Class II MHC binding

Gibbs ClusteringMassimo Andreatta, Morten Nielsen

CBS, Department of Systems biologyDTU, Denmark

Class II MHC binding• MHC class II binds peptides in the class II antigen presentation pathway• Binds peptides of length 9-18 (even whole proteins can bind!)• Binding cleft is open• Binding core is 9 aa

CENTER FO

R BIOLO

GICAL SEQ

CE ANALYSIS

Department of Systems BiologyTechnical University of Denmark 3

Conventional Gibbs samplingMHC class II binding

SLFIGLKGDIRESTVDGEEEVQLIAAVPGK VFRLKGGAPIKGVTF SFSCIAIGIITLYLG IDQVTIAGAKLRSLNWIQKETLVTFKNPHAKKQDV KMLLDNINTPEGIIP ELLEFHYYLSSKLNK LNKFISPKSVAGRFAESLHNPYPDYHWLRT

... DFAAQVDYPSTGLY

SLFIGLKGDIRESTVDGEEEVQLIAAVPGKVFRLKGGAPIKGVTFSFSCIAIGIITLYLGIDQVTIAGAKLRSLNWIQKETLVTFKNPHAKKQDVKMLLDNINTPEGIIPELLEFHYYLSSKLNKLNKFISPKSVAGRFAESLHNPYPDYHWLRT ... ... ... ...DFAAQVDYPSTGLY

CENTER FO

R BIOLO

GICAL SEQ

CE ANALYSIS

Gibbs sampling - sequence alignment

Why sampling? SLFIGLKGDIRESTVDGEEEVQLIAAVPGK VFRLKGGAPIKGVTF SFSCIAIGIITLYLG IDQVTIAGAKLRSLNWIQKETLVTFKNPHAKKQDV KMLLDNINTPEGIIP ELLEFHYYLSSKLNK LNKFISPKSVAGRFAESLHNPYPDYHWLRT

... DFAAQVDYPSTGLY

50 sequences 12 amino acids long

try all possible combinations with a 9-mer overlap

450 ~ 1030 possible combinations

...computationally unfeasible

CENTER FO

R BIOLO

GICAL SEQ

CE ANALYSIS

Gibbs sampling - sequence alignmentState transition

move to state +1

CENTER FO

R BIOLO

GICAL SEQ

CE ANALYSIS

move to state +1

Accept or reject the move?

Note that the probability of going to the new state depends on the previous state only

Gibbs sampling - sequence alignmentState transition

CENTER FO

R BIOLO

GICAL SEQ

CE ANALYSIS

move to state +1

Numerical example - 1

Accept move with Prob = 100%

CENTER FO

R BIOLO

GICAL SEQ

CE ANALYSIS

move to state +1

Numerical example - 2

Accept move with Prob = 63.8%

CENTER FO

R BIOLO

GICAL SEQ

CE ANALYSIS

What is the MC temperature?

iteration

it’s a scalar decreased during the simulation

CENTER FO

R BIOLO

GICAL SEQ

CE ANALYSIS

What is the MC temperature?

iteration

it’s a scalar decreased during the simulation E.g. same dE=-0.3 but at different temperatures

t1=0.4

t3=0.02

t2=0.1

CENTER FO

R BIOLO

GICAL SEQ

CE ANALYSIS

Move freely around states when the system is “warm”, then cool it off to force it into a state of high fitness

CENTER FO

R BIOLO

GICAL SEQ

CE ANALYSIS

Does it work?

