PeptX: using genetic algorithms to optimize peptides for MHC binding

APEX-pHLA: A novel method for accurate prediction of the binding between exogenous short peptides and HLA class I molecules

Human leukocyte antigen (HLA) molecules play critically significant role within the realm of immunotherapy due to their capacities to recognize and bind exogenous antigens such as peptides, subsequently delivering them to immune cells. Predicting the binding between peptides and HLA molecules (pHLA) can expedite the screening of immunogenic peptides and facilitate vaccine design. However, traditional experimental methods are time-consuming and inefficient. In this study, an efficient method based on deep learning was developed for predicting peptide-HLA binding, which treated peptide sequences as linguistic entities. It combined the architectures of textCNN and BiLSTM to create a deep neural network model called APEX-pHLA. This model operated without limitations related to HLA class I allele variants and peptide segment lengths, enabling efficient encoding of sequence features for both HLA and peptide segments. On the independent test set, the model achieved Accuracy, ROC_AUC, F1, and MCC is 0.9449, 0.9850, 0.9453, and 0.8899, respectively. Similarly, on an external test set, the results were 0.9803, 0.9574, 0.8835, and 0.7863, respectively. These findings outperformed fifteen methods previously reported in the literature. The accurate prediction capability of the APEX-pHLA model in peptide-HLA binding might provide valuable insights for future HLA vaccine design.

Physics-based generative model of curvature sensing peptides; distinguishing sensors from binders

Proteins can specifically bind to curved membranes through curvature-induced hydrophobic lipid packing defects. The chemical diversity among such curvature "sensors" challenges our understanding of how they differ from general membrane "binders" that bind without curvature selectivity. Here, we combine an evolutionary algorithm with coarse-grained molecular dynamics simulations (Evo-MD) to resolve the peptide sequences that optimally recognize the curvature of lipid membranes. We subsequently demonstrate how a synergy between Evo-MD and a neural network (NN) can enhance the identification and discovery of curvature sensing peptides and proteins. To this aim, we benchmark a physics-trained NN model against experimental data and show that we can correctly identify known sensors and binders. We illustrate that sensing and binding are phenomena that lie on the same thermodynamic continuum, with only subtle but explainable differences in membrane binding free energy, consistent with the serendipitous discovery of sensors.

Genetic Algorithms Applied to Thermodynamic Rational Design of Mimetic Antibodies Based on the GB1 Domain of Streptococcal Protein G: An Atomistic Simulation Study

The development of mimetic antibodies (MA) capable of combining the high affinity and selectivity of antibodies with the small size of the peptides has enormous potential for applications in current biotechnology. In this work, we demonstrate that in silico MA design is possible through genetic algorithms (GA) developed from shell scripts capable of combining software commonly used for atomistic simulation. Our results demonstrate that, using the GB1 domain of the streptococcal G protein as a model, it is possible to optimize the molecular recognition capacity of a large MA population in a few generations. In the first case, GA was able to generate 10 MA with binding free energy (BFE) less than the vascular endothelial cell growth factor conjugated with the fms-type tyrosine kinase receptor. In the second case, it generated 13 MA with BFE less than that of the hepatitis C-E2 viral envelope conjugate with the antibody. Through the GA developed in this work, we demonstrate the use of a new protocol, capable of guiding experimental methods for the design of bioactive peptides that can assist in the development of new therapeutic molecules.

Machine learning optimization of peptides for presentation by class II MHCs

T cells play a critical role in cellular immune responses to pathogens and cancer and can be activated and expanded by MHC-presented antigens contained in peptide vaccines. We present a machine learning method to optimize the presentation of peptides by class II MHCs by modifying their anchor residues. Our method first learns a model of peptide affinity for a class II MHC using an ensemble of deep residual networks, and then uses the model to propose anchor residue changes to improve peptide affinity. We use a high throughput yeast display assay to show that anchor residue optimization improves peptide binding. Supplementary information: Supplementary data are available at Bioinformatics online.

