Technology & Science

Premium CSS Template

Our technologies are based on world leading methods within epitope predictions, which are public available and thus constantly assessed by the global scientific community. The technologies are the product of more than 15 years of research and more than 30 scientific publications in the worlds leading bioinformatics journals – of which many are co-authored by one or more members of the EIR scientific team.
 
The epitope prediction methods that makes up the base for the EIR science technologies have been assessed multiple times by independent research organization and stands as the best performing methods available at the moment. EIR sciences is committed to open innovation and all research conducted at EIR sciences will be published in peer reviewed journals, thus ensuring the continued development and quality of our technologies.

Get an overview of the EIR Sciences technology here, have a look at the poster presented at the Coral Gables Symposium or read more below by expanding the sections

  • Immunogenicity of foreign molecules

    Introduction of peptides of foreign origin into the human organism have an inherent risk of inducing an immune response, in which case the peptide will act as an antigen. The exact mechanisms of such an event are presently not elucidated to a degree where it is possible to correlate a given degree of difference between a human native peptide and foreign peptide with the chance of inducing an immune response. Several factors, however, are known or strongly suspected to contribute to the induction of an immune response:

    • Creation of antibodies that are able to bind specifically to the antigen must exist.
    • Activation of helper T cells (Th cells). Th cells are activated by recognition of foreign peptides in complex with MHC II molecules presented by antigen presenting cells (APC)
    • Considering peptides or vaccines the cytotoxic immune response induced by the cytotoxic T lymphocytes (CTL) is important. CTL epitopes are required to be recognized in complex with a receptor, the major histocompatibility complex class I molecule (MHC I)
     
    Regarding humoral immune responses obviously the possibility of creating antibodies that are able to bind specifically to the antigen must exist. The part of proteins and peptides that antibodies bind are designated epitopes and thus identification of potential antigen specific epitopes will be one parameter that very likely will be useful in immunogenicity testing. For in silico screening tests this will mean the use of B-cell epitope prediction systems.

    Even though it is possible to generate a T-cell independent antibody response, the strongest and most long-lived humoral responses will be induced in concert with helper T cells (Th cells) identified as CD4+ T cells. Th cells are activated by recognition of foreign peptides in complex with major histocompatibility complex class II molecules (MHC II) presented by antigen presenting cells (APC), and recognized peptides are designated Th epitopes. A prerequisite of being presented is the ability of whole or a part of the antigen to bind to the MHC II, and identification of potential MHC II binding sub peptides within the antigen will thus be another parameter most likely to be useful for classifying immunogenic risk.

    Finally, when considering peptides the cytotoxic immune response induced by the cytotoxic T lymphocytes (CTL) may be a third parameter to consider as peptides have been shown to be able to function as CTL epitopes. CTL epitopes are required to be recognized in complex with a receptor, the major histocompatibility complex class I molecule (MHC I), thus prediction of potential MHC I binding sub peptides within a peptide will be a third parameter to consider with respect to induction of peptide drug immunogenicity.

  • Epitope Prediction

    Different types of epitope predictions

    What constitute a profound immune response is still not completely understood, however it is well established that the immune system recognize specific parts of a foreign molecule known as epitopes. During the past 15 years, research has been focusing on describing and developing tools for fast epitope discovery and today several methods exists for investigating and predicting epitopes. In general epitopes are divided into three types when investigated and predicted (see next slides for detailed description):
    • Cytotoxic T lymphocytes (CTL epitopes). CTL epitopes can be investigated by predicting the interaction between peptides and the MHC class 1 molecule. 
    • Helper T cell epitopes (Th epitopes). Th epitopes can be discovered by predicting the binding of 15-mer peptides, derived from the protein of interest, to the MHC class two complex.
    • B-cell epitopes. B-cell epitopes are defined as areas of a protein capable of binding to one or more antibodies. Prediction tools can pinpoint where on a protein surface the antibody is likely to bind, however most parts of a protein can constitute a B-cell epitopes hence complicating the correct analysis of these epitopes
     