HLA-DRB3*01:01

CENTER FO

R BIOLO

GICAL SEQ

CE ANALYSIS

It works

High Temperature Low Temperature

CENTER FO

R BIOLO

GICAL SEQ

CE ANALYSIS

Gibbs sampler. Prediction accuracy

CENTER FO

R BIOLO

GICAL SEQ

CE ANALYSIS

Next generation peptide-chip technology

CENTER FO

R BIOLO

GICAL SEQ

CE ANALYSIS

Identification of receptor binding motifs

CENTER FO

R BIOLO

GICAL SEQ

CE ANALYSIS

Data interpretationPeptide Signal

LMLSLFEQSLSCQAQ 9

QGTDATKSIIFEAER -12

RLEEAQAYLAAGQHD 10

EISELRTKVQEQQKQ 44

FAGAKKIFGSLAFLP 70

VRASSRVSGSFPEDS -7

CKAFFKRSIQGHNDY 86

CEGCKAFFKRSIQGH 100

RLSEADIRGFVAAVV -7

CENTER FO

R BIOLO

GICAL SEQ

CE ANALYSIS

Or with multiple specificities? Peptide Signal

LMLSLFEQSLSCQAQ 9

QGTDATKSIIFEAER -12

RLEEAQAYLAAGQHD 10

EISELRTKVQEQQKQ 44

FAGAKKIFGSLAFLP 70

VRASSRVSGSFPEDS -7

CKAFFKRSIQGHNDY 86

CEGCKAFFKRSIQGH 100

RLSEADIRGFVAAVV -7

CENTER FO

R BIOLO

GICAL SEQ

CE ANALYSIS

Gibbs clustering (multiple specificities)

--ELLEFHYYLSSKLNK----------LNKFISPKSVAGRFAESLHNPYPDYHWLRT-------NKVKSLRILNTRRKL-------MMGMFNMLSTVLGVS----AKSSPAYPSVLGQTI--------RHLIFCHSKKKCDELAAK-

----SLFIGLKGDIRESTV----DGEEEVQLIAAVPGK----------VFRLKGGAPIKGVTF---SFSCIAIGIITLYLG-------IDQVTIAGAKLRSLN--WIQKETLVTFKNPHAKKQDV-------KMLLDNINTPEGIIP

Cluster 2

Cluster 1SLFIGLKGDIRESTVDGEEEVQLIAAVPGKVFRLKGGAPIKGVTFSFSCIAIGIITLYLGIDQVTIAGAKLRSLNWIQKETLVTFKNPHAKKQDVKMLLDNINTPEGIIPELLEFHYYLSSKLNKLNKFISPKSVAGRFAESLHNPYPDYHWLRTNKVKSLRILNTRRKLMMGMFNMLSTVLGVSAKSSPAYPSVLGQTIRHLIFCHSKKKCDELAAK

Multiple motifs!

CENTER FO

R BIOLO

GICAL SEQ

CE ANALYSIS

The algorithm

SLFIGLKGDIRESTVDGEEEVQLIAAVPGKVFRLKGGAPIKGVTFSFSCIAIGIITLYLGIDQVTIAGAKLRSLN

WIQKETLVTFKNPHAKKQDVKMLLDNINTPEGIIPELLEFHYYLSSKLNKLNKFISPKSVAGRFAESLHNPYPDYHWLRTNKVKSLRILNTRRKLMMGMFNMLSTVLGVSAKSSPAYPSVLGQTI

RHLIFCHSKKKCDELAAK

1. List of peptides

CENTER FO

R BIOLO

GICAL SEQ

CE ANALYSIS

The algorithm

SLFIGLKGDIRESTVDGEEEVQLIAAVPGKVFRLKGGAPIKGVTF

SFSCIAIGIITLYLGIDQVTIAGAKLRSLN

WIQKETLVTFKNPHAKKQDV

KMLLDNINTPEGIIPELLEFHYYLSSKLNKLNKFISPKSVAGRFA

ESLHNPYPDYHWLRTNKVKSLRILNTRRKL

MMGMFNMLSTVLGVSAKSSPAYPSVLGQTI

RHLIFCHSKKKCDELAAK

1. List of peptides 2. create Nrandom groups

----IDQVTIAGAKLRSLN-WIQKETLVTFKNPHAKKQ

DV--ELLEFHYYLSSKLNK---

--MMGMFNMLSTVLGVS------AKSSPAYPSVLGQTI--

-SLFIGLKGDIRESTV-----SFSCIAIGIITLYLG

KMLLDNINTPEGIIP----LNKYVHGTWRSILP-----NKVKSLRILNTRRKL-

---ESLHNPYPDYHWLRT-RHLIFCHSKKKCDELAAK-----VFRLKGGAPIKGVTF--LNKFISPKSVAGRFA--DGEEEVQLIAAVPGK----

g1 g2 gN

3 Simple shift Remove peptide Phase shift

CENTER FO

R BIOLO

GICAL SEQ

CE ANALYSIS

The algorithm

RHLIFCHSKKKCDELAAK

----IDQVTIAGAKLRSLN-WIQKETLVTFKNPHAKKQDV--ELLEFHYYLSSKLNK-----MMGMFNMLSTVLGVS------AKSSPAYPSVLGQTI--

-SLFIGLKGDIRESTV-----SFSCIAIGIITLYLGKMLLDNINTPEGIIP----LNKYVHGTWRSILP-----NKVKSLRILNTRRKL-

g1 g2 gN

MMGMFNMLSTVLGVS

3 3. Finding the optimal configuration (the MC moves)

