Antibody therapies have been employed to address some of today's most challenging diseases, but must meet many criteria during drug development before reaching a patient. Humanization is a sequence optimization strategy that addresses one critical risk called immunogenicity -- a patient's immune response to the drug -- by making an antibody more'human-like' in the absence of a predictive lab-based test for immunogenicity. However, existing humanization strategies generally yield very few humanized candidates, which may have degraded biophysical properties or decreased drug efficacy. Here, we re-frame humanization as a conditional generative modeling task, where humanizing mutations are sampled from a language model trained on human antibody data. We describe a sampling process that incorporates models of therapeutic attributes, such as antigen binding affinity, to obtain candidate sequences that have both reduced immunogenicity risk and maintained or improved therapeutic properties, allowing this algorithm to be readily embedded into an iterative antibody optimization campaign. We demonstrate in silico and in lab validation that in real therapeutic programs our generative humanization method produces diverse sets of antibodies that are both (1) highly-human and (2) have favorable therapeutic properties, such as improved binding to target antigens. Antibodies are the fastest growing drug class, with approved molecules treating a breadth of disorders ranging from cancer to autoimmune disease to infectious disease (Carter & Lazar, 2018). Many candidate therapeutic antibodies are derived from non-human e.g., murine or camelid sources, and modern antibody formats such as multi-specifics or antibody-drug conjugates can require heavy sequence engineering after discovery. This increases the risk of immunogenicity, where Anti-Drug Antibodies (ADAs) result in either fast clearance of the drug or adverse events (Hwang & Foote, 2005). While antibody sequence humanness is only roughly correlated with immunogenicity, humanization is widely employed to decrease immunogenicity risk (Prihoda et al., 2022).

The immune checkpoint inhibitors have demonstrated promising clinical efficacy across various tumor types, yet the percentage of patients who benefit from them remains low. The binding affinity between antigens and HLA-I/TCR molecules plays a critical role in antigen presentation and T-cell activation. Some computational methods have been developed to predict antigen-HLA or antigen-TCR binding specificity, but they focus solely on one task at a time. In this paper, we propose UnifyImmun, a unified cross-attention transformer model designed to simultaneously predicts the binding of antigens to both HLA and TCR molecules, thereby providing more comprehensive evaluation of antigen immunogenicity. We devise a two-phase progressive training strategy that enables these two tasks to mutually reinforce each other, by compelling the encoders to extract more expressive features. To further enhance the model generalizability, we incorporate virtual adversarial training. Compared to over ten existing methods for predicting antigen-HLA and antigen-TCR binding, our method demonstrates better performance in both tasks. Notably, on a large-scale COVID-19 antigen-TCR binding test set, our method improves performance by at least 9% compared to the current state-of-the-art methods. The validation experiments on three clinical cohorts confirm that our approach effectively predicts immunotherapy response and clinical outcomes. Furthermore, the cross-attention scores reveal the amino acids sites critical for antigen binding to receptors. In essence, our approach marks a significant step towards comprehensive evaluation of antigen immunogenicity.

We investigate the potential of patent data for improving the antibody humanness prediction using a multi-stage, multi-loss training process. Humanness serves as a proxy for the immunogenic response to antibody therapeutics, one of the major causes of attrition in drug discovery and a challenging obstacle for their use in clinical settings. We pose the initial learning stage as a weakly-supervised contrastive-learning problem, where each antibody sequence is associated with possibly multiple identifiers of function and the objective is to learn an encoder that groups them according to their patented properties. We then freeze a part of the contrastive encoder and continue training it on the patent data using the cross-entropy loss to predict the humanness score of a given antibody sequence. We illustrate the utility of the patent data and our approach by performing inference on three different immunogenicity datasets, unseen during training. Our empirical results demonstrate that the learned model consistently outperforms the alternative baselines and establishes new state-of-the-art on five out of six inference tasks, irrespective of the used metric.

Traditional antibody optimization approaches involve screening a small subset of the available sequence space, often resulting in drug candidates with suboptimal binding affinity, developability or immunogenicity. Based on two distinct antibodies, we demonstrate that deep contextual language models trained on high-throughput affinity data can quantitatively predict binding of unseen antibody sequence variants. These variants span a KD range of three orders of magnitude over a large mutational space. Our models reveal strong epistatic effects, which highlight the need for intelligent screening approaches. In addition, we introduce the modeling of "naturalness", a metric that scores antibody variants for similarity to natural immunoglobulins.

This project looks into the application of Machine Learning (ML) techniques in the prediction of Immunogenicity (Categorical; Positive or Negative) based on a peptide and its associated amino acid properties. This study uses peptide data from the Immune Epitode Database (IEDB). The R package "Peptides" has been used to compute the amino acid properties and mashup with peptide data to enable the use of ML algorithms for immunogenicity analysis, particularly, the algorithms that are more efficient with numeric and categorical data instead of string sequence. Tensorflow is an open source software library ML that provides linear regression and classification algorithms (open sourced by Google in Nov 2015) for multi-dimensional arrays (aka "Tensors"). K-fold cross-validation as well as hold-out of test data was used to train and test the generated models.