Large-Scale Structure-Based Prediction of Stable Peptide Binding to Class I HLAs Using Random Forests

Prediction of stable peptide binding to Class I HLAs is an important component for designing immunotherapies. While the best performing predictors are based on machine learning algorithms trained on peptide-HLA (pHLA) sequences, the use of structure for training predictors deserves further exploration. Given enough pHLA structures, a predictor based on the residue-residue interactions found in these structures has the potential to generalize for alleles with little or no experimental data. We have previously developed APE-Gen, a modeling approach able to produce pHLA structures in a scalable manner. In this work we use APE-Gen to model over 150,000 pHLA structures, the largest dataset of its kind, which were used to train a structure-based pan-allele model. We extract simple, homogenous features based on residue-residue distances between peptide and HLA, and build a random forest model for predicting stable pHLA binding. Our model achieves competitive AUROC values on leave-one-allele-out validation tests using significantly less data when compared to popular sequence-based methods. Additionally, our model offers an interpretation analysis that can reveal how the model composes the features to arrive at any given prediction. This interpretation analysis can be used to check if the model is in line with chemical intuition, and we showcase particular examples. Our work is a significant step toward using structure to achieve generalizable and more interpretable prediction for stable pHLA binding.

A comprehensive review and performance evaluation of bioinformatics tools for HLA class I peptide-binding prediction

Human leukocyte antigen class I (HLA-I) molecules are encoded by major histocompatibility complex (MHC) class I loci in humans. The binding and interaction between HLA-I molecules and intracellular peptides derived from a variety of proteolytic mechanisms play a crucial role in subsequent T-cell recognition of target cells and the specificity of the immune response. In this context, tools that predict the likelihood for a peptide to bind to specific HLA class I allotypes are important for selecting the most promising antigenic targets for immunotherapy. In this article, we comprehensively review a variety of currently available tools for predicting the binding of peptides to a selection of HLA-I allomorphs. Specifically, we compare their calculation methods for the prediction score, employed algorithms, evaluation strategies and software functionalities. In addition, we have evaluated the prediction performance of the reviewed tools based on an independent validation data set, containing 21 101 experimentally verified ligands across 19 HLA-I allotypes. The benchmarking results show that MixMHCpred 2.0.1 achieves the best performance for predicting peptides binding to most of the HLA-I allomorphs studied, while NetMHCpan 4.0 and NetMHCcons 1.1 outperform the other machine learning-based and consensus-based tools, respectively. Importantly, it should be noted that a peptide predicted with a higher binding score for a specific HLA allotype does not necessarily imply it will be immunogenic. That said, peptide-binding predictors are still very useful in that they can help to significantly reduce the large number of epitope candidates that need to be experimentally verified. Several other factors, including susceptibility to proteasome cleavage, peptide transport into the endoplasmic reticulum and T-cell receptor repertoire, also contribute to the immunogenicity of peptide antigens, and some of them can be considered by some predictors. Therefore, integrating features derived from these additional factors together with HLA-binding properties by using machine-learning algorithms may increase the prediction accuracy of immunogenic peptides. As such, we anticipate that this review and benchmarking survey will assist researchers in selecting appropriate prediction tools that best suit their purposes and provide useful guidelines for the development of improved antigen predictors in the future.

CD4 T Cell Determinants in West Nile Virus Disease and Asymptomatic Infection

West Nile (WN) virus infection of humans is frequently asymptomatic, but can also lead to WN fever or neuroinvasive disease. CD4 T cells and B cells are critical in the defense against WN virus, and neutralizing antibodies, which are directed against the viral glycoprotein E, are an accepted correlate of protection. For the efficient production of these antibodies, B cells interact directly with CD4 helper T cells that recognize peptides from E or the two other structural proteins (capsid-C and membrane-prM/M) of the virus. However, the specific protein sites yielding such helper epitopes remain unknown. Here, we explored the CD4 T cell response in humans after WN virus infection using a comprehensive library of overlapping peptides covering all three structural proteins. By measuring T cell responses in 29 individuals with either WN virus disease or asymptomatic infection, we showed that CD4 T cells focus on peptides in specific structural elements of C and at the exposed surface of the pre- and postfusion forms of the E protein. Our data indicate that these immunodominant epitopes are recognized in the context of multiple different HLA molecules. Furthermore, we observed that immunodominant antigen regions are structurally conserved and similarly targeted in other mosquito-borne flaviviruses, including dengue, yellow fever, and Zika viruses. Together, these findings indicate a strong impact of virion protein structure on epitope selection and antigenicity, which is an important issue to consider in future vaccine design.

