Immune Epitope Database and Analysis Resource (IEDB)

www.IEDB.org

Bjoern Peters

bpeters@liai.org

La Jolla Institute for Allergy and Immunology

Buenos Aires, Oct 31, 2012

Overview

1. Introduction to the IEDB

2. Application: 2009 Swine-origin influenza virus

3. IEDB 2012+

What is the IEDB? NIH-sponsored free online resource

1. Database: repository of all published

experimentally-derived epitope information • Infectious disease

• Allergy

• Autoimmunity

• Transplantion/alloantigens

– Over 14,000 curated articles and direct submissions

– Over 90,000 unique epitopes

– Over 500,000 assays

2. Analysis Resource: tools to predict and model

immune responses

IEDB

www.iedb.org

Literature curation Epitope discovery

contract submission

Data Sources and Structure

Assay-Centric Data Representation

• IEDB captures the actual experimental assays relating to

– T cell responses

– B cell responses

– MHC Ligand Elution

– MHC Binding

• This allows searching in a variety of different ways

– By Epitope

– By Epitope Source

– By Immune Response

– By Host Organism

– By Assay …

Example query: All TB epitopes

recognized by T cells restricted

by MHC class II in humans

Advanced query

Only MTB epitopes recognized in

chronically infected humans and

detectable without in vitro restimulation

13

14

15

16

17

IEDB applications

Meta-Analyses

Prediction tool

development

IEDB Analysis Resource • Epitope prediction tools

– Machine learning algorithms that generalize the data

contained in the IEDB to predict new epitopes

• MHC class I & II binding and processing

• B cell epitope predictions

• Epitope analysis tools

– Conservancy analysis

– Population coverage

– Homology mapping

– Cluster analysis

Epitope Analysis Tools:

Add value to epitope datasets

Conservation of swine flu (S-OIV)

epitopes as an example

20

Swine flu project

• Initiated in spring 2009

• High mortality estimates based on first affected

population (Mexico)

• Novel combination of Swine and human influenza

strains

• Lack of neutralizing antibodies

fear of a global deadly pandemic

Swine flu project

• Question: Are there targets of pre-existing immunity in S-OIV?

How different is the pandemic virus from recent seasonal flu

viruses for the immune system?

• Query IEDB for all epitopes from influenza A

• Assemble sequences from recently circulating influenza strains

(=in the past 20 years),

• Epitopes contained in recently circulating strains are likely

targets of pre-existing immune responses

• Examine conservation of epitopes with likely pre-existing

immunity in seasonal flu strains vs. pandemic flu

• Follow up experimentally

50 B cell epitopes from

recent seasonal

influenza strains

55 sequences of

pandemic influenza

(10 antigens in each)

What does X number of

conserved epitopes in S-OIV

mean? Comparison to

seasonal flu 2008

The Number of Epitopes Described in the

Literature in Pre-2007 Years and Conserved

in Specific Strains

Influenza Strains B Cell T Cell – CD8+ T Cell – CD4+

Seasonal 2008 16 68 87

2009/S-OIV 8 54 57

Analysis of Conservation in S-OIV of

Known (pre-existing) Influenza

Responses

Hypothesis: Significant levels of preexisting immunity might exist in

the general population against S-OIV.