    CTL epitope prediction (MHC class I binding prediction)

    Background
    In the MHC class I pathway, peptides from endogenous antigens bound to class I MHCs are presented on the surface of all mammalian cells to the CTLs (CD8+ T cells). If a CTL receptor exists capable of binding to the MHC-peptide complex an immune response will be elicited and signals for apoptosis will be given to the infected cells. The MHC class I pathway is linked to the intracellular degradation of proteins and a precursor peptide is usually first generated by the proteasome, a large cytosomal protease complex. For further processing, the peptides must enter the endoplasmic reticulum (ER), by active transport mediated by the TAP transporter. During or after transport into the ER a potential epitope must bind to the MHC class I molecule in humans expressed by the human leukocyte antigen (HLA) –A –B, and –C genes, to be presented for CTLs on the cell surface. The binding to the MHC class I molecules is the most restrictive point in the pathway.
     
    Methods for prediction 
    Sequence based CTL epitope predictions have improved immensely in the last decade. From predictions of peptide binding to major histocompatibility complex molecules with moderate accuracy, limited allele coverage, and no good estimates of the other events in the anti-gen-processing pathway, the field has evolved significantly. Methods have now been developed that produce highly accurate binding predictions for many alleles. Moreover have so-called pan-specific methods been developed, which allow for prediction of peptide binding to MHCs expressed by HLA alleles characterized by limited or no peptide binding data. The NetMHCpan method used by EIR sciences (Hoof et al., 2009; Nielsen et al., 2007) thus allows prediction for all known HLA-A, -B, and –C alleles, as well as some non-human primate, mouse and pig HLA alleles.
     
    Validation
    In an attempt to perform a completely unbiased benchmark of different HLA binding prediction approaches, several groups have participated in a competition that has been held in connection with the ICANN 09 conference  (Lundegaard, Lund, & Nielsen, 2010; Zhang 2011). The binding to the HLA alleles HLA-A*01:01, HLA-A*02:01, and HLA-B*07:02 were to be predicted for a total of 177 10mer peptides and 265 9mers. The results of this competition placed NetMHC-3.2 and NetMHCpan-2.2 as the best performing methods on the benchmark set.
     
    Relevant publications
    • I. Hoof, B. Peters, J. Sidney, L. E. Pedersen, A. Sette, O. Lund, S. Buus, and M. Nielsen, ‘NetMHCpan, a method for MHC class I binding prediction beyond humans’, Immunogenetics, vol. 61, no. 1, pp. 1–13, Jan. 2009. (pubmed)
    • M. Nielsen, C. Lundegaard, T. Blicher, K. Lamberth, M. Harndahl, S. Justesen, G. Røder, B. Peters, A. Sette, O. Lund, and S. Buus, ‘NetMHCpan, a method for quantitative predictions of peptide binding to any HLA-A and -B locus protein of known sequence’, PLoS ONE, vol. 2, no. 8, p. e796, 2007  (pubmed)
    • T. Stranzl, M. V. Larsen, C. Lundegaard, and M. Nielsen, ‘NetCTLpan: pan-specific MHC class I pathway epitope predictions’, Immunogenetics, vol. 62, no. 6, pp. 357–368, Jun. 2010  (pubmed)
     

    Helper T-cell epitope prediction

    Background
    MHC II molecules sample peptides from the extracellular space, allowing the immune system to detect the presence of foreign microbes or proteins. Exogenous proteins normally enter the MHCII pathway by endocytosis and degradation in the endosome of antigen presenting cells. In the endosomes the protein is cleavage into 13-20mer peptides and then presented for the MHC class II complex. If the peptide binds the peptide-MHC complex will migrate to the cell surface and present the peptide for the Th-cells. A profound immune response is obtained upon recognition of the peptide-MHC complex by the Th-cell receptor (CD8).  Hence non-self peptides are recognized in concert between MHC class II binding and Th-cell receptor recognition. Little is known about the mechanism of which Th-celle receptor recognize peptide-MHC complexes and to date no prediction tool have been developed that predicts the interaction. Contrary, high-throughput technics have enables accurate prediction tools for MHC class II binding. (For further information please visit our website)
     