How peptide/MHC presence affects the dynamics of the LC13 T-cell receptor

The interaction between T-cell receptors (TCRs) of T-cells and potentially immunogenic peptides presented by MHCs of antigen presenting cells is one of the most important mechanisms of the adaptive human immune system. A large number of structural simulations of the TCR/peptide/MHC system have been carried out. However, to date no study has investigated the differences of the dynamics between free TCRs and pMHC bound TCRs on a large scale. Here we present a study totalling 37 100 ns investigating the LC13 TCR in its free form as well as in complex with HLA-B*08:01 and different peptides. Our results show that the dynamics of the bound and unbound LC13 TCR differ significantly. This is reflected in (a) expected results such as an increased flexibility and increased solvent accessible surface of the CDRs of unbound TCR simulations but also in (b) less expected results such as lower CDR distances and compactness as well as alteration in the hydrogen bond network around CDR3α of unbound TCR simulations. Our study further emphasises the structural flexibility of TCRs and confirms the importance of the CDR3 loops for the adoption to MHC.

Computational Opportunities and Challenges in Finding Cyclic Peptide Modulators of Protein-Protein Interactions

Peptide cyclization can improve stability, conformational constraint, and compactness. However, apart from beta-turn structures, which are well incorporated into cyclic peptides (CPs), many primary peptide structures and functions are markedly altered by cyclization. Accordingly, to mimic linear peptide interfaces with cyclic peptides, it can be beneficial to screen combinatorial cyclic peptide libraries. Computational methods have been developed to screen CPs, but face a number of challenges. Here, we review methods to develop in silico computational libraries, and the potential for screening naturally occurring libraries of CPs. The simplest and most rapid computational pharmacophore methods that estimate peptide three-dimensional structures to be screened versus targets are relatively easy to implement, and while the constraint on structure imposed by cyclization makes them more effective than the same approaches with linear peptides, there are a large number of limiting assumptions. In contrast, full molecular dynamics simulations of cyclic peptide structures not only are costly to implement, but also require careful attention to interpretation, so that not only is the computation time rate limiting, but the interpretation time is also rate limiting due to the analysis of the typically complex underlying conformational space of CPs. A challenge for the field of computational cyclic peptide screening is to bridge this gap effectively. Natural compound libraries of short cyclic peptides, and short cyclized regions of proteins, encoded in the genomes of many organisms present a potential treasure trove of novel functionality which may be screened via combined computational and experimental screening approaches.

Avoiding False Positive Conclusions in Molecular Simulation: The Importance of Replicas

Molecular simulations are a computational technique used to investigate the dynamics of proteins and other molecules. The free energy landscape of these simulations is often rugged, and minor differences in the initial velocities, floating-point precision, or underlying hardware can cause identical simulations (replicas) to take different paths in the landscape. In this study we investigated the magnitude of these effects based on 310 000 ns of simulation time. We performed 100 identically parametrized replicas of 3000 ns each for a small 10 amino acid system as well as 100 identically parametrized replicas of 100 ns each for an 827 residue T-cell receptor/MHC system. Comparing randomly chosen subgroups within these replica sets, we estimated the reproducibility and reliability that can be achieved by a given number of replicas at a given simulation time. These results demonstrate that conclusions drawn from single simulations are often not reproducible and that conclusions drawn from multiple shorter replicas are more reliable than those from a single longer simulation. The actual number of replicas needed will always depend on the question asked and the level of reliability sought. On the basis of our data, it appears that a good rule of thumb is to perform a minimum of five to 10 replicas.

Structure-based Methods for Binding Mode and Binding Affinity Prediction for Peptide-MHC Complexes

Understanding the mechanisms involved in the activation of an immune response is essential to many fields in human health, including vaccine development and personalized cancer immunotherapy. A central step in the activation of the adaptive immune response is the recognition, by T-cell lymphocytes, of peptides displayed by a special type of receptor known as Major Histocompatibility Complex (MHC). Considering the key role of MHC receptors in T-cell activation, the computational prediction of peptide binding to MHC has been an important goal for many immunological applications. Sequence- based methods have become the gold standard for peptide-MHC binding affinity prediction, but structure-based methods are expected to provide more general predictions (i.e., predictions applicable to all types of MHC receptors). In addition, structural modeling of peptide-MHC complexes has the potential to uncover yet unknown drivers of T-cell activation, thus allowing for the development of better and safer therapies. In this review, we discuss the use of computational methods for the structural modeling of peptide-MHC complexes (i.e., binding mode prediction) and for the structure-based prediction of binding affinity.