Greenbaum et al, PNAS, 2009

Preexisting T Cell Immunity

Against S-OIV in the General

Population

Greenbaum et al, PNAS, 2009

Conclusions: Swine Flu epitope

conservation

• The conservancy tool predicted that pre-existing immunity

exists in the general population at the T cell (but less at

the B cell) level

• Experimentally measured T cell responses confirmed that

preexisting memory against S-OIV epitopes were similar

in magnitude compared to new seasonal influenza

Analysis tools available in the

IEDB-AR

• Conservancy analysis Analyze if epitopes are found

conserved across different protein sequences

• Population coverage Analyze how many T cell epitopes with

known HLA restriction will be recognized in a human population

based on HLA frequencies

• Homology mapping Analyze the structure of an epitope in its

source antigen based on homology mapping

• Cluster analysis Analyze how many epitopes in a set have

significant sequence homology

30

Summary IEDB Introduction

+ S-OIV meta-analysis

• The IEDB catalogs all experiments characterizing

epitopes

• Multiple query mechanisms allow definition of

custom epitope sets

• IEDB epitope data is used to develop prediction

algorithms and perform Meta-Analysis

• IEDB Analysis tools help to examine existing sets of

epitopes and gain new knowledge

• Without the IEDB such meta-analysis would cost

much more time and effort 31

IEDB 2012+

• First funding period 2004 – 2011

• Renewal awarded for 7 more years

Update on priorities in the second funding

period

– Populating the IEDB

– Query enhancements

– Reporting enhancements

Populating the IEDB

PubMed Epitope

References

Over 21 Million 171,639

Relevant

References

29,559

Infectious Disease Allergy

Autoimmunity Transplantation

Cancer HIV

Others

Curation

Query Automatic

Abstract scans

Domain Classification

18,104

We finally caught up!

Curation from now on

• Reduced effort allows to cut expense and

refocus on other areas

• Implementation of biweekly update process

– data will appear faster

– You will be able to rely on the IEDB as a source of

current in addition to historical information

The IEDB 2012+:

Hierarchical queries using

ontologies

Example: Hierarchical query tree for

proteins

Queries for non-peptidic molecules

and diseases

Finders require replacing IEDB controlled

vocabularies with ontology classes

• Where available, re-use existing ontologies

• As necessary, contribute to building ontologies

• Benefits:

– Increase consistency in data curation

– Avoid duplicates

– Improve documentation to external users

– Enhance search capabilities

The IEDB 2012+:

Aggregate reporting using Immunome

Browser

Problem: Existing ways of displaying

immune epitope data have limitations

41

Query: T cell

epitopes in TB

Solution: Immunome Browser

A web application, integrated into the IEDB

Maps epitopes onto antigens from a reference

genome – minimize redundancy, consistent use

of antigen names

42

Mapping epitopes onto antigens

43

M.bovine >gi|398979|A85B_MYCTU Antigen 85-B precursor

MTDVSRKIRAWGRRLMIGTAAAVVLPGLVGLAGGAATAGAFSRPGLPVEYLQVPSPSM

GRDIKVQFQSGGNNSPAVYLLDGLRAQDDYNGWDINTPAFEWYYQSGLSIVMPVGGQS

SFYSDWYSPACGKAGCQTYKWETFLTSELPQWLSANRAVKPTGSAAIGLSMAGSSAMI

LAAYHPQQFIYAGSLSALLDPSQGMGPSLIGLAMGDAGGYKAADMWGPSSDPAWERND

PTQQIPKLVANNTRLWVYCGNGTPNELGGANIPAEFLENFVRSSNLKFQDAYNAAGGH

NAVFNFPPNGTHSWEYWGAQLNAMKGDLQSSLGAG

Epitope #1: AEFLENFVRSSNLKFQDA from antigen “Antigen 85-B precursor“ of

M.bovis BCG strain

Epitope #2: VFNFPPNGTHSWEYWGAQ from antigen “alpha-antigen“ of

M.tuberculosis H37Rv strain

44

45

Identification of 'antigenic regions' in TB

46

IEDB Conclusions

• IEDB has established a versatile database

structure and curation processes to capture

diverse immune epitopes

• Literature curation is current in all categories

and new articles are typically curated within 4

weeks from entry in PubMed

• Focus of renewal period is on improving the

query and reporting mechanisms, coming online

in the next months

Acknowledgments La Jolla Institute for Allergy & Immunology

San Diego Supercomputer Center • Phil Bourne

• Julia Ponomarenko

Consultants • Laura Zarebski (Buenos Aires)

• David Nemazee (Scripps)

• Ralph Kubo (KKC)

• Chemical Entities of Biological Interest (ChEBI)

Science Applications International Corporation

CBS / UC team

La Jolla Institute for Allergy and Immunology

• non-profit research institute

• focused exclusively on immune system research

• 21 faculty (20 experimental, 1 bioinformatic)

• >100 postdoctoral employees

Always open positions for

bright + enthusiastic

students / postdocs /

visiting scholars