    MHC class II prediction
    Few methods other than the pioneering TEPITOPE/ProPred method  (Hammer et al., 1994) have been developed for MHC II peptide binding predictions. Despite recent progress in method development, the predictive performance for MHC II peptide binding remains significantly lower than what can be obtained for MHC I. One reason for this is that the MHC II molecule is open at both ends allowing binding of peptides extending out of the groove. The binding core of MHC II-bound peptides is therefore not known a priori and the binding motif is hence not readily discernible. Recent progress has been obtained by including the flanking residues in the predictions. All attempts to make ab initio predictions based on protein structure have failed to reach predictive performances similar to those that can be obtained by data-driven methods. An MHC II peptide binding prediction algorithm aiming at dealing with these challenges is the NetMHCIIpan method used by EIR sciences  (Nielsen et al., 2010). The method is a pan-specific version of the earlier published allele-specific NN-align algorithm  (Nielsen & Lund, 2009) and does not require any pre-alignment of the input data. This allows the method to benefit also from information from alleles covered by limited binding data.
     
    Validation of MHC class II binding algorithms
    As only a few methods exists for MHC class II prediction no external benchmark have been conducted in recent years. However, in the publication related to netMHCIIpan the method is was proven significantly better than the NN-align (Nielsen et al., 2009 (pubmed)) and TEPITOPE (Stumiole et al., 1999 (pubmed)).
     
    Relevant publications
    • M. Nielsen, S. Justesen, O. Lund, C. Lundegaard, and S. Buus, ‘NetMHCIIpan-2.0 - Improved pan-specific HLA-DR predictions using a novel concurrent alignment and weight optimization training procedure’, Immunome Res, vol. 6, p. 9, 2010 (pubmed)
    • M. Nielsen and O. Lund, ‘NN-align. An artificial neural network-based alignment algorithm for MHC class II peptide binding prediction’, BMC Bioinformatics, vol. 10, p. 296, 2009 (pubmed)
     

    B-cell epitope prediction

    Background
    The interaction between antibodies and antigens is one of the most important events of the immune system and is involved in clearing infectious organisms from the host. Antibodies bind to antigens at sites known as antigenic determinant regions, also referred to as B-cell epitopes. Identification of the location of B-cell epitopes on the antigen surface is essential in several biomedical application such as; rational vaccine design, disease diagnostic, immuno-therapeutics, and estimation of immunogenicity of protein and peptide derived pharmaceuticals.
     
    Methods for predicting B-cell epitopes
    To date, the performance of methods for in silico mapping of B-cell epitopes has been moderate, with methods using structural data and epitope amino acid composition showing the most promising results. The method used by EIR sciencs; DiscoTope is developed for predictions of structural epitopes and uses a combination of amino acid statistics, spatial information, and surface exposure. It is trained on a compiled data set of discontinuous epitopes from 76 X-ray structures of antibody/antigen protein complexes. By default DiscoTope takes a protein structure as input, however if the protein structure is unknown only the structure independent part, the propensity scores are used for prediction 
     
    Validation of B-cell epitope prediction
    No independent evaluation of structural epitope prediction tools has been conducted in recent years, however the publication related to the DiscoTope-2.0 method present a benchmark study of the DiscoTope-2.0 (Kringelum et al., 2012 (pubmed)), PEPITO (Sweredoski and Baldi 2008 (pubmed)), ElliPro (Ponomarenko et al., 2008 (pubmed)), SEPPA (Sun et al., 2009 (pubmed)), Epitopia (Rubinstein et al., 2009 (pubmed)), EPCES (Liang et al., 2009 (pubmed)) and EPSVR (Liang et al., 2010 (pubmed)) methods on an independent dataset of 52 antigen structures. On this dataset DiscoTope-2.0 showed significantly improved performance compared ElliPro, Epitopia and EPSVR, improved performance compared to SEPPA and EPCES and comparable results to the PEPITO method. The performance of B-cell epitopes prediction is however still significantly lower than performance of both CTL and Th-cell epitopes.
     
