Building Representation Learning Models for Antibody Comprehension.

Antibodies are versatile proteins with both the capacity to bind a broad range of targets and a proven track record as some of the most successful therapeutics. However, the development of novel antibody therapeutics is a lengthy and costly process. It is challenging to predict the functional and biophysical properties of antibodies from their amino acid sequence alone, requiring numerous experiments for full characterization. Machine learning, specifically deep representation learning, has emerged as a family of methods that can complement wet lab approaches and accelerate the overall discovery and engineering process. Here, we review advances in antibody sequence representation learning, and how this has improved antibody structure prediction and facilitated antibody optimization. We discuss challenges in the development and implementation of such models, such as the lack of publicly available, well-curated antibody function data and highlight opportunities for improvement. These and future advances in machine learning for antibody sequences have the potential to increase the success rate in developing new therapeutics, resulting in broader access to transformative medicines and improved patient outcomes.

Becoming fluent in proteins.

Large language models have emerged as a new compass for navigating the complex landscapes of protein engineering. This issue of Cell Systems features ProGen2 and IgLM-two protein language models (PLMs) that use subtly different approaches to design proteins.

Structural modeling of antibody variable regions using deep learning-progress and perspectives on drug discovery.

AlphaFold2 has hallmarked a generational improvement in protein structure prediction. In particular, advances in antibody structure prediction have provided a highly translatable impact on drug discovery. Though AlphaFold2 laid the groundwork for all proteins, antibody-specific applications require adjustments tailored to these molecules, which has resulted in a handful of deep learning antibody structure predictors. Herein, we review the recent advances in antibody structure prediction and relate them to their role in advancing biologics discovery.

Deciphering the language of antibodies using self-supervised learning.

An individual's B cell receptor (BCR) repertoire encodes information about past immune responses and potential for future disease protection. Deciphering the information stored in BCR sequence datasets will transform our understanding of disease and enable discovery of novel diagnostics and antibody therapeutics. A key challenge of BCR sequence analysis is the prediction of BCR properties from their amino acid sequence alone. Here, we present an antibody-specific language model, Antibody-specific Bidirectional Encoder Representation from Transformers (AntiBERTa), which provides a contextualized representation of BCR sequences. Following pre-training, we show that AntiBERTa embeddings capture biologically relevant information, generalizable to a range of applications. As a case study, we fine-tune AntiBERTa to predict paratope positions from an antibody sequence, outperforming public tools across multiple metrics. To our knowledge, AntiBERTa is the deepest protein-family-specific language model, providing a rich representation of BCRs. AntiBERTa embeddings are primed for multiple downstream tasks and can improve our understanding of the language of antibodies.

AbDiver: a tool to explore the natural antibody landscape to aid therapeutic design.

MOTIVATION: Rational design of therapeutic antibodies can be improved by harnessing the natural sequence diversity of these molecules. Our understanding of the diversity of antibodies has recently been greatly facilitated through the deposition of hundreds of millions of human antibody sequences in next-generation sequencing (NGS) repositories. Contrasting a query therapeutic antibody sequence to naturally observed diversity in similar antibody sequences from NGS can provide a mutational roadmap for antibody engineers designing biotherapeutics. Because of the sheer scale of the antibody NGS datasets, performing queries across them is computationally challenging. RESULTS: To facilitate harnessing antibody NGS data, we developed AbDiver (http://naturalantibody.com/abdiver), a free portal allowing users to compare their query sequences to those observed in the natural repertoires. AbDiver offers three antibody-specific use-cases: (i) compare a query antibody to positional variability statistics precomputed from multiple independent studies, (ii) retrieve close full variable sequence matches to a query antibody and (iii) retrieve CDR3 or clonotype matches to a query antibody. We applied our system to a set of 742 therapeutic antibodies, demonstrating that for each use-case our system can retrieve relevant results for most sequences. AbDiver facilitates the navigation of vast antibody mutation space for the purpose of rational therapeutic antibody design. AVAILABILITY AND IMPLEMENTATION: AbDiver is freely accessible at http://naturalantibody.com/abdiver. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

TCRBuilder: multi-state T-cell receptor structure prediction.

MOTIVATION: T-cell receptors (TCRs) are immune proteins that primarily target peptide antigens presented by the major histocompatibility complex. They tend to have lower specificity and affinity than their antibody counterparts, and their binding sites have been shown to adopt multiple conformations, which is potentially an important factor for their polyspecificity. None of the current TCR-modelling tools predict this variability which limits our ability to accurately predict TCR binding. RESULTS: We present TCRBuilder, a multi-state TCR structure prediction tool. Given a paired αßTCR sequence, TCRBuilder returns a model or an ensemble of models covering the potential conformations of the binding site. This enables the analysis of structurally driven polyspecificity in TCRs, which is not possible with existing tools. AVAILABILITY AND IMPLEMENTATION: http://opig.stats.ox.ac.uk/resources. CONTACT: deane@stats.ox.ac.uk. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Comparative Analysis of the CDR Loops of Antigen Receptors.

