iR+ Seminar Series: Theory and Applications of Immune Repertoires

The iReceptor Plus Seminar Series highlights the Theory and Applications of immune repertoires. The overall goal of the iReceptor Plus Consortium is to facilitate sharing and analysis of Adaptive Immune Receptor Repertoire Data. These immune repertoires are rapidly increasing in size and importance to all aspects of immunotherapy, and the Consortium is dedicated to improving analysis and curation tools that will allow the entire immunology community to maximize the benefit of these data for biomedical research and patient care.  In the seminar Series we will invite one senior and one early career scientist each month to present their most exciting and impactful research in this area.

iR+ Seminar Series - November 26, 2020

Keynote: Gunilla Karlsson Hedestam, Karolinska Institutet, Stockholm, Sweden

Title: Generation of a comprehensive database of rhesus and cynomolgus macaque IGH alleles

Abstract: We applied the immunoglobulin (IG) gene inference tool, IgDiscover, to define germline VDJ alleles in 45 macaques of different origins. Our analysis resulted in a comprehensive database comprising 1198 IGHV alleles, of which around 70% were not previously described. Haplotype analysis of the animals revealed a considerable level of structural variation in the IGH locus. This work will facilitate high-quality B cell studies in rhesus and cynomolgus macaques.

Early Career Scientist: Vanessa Mhanna, Sorbonne Universite, Paris, France

Title: Exhaustion of the regulatory T cell receptor repertoire instigates diabetes in NOD mice

Abstract: Non-obese diabetic (NOD) mice spontaneously develop autoimmune diabetes. We aimed to analyze their TCR repertoire to better understand NOD autoimmunity. We performed next-generation sequencing of TCRs from splenocytes of prediabetic NOD and normal B6 mice. We analyzed the repertoire of CD4+ effector T cells (Teffs), CD44low CD62Lhigh naïve regulatory T cells (nTregs) and CD44high CD62Llow activated/memory Tregs (amTregs). These latter are known to respond to self-antigens and to be involved in protection against autoimmune diseases.

NOD and B6 nTreg TCR β repertoires were very diverse and mostly composed of unexpanded clonotypes. In contrast, B6 amTregs contained frequent expanded clonotypes that were lost in NOD amTregs resulting in an increased diversity of their repertoire. This was also seen, albeit to a lesser extent, in NOD Teffs. These observations suggested that NOD mice had lost the amTreg clonotypes that could protect them from diabetes.

As IL-2 administration leads to Tregs expansion and activation, and correlatively to protection from diabetes occurrence, we investigated the effects of IL-2 on NOD TCR repertoires. IL-2 administration to NOD mice restored amTreg clonotype expansions and led to few and no clonotype expansions of nTreg and Teffs, respectively. Noteworthily, IL-2-expanded amTreg and nTreg clonotypes were markedly enriched for islet-antigen specific TCRs.

Altogether, our results establish a causal link between an IL-2-mediated impoverishment of Treg repertoires affecting self‑antigen-specific TCRs and the development of autoimmune disease.

iR+ Seminar Series - January 28, 2021

Keynote Speaker: Steven H. Kleinstein, Yale School of Medicine, New Haven, CT, USA

Title: Analysis of B cell antibody repertoires from next-generation sequencing (in infection, vaccination and autoimmunity)

Abstract: Next-generation sequencing (NGS) technologies have revolutionized our ability to carry out large-scale adaptive immune receptor repertoire sequencing (AIRR-Seq) experiments. AIRR-Seq is increasingly being applied to profile B cell receptor (BCR) repertoires and gain insights into immune responses in healthy individuals and those with a range of diseases. As NGS technologies improve, these experiments are producing ever larger datasets, with tens- to hundreds-of-millions of BCR sequences. Although promising, repertoire-scale data present fundamental challenges for analysis requiring the development of new techniques and the rethinking of existing methods that are not scalable to the large number of sequences being generated [1].

To address these challenges, we have developed computational tools and methods that we currently make available to the wider scientific community through the Immcantation tool suite. This includes: raw read processing, novel V gene allele detection, subject-specific germline genotype identification, B cell clone assignment, lineage tree construction and analysis, somatic mutation profiling and selection analysis. Along with the underlying computational methodology, this presentation will discuss applications of BCR repertoire sequencing and lineage analysis to infection (Lyme disease, COVID-19 and West Nile Virus), vaccination (Influenza), autoimmunity (Multiple sclerosis, Myasthenia Gravis) and allergy/asthma.

  1. Yaari G, Kleinstein SH. Practical guidelines for B-cell receptor repertoire sequencing analysis. Genome Med. 2015 Nov 20;7:121. doi: 10.1186/s13073-015-0243-2.

Early Career Scientist: Milena Pavlovic, University of Oslo, Oslo, Norway

Title: immuneML: an open-source ecosystem for machine learning analysis of adaptive immune receptor repertoire data

Abstract: Adaptive immune receptor repertoires (AIRR) are key targets for immunological and pharmacological research as they provide a DNA-sequence record of all past and ongoing adaptive immune responses in health, disease, infection and vaccination. The capacity of machine learning (ML) to learn complex discriminative sequence patterns has led to its increasing use for AIRR-based diagnostics and therapeutics discovery. Previous developments have however been highly heterogeneous in terms of technical solutions, domain assumptions and user-interaction options, hampering transparent comparative evaluation and the ability to explore and select ML methodology most appropriate for a given study.

immuneML addresses these previous concerns by covering all major steps in AIRR ML within an open-source ecosystem and online user interface: AIRR sequence data read-in and encoding, training ML models of antigen specificity or immune state prediction as well as model assessment and interpretation. We demonstrate the broad applicability of immuneML for AIRR ML research by (i) replicating inside immuneML a published large-scale study on AIRR-based immune state prediction, (ii) applying a novel ML method for AIRR-based antigen specificity prediction and (iii) showcasing how immuneML may be used for AIRR ML method benchmarking. immuneML promotes reproducibility, customizability and shareability by (i) providing infrastructure for sharing complete ML workflows and intermediate steps, (ii) useful default parameters and workflows that shield beginners from common ML mistakes, (iii) and a user-friendly design based on the Galaxy framework.

Lab website: Sandve Lab, Greiff Lab 

Past Seminars