Exploring peptide/MHC detachment processes using hierarchical natural move Monte Carlo

MOTIVATION

The binding between a peptide and a major histocompatibility complex (MHC) is one of the most important processes for the induction of an adaptive immune response. Many algorithms have been developed to predict peptide/MHC (pMHC) binding. However, no approach has yet been able to give structural insight into how peptides detach from the MHC.

RESULTS

In this study, we used a combination of coarse graining, hierarchical natural move Monte Carlo and stochastic conformational optimization to explore the detachment processes of 32 different peptides from HLA-A*02:01. We performed 100 independent repeats of each stochastic simulation and found that the presence of experimentally known anchor amino acids affects the detachment trajectories of our peptides. Comparison with experimental binding affinity data indicates the reliability of our approach (area under the receiver operating characteristic curve 0.85). We also compared to a 1000 ns molecular dynamics simulation of a non-binding peptide (AAAKTPVIV) and HLA-A*02:01. Even in this simulation, the longest published for pMHC, the peptide does not fully detach. Our approach is orders of magnitude faster and as such allows us to explore pMHC detachment processes in a way not possible with all-atom molecular dynamics simulations.

AVAILABILITY AND IMPLEMENTATION

The source code is freely available for download at http://www.cs.ox.ac.uk/mosaics/.

CONTACT

bernhard.knapp@stats.ox.ac.uk

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

T-Cell Receptor Binding Affects the Dynamics of the Peptide/MHC-I Complex

The recognition of peptide/MHC by T-cell receptors is one of the most important interactions in the adaptive immune system. A large number of computational studies have investigated the structural dynamics of this interaction. However, to date only limited attention has been paid to differences between the dynamics of peptide/MHC with the T-cell receptor bound and unbound. Here we present the first large-scale molecular dynamics simulation study of this type investigating HLA-B*08:01 in complex with the Epstein-Barr virus peptide FLRGRAYGL and all possible single-point mutations (n = 172). All of the simulations were performed with and without the LC 13 T-cell receptor for a simulation time of 100 ns, yielding 344 simulations and a total simulation time of 34 400 ns. Our study is 2 orders of magnitude larger than the average T-cell receptor/peptide/MHC molecular dynamics simulation study. This data set provides reliable insights into alterations of the peptide/MHC-I dynamics caused by the presence of the T-cell receptor. We found that simulations in the presence of T-cell receptors have more hydrogen bonds between the peptide and MHC, altered flexibility patterns in the MHC helices and the peptide, a lower MHC groove width range, and altered solvent-accessible surface areas. This indicates that without a T-cell receptor the MHC binding groove can open and close, while the presence of the T-cell receptor inhibits these breathing-like motions.

Rapid, Precise, and Reproducible Prediction of Peptide-MHC Binding Affinities from Molecular Dynamics That Correlate Well with Experiment

The presentation of potentially pathogenic peptides by major histocompatibility complex (MHC) molecules is one of the most important processes in adaptive immune defense. Prediction of peptide-MHC (pMHC) binding affinities is therefore a principal objective of theoretical immunology. Machine learning techniques achieve good results if substantial experimental training data are available. Approaches based on structural information become necessary if sufficiently similar training data are unavailable for a specific MHC allele, although they have often been deemed to lack accuracy. In this study, we use a free energy method to rank the binding affinities of 12 diverse peptides bound by a class I MHC molecule HLA-A*02:01. The method is based on enhanced sampling of molecular dynamics calculations in combination with a continuum solvent approximation and includes estimates of the configurational entropy based on either a one or a three trajectory protocol. It produces precise and reproducible free energy estimates which correlate well with experimental measurements. If the results are combined with an amino acid hydrophobicity scale, then an extremely good ranking of peptide binding affinities emerges. Our approach is rapid, robust, and applicable to a wide range of ligand-receptor interactions without further adjustment.