CENTER FO

R BIOLO

GICAL SEQ

CE ANALYSIS

The algorithm

RHLIFCHSKKKCDELAAK

----IDQVTIAGAKLRSLN-WIQKETLVTFKNPHAKKQDV--ELLEFHYYLSSKLNK------AKSSPAYPSVLGQTI--

g1 g2 gN

MMGMFNMLSTVLGVS

4. Random shift of the coreMMGMFNMLSTVLGVS

3. Finding the optimal configuration (the MC moves)

5. Score new core to log-odds

matrices6. Accept or reject move

CENTER FO

R BIOLO

GICAL SEQ

CE ANALYSIS

The algorithm

RHLIFCHSKKKCDELAAK

DV--ELLEFHYYLSSKLNK---

--MMGMFNMLSTVLGVS------AKSSPAYPSVLGQTI--

g1 g2 gN

MMGMFNMLSTVLGVS

CENTER FO

R BIOLO

GICAL SEQ

CE ANALYSIS

The algorithm

RHLIFCHSKKKCDELAAK

DV--ELLEFHYYLSSKLNK------AKSSPAYPSVLGQTI--

g1 g2 gN

MMGMFNMLSTVLGVS

4b. Random shift of the coreMMGMFNMLSTVLGVS

CENTER FO

R BIOLO

GICAL SEQ

CE ANALYSIS

The algorithm

RHLIFCHSKKKCDELAAK

g1 g2 gN

MMGMFNMLSTVLGVS

5b. Score new core to all log-odds

matrices

E1 E2E3

CENTER FO

R BIOLO

GICAL SEQ

CE ANALYSIS

The algorithm

RHLIFCHSKKKCDELAAK

-MMGMFNMLSTVLGVS---

g1 g2 gN

MMGMFNMLSTVLGVS

5b. Score new core to all log-odds

matrices

E1 E2E3

6b. Accept or reject move

CENTER FO

R BIOLO

GICAL SEQ

CE ANALYSIS

The scoring function

-SLFIGLKGDIRESTV---SFSCIAIGIITLYLG--KMLLDNINTPEGIIP----LNKYVHGTWRSILP-----NKVKSLRILNTRRKL-

P = Probability of accepting the

Avoid small specialized clusters

(s = 10)Maximize intra cluster similarity

whilst minimize inter cluster similarity

CENTER FO

R BIOLO

GICAL SEQ

CE ANALYSIS

A mixture of 500 9mer peptides. How many motifs?

CENTER FO

R BIOLO

GICAL SEQ

CE ANALYSIS

A mixture of 500 9mer peptides. How many motifs?

CENTER FO

R BIOLO

GICAL SEQ

CE ANALYSIS

Mixture of 100 binders for the two allelesATDKAAAAY A*0101EVDQTKIQY A*0101AETGSQGVY B*4402ITDITKYLY A*0101AEMKTDAAT B*4402FEIKSAKKF B*4402LSEMLNKEY A*0101GELDRWEKI B*4402LTDSSTLLV A*0101FTIDFKLKY A*0101TTTIKPVSY A*0101EEKAFSPEV B*4402AENLWVPVY B*4402

Two MHC class I alleles: HLA-A*0101 and HLA-B*4402

CENTER FO

R BIOLO

GICAL SEQ

CE ANALYSIS

A0101 B4402

CENTER FO

R BIOLO

GICAL SEQ

CE ANALYSIS

A0101 B4402

Resolved

CENTER FO

R BIOLO

GICAL SEQ

CE ANALYSIS

Five MHC class I allelesA0101

A0201 A0301 B0702 B4402

CENTER FO

R BIOLO

GICAL SEQ

CE ANALYSIS

Five MHC class I allelesA0101

0 1 76 1 02 4 0 0 955 87 5 1 093 2 19 0 20 6 0 98 3

A0201 A0301 B0702 B4402

HLA-B440294%

HLA-B070292%

HLA-A020189%

HLA-A010180%

HLA-A030197%

CENTER FO

R BIOLO

GICAL SEQ

CE ANALYSIS

The right number of specificities

CENTER FO

R BIOLO

GICAL SEQ

CE ANALYSIS

How many motifs?