    Relevant publications
    •  P. H. Andersen, M. Nielsen, and O. Lund, ‘Prediction of residues in discontinuous B-cell epitopes using protein 3D structures’, Protein Sci., vol. 15, no. 11, pp. 2558–2567, Nov. 2006. (pubmed)
    • J. V. Kringelum, C. Lundegaard, O. Lund, and M. Nielsen, ‘Reliable B cell epitope predictions: impacts of method development and improved benchmarking’, PLoS Comput. Biol., vol. 8, no. 12, p. e1002829, Dec. 2012. (pubmed)
    • J. E. P. Larsen, O. Lund, and M. Nielsen, ‘Improved method for predicting linear B-cell epitopes’, Immunome Res, vol. 2, p. 2, 2006 (pubmed)

  • EIR::predTM [Epitope Immune Resolver]

    Technology description

    The EIR::predTM technology enables fast and efficient discovery of potential immunogenic regions in proteins. The technology is based on world leading methods within epitope predicting and protein pattern recognition. These methods target different parts of the non-self recognition mechanism that constitute the immune system - including the MHC class I & II pathways that recognizing foreign endogenous and exogenous peptides respectively and antibody recognition (B-cell epitopes) - thus providing a detailed overview of potential risk factors.  Refer to the sections CTL epitope prediction, Helper T-cell epitope prediction and B-cell epitope prediction for scientific description of the epitope prediction methods used.
     

    Method overview (click to read abstract)

  • EIR::evalTM [Evaluate Immunogenicity Risk]

    Technology description

    The EIR::evalTM technology provide relevant immunogenicity predictions which aid the decision process before, during and after clinical testing of novel or existing biopharmaceutical drugs. Using the EIR::pred technology CTL, helper T-cell and B-cell epitopes (see separate section) are predicted and a combined risk score are calculated. Depending on the objective of the study, the risk score can be customized by e.g. including variation in MHC molecules between ethnic clusters or focus on one type of epitope prediction. Other risk factor might also be included.

    Method overview (click to read abstract)

  • EIR::elimTM [Eliminate Immunogenicity Risk]

    Technology description

    The EIR::elimTM technology analyze the immunogenicity risk profile of a protein drug, peptide or antibody and suggest specific mutations that minimizes the risk of evoking an immune response. Using the EIR::eval technology the risk of evoking an immune response is evaluated and random mutations are introduced in areas of high risk to decrease the overall immune response. Using an iterative Monte Carlo like optimization process, potential important CTL, helper T-cell and B-cell epitopes can be removed and the immunogenicity profile tailored to meet the needs of our customers. Protein secondary structure predictions are included in the optimization process to preserve the protein structure and function

    Method overview (click to read abstract)

  • EIR::optiTM [Enhance Immunogenicity Response]

    Technology description

    The EIR::optiTM technology analyze the immunogenicity risk profile of a protein, peptide or antibody and suggest specific mutations that maximize the risk of evoking an immune response. Using the EIR::eval technology the risk of evoking an immune response is evaluated and random mutations are introduced in areas of low risk to increase the overall immune response. Using an iterative Monte Carlo like optimization process, potential important CTL, helper T-cell and B-cell epitopes are introduced and the immunogenicity profile tailored to meet the needs of our customers. Protein secondary structure predictions are included in the optimization process to preserve the protein structure and function    

    Method overview (click to read abstract)

  •  

     

  •