Computational approaches to developing short cyclic peptide modulators of protein-protein interactions

Cyclic peptides are a promising class of bioactive molecules potentially capable of modulating "difficult" targets, such as protein-protein interactions. Cyclic peptides have long been used as therapeutics derived from natural product derivatives, but remain an underexplored class of compounds from the perspective of rational drug design, possibly due to the known weaknesses of peptide drugs in general. While cyclic peptides are non"druglike" by the accepted empirical rules, their unique structure may lend itself to both membrane permeability and proteolytic resistance-the main barriers to oral delivery. The constrained shape of cyclic peptides also lends itself better to virtual screening approaches, and new tools and successes in this area have been recently noted. An increasing number of strategies are available, both to generate and screen cyclic peptide libraries, and best practises and current successes are described within. This chapter will describe various computational strategies for virtual screening cyclic peptides, along with known implementations and applications. We will explore the generation and screening of diverse combinatorial virtual libraries, incorporating a range of cyclization strategies and structural modifications. More advanced approaches covered include evolutionary algorithms designed to aid in screening large structural libraries, machine learning approaches, and harnessing bioinformatics resources to bias cyclic peptide virtual libraries towards known bioactive structures.

Large scale characterization of the LC13 TCR and HLA-B8 structural landscape in reaction to 172 altered peptide ligands: a molecular dynamics simulation study

The interplay between T cell receptors (TCRs) and peptides bound by major histocompatibility complexes (MHCs) is one of the most important interactions in the adaptive immune system. Several previous studies have computationally investigated their structural dynamics. On the basis of these simulations several structural and dynamical properties have been proposed as effectors of the immunogenicity. Here we present the results of a large scale Molecular Dynamics simulation study consisting of 100 ns simulations of 172 different complexes. These complexes consisted of all possible point mutations of the Epstein Barr Virus peptide FLRGRAYGL bound by HLA-B*08:01 and presented to the LC13 TCR. We compare the results of these 172 structural simulations with experimental immunogenicity data. We found that simulations with more immunogenic peptides and those with less immunogenic peptides are in fact highly similar and on average only minor differences in the hydrogen binding footprints, interface distances, and the relative orientation between the TCR chains are present. Thus our large scale data analysis shows that many previously suggested dynamical and structural properties of the TCR/peptide/MHC interface are unlikely to be conserved causal factors for peptide immunogenicity.

Specificities of human CD4+ T cell responses to an inactivated flavivirus vaccine and infection: correlation with structure and epitope prediction

Tick-borne encephalitis (TBE) virus is endemic in large parts of Europe and Central and Eastern Asia and causes more than 10,000 annual cases of neurological disease in humans. It is closely related to the mosquito-borne yellow fever, dengue, Japanese encephalitis, and West Nile viruses, and vaccination with an inactivated whole-virus vaccine can effectively prevent clinical disease. Neutralizing antibodies are directed to the viral envelope protein (E) and an accepted correlate of immunity. However, data on the specificities of CD4(+) T cells that recognize epitopes in the viral structural proteins and thus can provide direct help to the B cells producing E-specific antibodies are lacking. We therefore conducted a study on the CD4(+) T cell response against the virion proteins in vaccinated people in comparison to TBE patients. The data obtained with overlapping peptides in interleukin-2 (IL-2) enzyme-linked immunosorbent spot (ELISpot) assays were analyzed in relation to the three-dimensional structures of the capsid (C) and E proteins as well as to epitope predictions based on major histocompatibility complex (MHC) class II peptide affinities. In the C protein, peptides corresponding to two out of four alpha helices dominated the response in both vaccinees and patients, whereas in the E protein concordance of immunodominance was restricted to peptides of a single domain (domain III). Epitope predictions were much better for C than for E and were especially erroneous for the transmembrane regions. Our data provide evidence for a strong impact of protein structural features that influence peptide processing, contributing to the discrepancies observed between experimentally determined and computer-predicted CD4(+) T cell epitopes. Importance: Tick-borne encephalitis virus is endemic in large parts of Europe and Asia and causes more than 10,000 annual cases of neurological disease in humans. It is closely related to yellow fever, dengue, Japanese encephalitis, and West Nile viruses, and vaccination with an inactivated vaccine can effectively prevent disease. Both vaccination and natural infection induce the formation of antibodies to a viral surface protein that neutralize the infectivity of the virus and mediate protection. B lymphocytes synthesizing these antibodies require help from other lymphocytes (helper T cells) which recognize small peptides derived from proteins contained in the viral particle. Which of these peptides dominate immune responses to vaccination and infection, however, was unknown. In our study we demonstrate which parts of the proteins contribute most strongly to the helper T cell response, highlight specific weaknesses of currently available approaches for their prediction, and demonstrate similarities and differences between vaccination and infection.