CENTER FO

R BIOLO

GICAL SEQ

CE ANALYSIS

How many motifs?

CENTER FO

R BIOLO

GICAL SEQ

CE ANALYSIS

Adding in alignment (MHC class II)

HLA-DRB1*03:01HLA-DRB1*04:01

SLFIGLKGDIRESTVDGEEEVQLIAAVPGKVFRLKGGAPIKGVTFSFSCIAIGIITLYLGIDQVTIAGAKLRSLNWIQKETLVTFKNPHAKKQDVKMLLDNINTPEGIIPELLEFHYYLSSKLNKLNKFISPKSVAGRFAESLHNPYPDYHWLRTNKVKSLRILNTRRKLMMGMFNMLSTVLGVSAKSSPAYPSVLGQTIRHLIFCHSKKKCDELAAK

CENTER FO

R BIOLO

GICAL SEQ

CE ANALYSIS

Dealing with noisy data

• Experimental data often contain false positives

• Outliers do not match any recurrent motif

• Introduce a garbage bin to collect outliers

CENTER FO

R BIOLO

GICAL SEQ

CE ANALYSIS

• Introduce a garbage bin to collect outliersSLFIGLKGDIRESTVDGEEEVQLIAAVPGKVFRLKGGAPIKGVTFSFSCIAIGIITLYLGIDQVTIAGAKLRSLNWIQKETLVTFKNPHAKKQDVKMLLDNINTPEGIIPELLEFHYYLSSKLNKLNKFISPKSVAGRFAESLHNPYPDYHWLRTNKVKSLRILNTRRKLMMGMFNMLSTVLGVSAKSSPAYPSVLGQTIRHLIFCHSKKKCDELAAK

g1 g2 gn trash

Move to the trash cluster peptides that do not match any motif

CENTER FO

R BIOLO

GICAL SEQ

CE ANALYSIS

50 random sequences are added to the data set

200 binders to 3 MHC class I alleles

3 alleles HLA-A0101 HLA-B0702 HLA-B4001 Random g0 0 197 2 6 205g1 199 1 0 2 202g2 0 0 196 5 201trash 1 2 2 37 42

200 200 200 50

CENTER FO

R BIOLO

GICAL SEQ

CE ANALYSIS

50 random sequences are added to the data set

200 binders to 3 MHC class I alleles

3 alleles HLA-A0101 HLA-B0702 HLA-B4001 Random g0 0 197 2 6 205g1 199 1 0 2 202g2 0 0 196 5 201trash 1 2 2 37 42

200 200 200 50

DHHFTPQIINAFGWENAYSQTSYQYLI

ELPIVTPALADKNLIKCS

CENTER FO

R BIOLO

GICAL SEQ

CE ANALYSIS

CENTER FO

R BIOLO

GICAL SEQ

CE ANALYSIS

SH3 domains

- SH3 domains are small interaction modules abundantly found in eukaryotes

- They preferentially bind proline-rich regions (minimum consensus sequence is PxxP)

- Two main classes of ligands exist:

Surface representation of the Hck SH3 domain bound to an artificial high-affinity peptide ligand, in green. Schmidt et al. (2007)

class I+xPxP

class IIPxPx+

+: R or K: hydrophobicx: any amino acid

CENTER FO

R BIOLO

GICAL SEQ

CE ANALYSIS

SH3 domains - Gibbs clustering

- Large data set of 2,457 unique peptide sequences* binding to Src SH3 domain

- 12 amino acids long peptides, unaligned with respect to the binding core

WVTAPRSLPVLPGSWVVDISNVEDNYSGNRPLPGIWRSPIVRQLPSLPRALPVMPTNGPMAKSRPLPMVGLVVRPLPSPEGFGQYRGRMLPVIWGTRALPLPRAYEGIVPLLPIRNGAVNPVRVLPNIPLSV...

*Kim, T., et al. (2011) MUSI: an integrated system for identifying multiple specificity from large peptide or nucleic acid data sets, Nucleic Acids Res, 1-8.