The adaptive immune system uses two main types of antigen receptors: T-cell receptors (TCRs) and antibodies. While both proteins share a globally similar ß-sandwich architecture, TCRs are specialized to recognize peptide antigens in the binding groove of the major histocompatibility complex, while antibodies can bind an almost infinite range of molecules. For both proteins, the main determinants of target recognition are the complementarity-determining region (CDR) loops. Five of the six CDRs adopt a limited number of backbone conformations, known as the "canonical classes"; the remaining CDR (ß3in TCRs and H3 in antibodies) is more structurally diverse. In this paper, we first update the definition of canonical forms in TCRs, build an auto-updating sequence-based prediction tool (available at http://opig.stats.ox.ac.uk/resources) and demonstrate its application on large scale sequencing studies. Given the global similarity of TCRs and antibodies, we then examine the structural similarity of their CDRs. We find that TCR and antibody CDRs tend to have different length distributions, and where they have similar lengths, they mostly occupy distinct structural spaces. In the rare cases where we found structural similarity, the underlying sequence patterns for the TCR and antibody version are different. Finally, where multiple structures have been solved for the same CDR sequence, the structural variability in TCR loops is higher than that in antibodies, suggesting TCR CDRs are more flexible. These structural differences between TCR and antibody CDRs may be important to their different biological functions.

High-Throughput Antibody Structure Modeling and Design Using ABodyBuilder.

Antibodies are proteins of the adaptive immune system; they can be designed to bind almost any molecule, and are increasingly being used as biotherapeutics. Experimental antibody design is an expensive and time-consuming process, and computational antibody design methods can now be used to help develop new therapeutics and diagnostics. Within the design pipeline, accurate antibody structure modeling is essential, as it provides the basis for antibody-antigen docking, binding affinity prediction, and estimating thermal stability. Ideally, models should be rapidly generated, allowing the exploration of the breadth of antibody space. This allows methods to replicate the natural processes of antibody diversification (e.g., V(D)J recombination and somatic hypermutation), and cope with large volumes of data that are typical of next-generation sequencing datasets. Here we describe ABodyBuilder and PEARS, algorithms that build and mutate antibody model structures. These methods take ~30 s to generate a model antibody structure.

SCALOP: sequence-based antibody canonical loop structure annotation.

MOTIVATION: Canonical forms of the antibody complementarity-determining regions (CDRs) were first described in 1987 and have been redefined on multiple occasions since. The canonical forms are often used to approximate the antibody binding site shape as they can be predicted from sequence. A rapid predictor would facilitate the annotation of CDR structures in the large amounts of repertoire data now becoming available from next generation sequencing experiments. RESULTS: SCALOP annotates CDR canonical forms for antibody sequences, supported by an auto-updating database to capture the latest cluster information. Its accuracy is comparable to that of a standard structural predictor but it is 800 times faster. The auto-updating nature of SCALOP ensures that it always attains the best possible coverage. AVAILABILITY AND IMPLEMENTATION: SCALOP is available as a web application and for download under a GPLv3 license at opig.stats.ox.ac.uk/webapps/scalop. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Observed Antibody Space: A Resource for Data Mining Next-Generation Sequencing of Antibody Repertoires.

Abs are immune system proteins that recognize noxious molecules for elimination. Their sequence diversity and binding versatility have made Abs the primary class of biopharmaceuticals. Recently, it has become possible to query their immense natural diversity using next-generation sequencing of Ig gene repertoires (Ig-seq). However, Ig-seq outputs are currently fragmented across repositories and tend to be presented as raw nucleotide reads, which means nontrivial effort is required to reuse the data for analysis. To address this issue, we have collected Ig-seq outputs from 55 studies, covering more than half a billion Ab sequences across diverse immune states, organisms (primarily human and mouse), and individuals. We have sorted, cleaned, annotated, translated, and numbered these sequences and make the data available via our Observed Antibody Space (OAS) resource at http://antibodymap.org The data within OAS will be regularly updated with newly released Ig-seq datasets. We believe OAS will facilitate data mining of immune repertoires for improved understanding of the immune system and development of better biotherapeutics.

Structurally Mapping Antibody Repertoires.

Every human possesses millions of distinct antibodies. It is now possible to analyze this diversity via next-generation sequencing of immunoglobulin genes (Ig-seq). This technique produces large volume sequence snapshots of B-cell receptors that are indicative of the antibody repertoire. In this paper, we enrich these large-scale sequence datasets with structural information. Enriching a sequence with its structural data allows better approximation of many vital features, such as its binding site and specificity. Here, we describe the structural annotation of antibodies pipeline that maps the outputs of large Ig-seq experiments to known antibody structures. We demonstrate the viability of our protocol on five separate Ig-seq datasets covering ca. 35 m unique amino acid sequences from ca. 600 individuals. Despite the great theoretical diversity of antibodies, we find that the majority of sequences coming from such studies can be reliably mapped to an existing structure.

Antibody side chain conformations are position-dependent.