Using a Genetic Algorithm and Markov Clustering on Protein–Protein Interaction Graphs

In this paper, a Genetic Algorithm is applied on the filter of the Enhanced Markov Clustering algorithm to optimize the selection of clusters having a high probability to represent protein complexes. The filter was applied on the results (obtained by experiments made on five different yeast datasets) of three different algorithms known for their efficiency on protein complex detection through protein interaction graphs. The results are compared with three popular clustering algorithms, proving the efficiency of the proposed method according to metrics such as successful prediction rate and geometrical accuracy.

Molecular evolution of peptide ligands with custom-tailored characteristics for targeting of glycostructures

As an advanced approach to identify suitable targeting molecules required for various diagnostic and therapeutic interventions, we developed a procedure to devise peptides with customizable features by an iterative computer-assisted optimization strategy. An evolutionary algorithm was utilized to breed peptides in silico and the “fitness” of peptides was determined in an appropriate laboratory in vitro assay. The influence of different evolutional parameters and mechanisms such as mutation rate, crossover probability, gaussian variation and fitness value scaling on the course of this artificial evolutional process was investigated. As a proof of concept peptidic ligands for a model target molecule, the cell surface glycolipid ganglioside G_M1, were identified. Consensus sequences describing local fitness optima were reached from diverse sets of L- and proteolytically stable D lead peptides. Ten rounds of evolutional optimization encompassing a total of just 4400 peptides lead to an increase in affinity of the peptides towards fluorescently labeled ganglioside G_M1 by a factor of 100 for L- and 400 for D-peptides.

A clever identification procedure is crucial when peptidic ligands for diagnostic and therapeutic techniques such as in vivo imaging or drug targeting are to be developed. Here, we present a propitious and versatile approach for the discovery of peptide sequences with custom features that is based on an iterative computer-assisted optimization process. The methodology smartly combines in silico evolution with in vitro testing to quickly obtain promising peptide ligand candidates with desired properties. To validate our method in a proof of concept we tried to identify peptide sequences that can bind to a glycosidic cell membrane component. We applied the evolution process by starting out with a small population of peptide lead sequences and achieved a constant increase in affinity between the peptide candidates and their target molecule with each generation. After 10 rounds and a total number of only 4400 peptides synthesized and tested, a more than 100fold improvement in target recognition could be achieved. Since all kinds of building blocks useable in chemical solid phase peptide synthesis can in principle be employed in this evolutionary optimization process, our method should prove a most versatile approach for the optimization of peptides, peptoids and peptomers towards a preset functionality.

Epitope-specificity of recombinant antibodies reveals promiscuous peptide-binding properties

Protein-peptide interactions are a common occurrence and essential for numerous cellular processes, and frequently explored in broad applications within biology, medicine, and proteomics. Therefore, understanding the molecular mechanism(s) of protein-peptide recognition, specificity, and binding interactions will be essential. In this study, we report the first detailed analysis of antibody-peptide interaction characteristics, by combining large-scale experimental peptide binding data with the structural analysis of eight human recombinant antibodies and numerous peptides, targeting tryptic mammalian and eukaryote proteomes. The results consistently revealed that promiscuous peptide-binding interactions, that is, both specific and degenerate binding, were exhibited by all antibodies, and the discovery was corroborated by orthogonal data, indicating that this might be a general phenomenon for low-affinity antibody-peptide interactions. The molecular mechanism for the degenerate peptide-binding specificity appeared to be executed through the use of 2-3 semi-conserved anchor residues in the C-terminal part of the peptides, in analogue to the mechanism utilized by the major histocompatibility complex-peptide complexes. In the long-term, this knowledge will be instrumental for advancing our fundamental understanding of protein-peptide interactions, as well as for designing, generating, and applying peptide specific antibodies, or peptide-binding proteins in general, in various biotechnical and medical applications.