CENTER FO

R BIOLO

GICAL SEQ

CE ANALYSIS

SH3 domains

Gibbs clustering - 1 cluster

class I+xPxP

class IIPxPx+

2,350 sequences(107 to trash)

CENTER FO

R BIOLO

GICAL SEQ

CE ANALYSIS

SH3 domains

Gibbs clustering - 2 clusters

class I+xPxP

class IIPxPx+ (63 to trash)

466 sequences

Class II

1,928 sequences

Class I

CENTER FO

R BIOLO

GICAL SEQ

CE ANALYSIS

SH3 domains

Gibbs clustering - 3 clusters

class I+xPxP

class IIPxPx+

1,404 sequences 560 sequences

(48 to trash)

445 sequences

CENTER FO

R BIOLO

GICAL SEQ

CE ANALYSIS

SH3 domains

class I+xPxPclass II

Optimal solution has two

specificities

CENTER FO

R BIOLO

GICAL SEQ

CE ANALYSIS

http://www.cbs.dtu.dk/services/GibbsCluster

CENTER FO

R BIOLO

GICAL SEQ

CE ANALYSIS

Acknowledgements

• Massimo Andreatta, PhD (for doing all the work)

• John Sidney (La Jolla Institute for Allergy and Immunology, California, USA) for the binding affinity validation of the predicted MHC class I outliers.

• The European Union 7th Framework Program FP7/2007-2013 [grant number 222773]

Class II MHC binding

Documents

Antigen Loading of MHC Class I Molecules in the Endocytic ... · to MHC class II molecules, occur in the endocytic system (8,9). MHC class II molecules, consisting of two transmem-brane

TOTAL PREDICTED MHC-I ASSOCIATED WITH POPULATION … · 5/8/2020 · resolved a few hundred high-conﬁdence MHC-I peptides. By weighing individual MHC allele SARS-CoV-2 binding

Adaptive immunity How T cells recognize antigen · MHC Class I MHC Class II Peptides binding class II molecules vary in length, are anchored in the middle, but are also always oriented

NetMHCpan 4.0: Improved peptide-MHC class I interaction ... · NetMHC pan 4.0: Improved peptide-MHC class I interaction predictions integrating eluted ligand and peptide binding affinity

MHC class II-dependent presentation€¦ · The MHC class II-like molecule HLA-DOcan subtly modulate the peptide repertoireloaded onto MHC class II molecules by HLA-DM. A new study

The Major Histocompatibility Complex, MHC€¦ · MHC class I Molecules •αchain +β-2 microglobulin –αchain: polymorphic, transmembrane •α1 and α2 domains form peptide-binding

Footprints of antigen processing boost MHC class II ... · MHC-II binding groove, unlike MHC class I, is open at both ends. This attribute facilitates peptide protrusion out of the

The MHC Motif Viewer: A Visualizationtheory.bio.uu.nl/pdf/Rapin.cpi10-0.pdfThe MHC Motif Viewer: A Visualization UNIT 18.17 Tool for MHC Binding Motifs Nicolas Rapin,1 Ilka Hoof,2,3

MHC binding and MHC polymorphism. MHC-I molecules present peptides on the surface of most cells

Improved methods for predicting peptide binding affinity ... · Major histocompatibility complex class II (MHC-II) molecules are found on the surface of antigen presenting cells where

Targeted disruption of the MHC class II Aa gene in C57BL/6 ...doc.rero.ch/record/290416/files/5-8-957.pdf · MHC class II, thymocyte maturation Abstract The MHC class II gene Aa was

MHC Class I Related Antigen-Processing Machinery

Quiz 2 Review - WordPress.com · Quiz 2 Review • Immunology Overview class II MHC molecule class I MHC molecule class II MHC genes class III MHC genes class I MHC genes MHC complex

MHC class II binding predictions NN-align …...Class II MHC binding •Binds peptides of length 9-18 (even whole proteins can bind!) •Binding cleft is open •Binding core is 9

p-MHC - Columbia University...MHC class I and II molecules have homologous domain organization, but different chain structure The Structure of MHC Molecules: MHC Class I • ~ 70 AA

MHC binding and MHC polymorphism

Christina Sylvester-Hvid, IMMI, Panum “Experimental and Theoretical description of Peptide-MHC binding” “Theoretical and Experimental description of Peptide-MHC

Chaperon-like function of MHC class II molecules …...Chaperon-like function of MHC class II molecules for misfolded proteins : A new mechanism for autoimmune diseases MHC mol.,ul.s

In silico design of MHC class I high binding affinity

Different MHC alleles confer different functional ... · I system, since the peptide terminii are not constrained by MHC class II • A chaperone that complexes with MHC class II