Development and use of machine learning algorithms in vaccine target selection.

Computer-aided discovery of vaccine targets has become a cornerstone of rational vaccine design. In this article, I discuss how Machine Learning (ML) can inform and guide key computational steps in rational vaccine design concerned with the identification of B and T cell epitopes and correlates of protection. I provide examples of ML models, as well as types of data and predictions for which they are built. I argue that interpretable ML has the potential to improve the identification of immunogens also as a tool for scientific discovery, by helping elucidate the molecular processes underlying vaccine-induced immune responses. I outline the limitations and challenges in terms of data availability and method development that need to be addressed to bridge the gap between advances in ML predictions and their translational application to vaccine design.

A transfer-learning approach to predict antigen immunogenicity and T-cell receptor specificity.

Antigen immunogenicity and the specificity of binding of T-cell receptors to antigens are key properties underlying effective immune responses. Here we propose diffRBM, an approach based on transfer learning and Restricted Boltzmann Machines, to build sequence-based predictive models of these properties. DiffRBM is designed to learn the distinctive patterns in amino-acid composition that, on the one hand, underlie the antigen's probability of triggering a response, and on the other hand the T-cell receptor's ability to bind to a given antigen. We show that the patterns learnt by diffRBM allow us to predict putative contact sites of the antigen-receptor complex. We also discriminate immunogenic and non-immunogenic antigens, antigen-specific and generic receptors, reaching performances that compare favorably to existing sequence-based predictors of antigen immunogenicity and T-cell receptor specificity.

Neoantigen quality predicts immunoediting in survivors of pancreatic cancer.

Cancer immunoediting1 is a hallmark of cancer2 that predicts that lymphocytes kill more immunogenic cancer cells to cause less immunogenic clones to dominate a population. Although proven in mice1,3, whether immunoediting occurs naturally in human cancers remains unclear. Here, to address this, we investigate how 70 human pancreatic cancers evolved over 10 years. We find that, despite having more time to accumulate mutations, rare long-term survivors of pancreatic cancer who have stronger T cell activity in primary tumours develop genetically less heterogeneous recurrent tumours with fewer immunogenic mutations (neoantigens). To quantify whether immunoediting underlies these observations, we infer that a neoantigen is immunogenic (high-quality) by two features-'non-selfness'  based on neoantigen similarity to known antigens4,5, and 'selfness'  based on the antigenic distance required for a neoantigen to differentially bind to the MHC or activate a T cell compared with its wild-type peptide. Using these features, we estimate cancer clone fitness as the aggregate cost of T cells recognizing high-quality neoantigens offset by gains from oncogenic mutations. With this model, we predict the clonal evolution of tumours to reveal that long-term survivors of pancreatic cancer develop recurrent tumours with fewer high-quality neoantigens. Thus, we submit evidence that that the human immune system naturally edits neoantigens. Furthermore, we present a model to predict how immune pressure induces cancer cell populations to evolve over time. More broadly, our results argue that the immune system fundamentally surveils host genetic changes to suppress cancer.

Probing T-cell response by sequence-based probabilistic modeling.

With the increasing ability to use high-throughput next-generation sequencing to quantify the diversity of the human T cell receptor (TCR) repertoire, the ability to use TCR sequences to infer antigen-specificity could greatly aid potential diagnostics and therapeutics. Here, we use a machine-learning approach known as Restricted Boltzmann Machine to develop a sequence-based inference approach to identify antigen-specific TCRs. Our approach combines probabilistic models of TCR sequences with clone abundance information to extract TCR sequence motifs central to an antigen-specific response. We use this model to identify patient personalized TCR motifs that respond to individual tumor and infectious disease antigens, and to accurately discriminate specific from non-specific responses. Furthermore, the hidden structure of the model results in an interpretable representation space where TCRs responding to the same antigen cluster, correctly discriminating the response of TCR to different viral epitopes. The model can be used to identify condition specific responding TCRs. We focus on the examples of TCRs reactive to candidate neoantigens and selected epitopes in experiments of stimulated TCR clone expansion.

RBM-MHC: A Semi-Supervised Machine-Learning Method for Sample-Specific Prediction of Antigen Presentation by HLA-I Alleles.

The recent increase of immunopeptidomics data, obtained by mass spectrometry or binding assays, opens up possibilities for investigating endogenous antigen presentation by the highly polymorphic human leukocyte antigen class I (HLA-I) protein. State-of-the-art methods predict with high accuracy presentation by HLA alleles that are well represented in databases at the time of release but have a poorer performance for rarer and less characterized alleles. Here, we introduce a method based on Restricted Boltzmann Machines (RBMs) for prediction of antigens presented on the Major Histocompatibility Complex (MHC) encoded by HLA genes-RBM-MHC. RBM-MHC can be trained on custom and newly available samples with no or a small amount of HLA annotations. RBM-MHC ensures improved predictions for rare alleles and matches state-of-the-art performance for well-characterized alleles while being less data demanding. RBM-MHC is shown to be a flexible and easily interpretable method that can be used as a predictor of cancer neoantigens and viral epitopes, as a tool for feature discovery, and to reconstruct peptide motifs presented on specific HLA molecules.

Systematic model reduction captures the dynamics of extrinsic noise in biochemical subnetworks.

We consider the general problem of describing the dynamics of subnetworks of larger biochemical reaction networks, e.g., protein interaction networks involving complex formation and dissociation reactions. We propose the use of model reduction strategies to understand the "extrinsic" sources of stochasticity arising from the rest of the network. Our approaches are based on subnetwork dynamical equations derived by projection methods and path integrals. The results provide a principled derivation of different components of the extrinsic noise that is observed experimentally in cellular biochemical reactions, over and above the intrinsic noise from the stochasticity of biochemical events in the subnetwork. We explore several intermediate approximations to assess systematically the relative importance of different extrinsic noise components, including initial transients, long-time plateaus, temporal correlations, multiplicative noise terms, and nonlinear noise propagation. The best approximations achieve excellent accuracy in quantitative tests on a simple protein network and on the epidermal growth factor receptor signaling network.

Direct coupling analysis of epistasis in allosteric materials.

In allosteric proteins, the binding of a ligand modifies function at a distant active site. Such allosteric pathways can be used as target for drug design, generating considerable interest in inferring them from sequence alignment data. Currently, different methods lead to conflicting results, in particular on the existence of long-range evolutionary couplings between distant amino-acids mediating allostery. Here we propose a resolution of this conundrum, by studying epistasis and its inference in models where an allosteric material is evolved in silico to perform a mechanical task. We find in our model the four types of epistasis (Synergistic, Sign, Antagonistic, Saturation), which can be both short or long-range and have a simple mechanical interpretation. We perform a Direct Coupling Analysis (DCA) and find that DCA predicts well the cost of point mutations but is a rather poor generative model. Strikingly, it can predict short-range epistasis but fails to capture long-range epistasis, in consistence with empirical findings. We propose that such failure is generic when function requires subparts to work in concert. We illustrate this idea with a simple model, which suggests that other methods may be better suited to capture long-range effects.