/NetChop-3.0
Predicting TAP affinity
9 meric peptides >9 meric
Peters et el., 2003. JI, 171: 1741.
ILRGTSFVYV-0.11 + 0.09 - 0.42 - 0.3 = -0.74
Integrating all three steps (protesaomal cleavage, TAP transport and MHC binding) should lead to improved identification of peptides capable of eliciting CTL responses
Integration?
Identifying CTL epitopes
1 EBN3_EBV YQAYSSWMY 2.56 1.00 0.03 0.34 0.99 0.02 0.01 0.75 0.94 0.92 2.97 0 2.802 EBN3_EBV QSDETATSH 2.22 0.01 0.28 0.88 0.04 0.83 0.51 0.30 0.11 0.99 -0.80 0 2.283 EBN3_EBV PVSPAVNQY 1.55 0.01 0.97 0.01 0.22 0.21 1.00 0.02 0.04 1.00 2.63 0 1.784 EBN3_EBV AYSSWMYSY 1.31 0.34 0.99 0.02 0.01 0.75 0.94 0.92 0.09 1.00 3.28 1 1.585 EBN3_EBV LAAGWPMGY 1.02 1.00 0.97 0.22 0.01 0.18 0.01 0.06 0.01 1.00 3.01 0 1.276 EBN3_EBV IVQSCNPRY 0.99 0.10 0.97 0.50 0.05 0.01 0.01 0.01 0.02 0.93 3.19 0 1.247 EBN3_EBV FLQRTDLSY 0.94 0.46 0.99 0.02 0.82 0.07 0.01 0.63 0.01 0.96 2.79 0 1.188 EBN3_EBV YTDHQTTPT 1.15 1.00 0.01 0.42 0.02 0.04 0.01 0.02 0.54 0.14 -0.87 0 1.129 EBN3_EBV GTDVVQHQL 0.96 0.01 0.02 0.03 0.99 1.00 0.02 0.46 0.30 1.00 0.53 0 1.09...
HLA affinity
Proteasomal cleavage
TAP affinity
Case II:Discovery of conserved Class I epitopes in Human Influenza Virus H1N1
Wang et al., Vaccine 2007
Influenza
• We selected the Influenza peptides with the top 15 combined scores with conservation p9 > 70% for each pf the 12 supertypes.
• 180 peptides selected
• 167 tested for binding and CTL response
• 89 (53%) of the influenza peptides tested have an affinity better than 500nM
Donors
•35 normal healthy blood donors
•35-65 years old
•Expected to have had influenza more than 3 times
•HLA typed by SBT for HLA A and B
ELISPOT assay
•Measure number of white blood cells that in vitro produce interferon- in response to a peptide
•A positive result means that the immune system have earlier reacted to the peptide (during a response of a vaccine/natural infection)
FLDVMESM
Two spots
FLDVMESM
FLDVMESMFLDVMESMFLDVMESM
FLDVMESM
Conservation of epitopes
• Number of 9mers 100% conserved:
• 10/12 conserved in Influenza A virus (A/Goose/Guangdong/1/96(H5N1))
• 11/12 conserved in Influenza A virus (A/chicken/Jilin/9/2004(H5N1))
EpiSelect
Genotype 1
Top Scoring Peptides Top Scoring Peptides
Genotype 2
Genotype 3
Genotype 4
Genotype 5
Genotype 6
Select peptide with maximal coverage
Select peptide with maximal coverage preferring uncovered strains
Select peptide with maximal coverage preferring
lowest covered strains
Repeat until the Repeat until the desired number of desired number of
peptides is peptides is selectedselected
HCV Results - B7
Genotype 1
Genotype 2
Genotype 3
Genotype 4
Genotype 5
Genotype 6
QPRGRRQPIQPRGRRQPI
PeptidePeptide Predicted Predicted affinity (nM)affinity (nM)
55
SPRGSRPSWSPRGSRPSW 4343
GenomeGenomeCoverageCoverage
55
44
DPRRRSRNLDPRRRSRNL** 336666
RARAVRAKLRARAVRAKL
PeptidesPeptides
66 33
TPAETTVRLTPAETTVRL** 3838 33
33
33
22
33
44
33
* Verified B7 supertype restricted CD8+ epitope in the Los Alamos HCV epitope database
Ongoing work
• Selection of epitopes covering host (HLA) and pathogen variability
• Selection of diagnostic peptides in TB• Predict cross reactivity (T and B cell)
– Applications in epitope prediction, autoimmune diseases, transplantation
• Virulence factor discovery by comparative genomics
• Function-antigenecity studies• Bioinformatics immune system simulation
Recommended