Side chain prediction is an integral component of computational antibody design and structure prediction. Current antibody modelling tools use backbone-dependent rotamer libraries with conformations taken from general proteins. Here we present our antibody-specific rotamer library, where rotamers are binned according to their immunogenetics (IMGT) position, rather than their local backbone geometry. We find that for some amino acid types at certain positions, only a restricted number of side chain conformations are ever observed. Using this information, we are able to reduce the breadth of the rotamer sampling space. Based on our rotamer library, we built a side chain predictor, position-dependent antibody rotamer swapper (PEARS). On a blind test set of 95 antibody model structures, PEARS had the highest average χ1 and χ1+2 accuracy (78.7% and 64.8%) compared to three leading backbone-dependent side chain predictors. Our use of IMGT position, rather than backbone ϕ/ψ, meant that PEARS was more robust to errors in the backbone of the model structure. PEARS also achieved the lowest number of side chain-side chain clashes. PEARS is freely available as a web application at http://opig.stats.ox.ac.uk/webapps/pears.

STCRDab: the structural T-cell receptor database.

The Structural T-cell Receptor Database (STCRDab; http://opig.stats.ox.ac.uk/webapps/stcrdab) is an online resource that automatically collects and curates TCR structural data from the Protein Data Bank. For each entry, the database provides annotations, such as the α/ß or γ/δ chain pairings, major histocompatibility complex details, and where available, antigen binding affinities. In addition, the orientation between the variable domains and the canonical forms of the complementarity-determining region loops are also provided. Users can select, view, and download individual or bulk sets of structures based on these criteria. Where available, STCRDab also finds antibody structures that are similar to TCRs, helping users explore the relationship between TCRs and antibodies.

ABodyBuilder: Automated antibody structure prediction with data-driven accuracy estimation.

Computational modeling of antibody structures plays a critical role in therapeutic antibody design. Several antibody modeling pipelines exist, but no freely available methods currently model nanobodies, provide estimates of expected model accuracy, or highlight potential issues with the antibody's experimental development. Here, we describe our automated antibody modeling pipeline, ABodyBuilder, designed to overcome these issues. The algorithm itself follows the standard 4 steps of template selection, orientation prediction, complementarity-determining region (CDR) loop modeling, and side chain prediction. ABodyBuilder then annotates the 'confidence' of the model as a probability that a component of the antibody (e.g., CDRL3 loop) will be modeled within a root-mean square deviation threshold. It also flags structural motifs on the model that are known to cause issues during in vitro development. ABodyBuilder was tested on 4 separate datasets, including the 11 antibodies from the Antibody Modeling Assessment-II competition. ABodyBuilder builds models that are of similar quality to other methodologies, with sub-Angstrom predictions for the 'canonical' CDR loops. Its ability to model nanobodies, and rapidly generate models (∼30 seconds per model) widens its potential usage. ABodyBuilder can also help users in decision-making for the development of novel antibodies because it provides model confidence and potential sequence liabilities. ABodyBuilder is freely available at http://opig.stats.ox.ac.uk/webapps/abodybuilder .

SAbPred: a structure-based antibody prediction server.

SAbPred is a server that makes predictions of the properties of antibodies focusing on their structures. Antibody informatics tools can help improve our understanding of immune responses to disease and aid in the design and engineering of therapeutic molecules. SAbPred is a single platform containing multiple applications which can: number and align sequences; automatically generate antibody variable fragment homology models; annotate such models with estimated accuracy alongside sequence and structural properties including potential developability issues; predict paratope residues; and predict epitope patches on protein antigens. The server is available at http://opig.stats.ox.ac.uk/webapps/sabpred.

SAbDab: the structural antibody database.

Structural antibody database (SAbDab; http://opig.stats.ox.ac.uk/webapps/sabdab) is an online resource containing all the publicly available antibody structures annotated and presented in a consistent fashion. The data are annotated with several properties including experimental information, gene details, correct heavy and light chain pairings, antigen details and, where available, antibody-antigen binding affinity. The user can select structures, according to these attributes as well as structural properties such as complementarity determining region loop conformation and variable domain orientation. Individual structures, datasets and the complete database can be downloaded.

Trapped platelets activated in ischemia initiate ventricular fibrillation.

BACKGROUND: We tested the hypothesis that ischemia-induced ventricular fibrillation (VF) is facilitated by platelets, trapped regionally in the ischemic zone and activated to release arrhythmogenic secretome. METHODS AND RESULTS: In a randomized study in blood-free, buffer-perfused isolated rat hearts, ischemic zone territory (34±1% of left ventricle) was selected so that ischemia evoked VF in only 42% of controls. VF incidence was increased to 91% (P<0.05) by coronary ligation-induced trapping of freshly prepared autologous platelets (infused before and during coronary ligation, with trapping confirmed by 111In-labeled platelet autoradiographic imaging). Trapping of platelet secretome prepared ex vivo, or platelet-sized fluorospheres, did not increase ischemia-induced VF incidence. Secretome alone did, however, evoke VF in 2 sham coronary-ligated hearts. Perfusion did not activate infused platelets in sham coronary-ligated hearts, whereas ligation activated trapped platelets (assessed by thromboxane release). In a separate study, trapping whole-heparinized blood mimicked the ability of trapped platelets to increase VF incidence. This effect was not prevented by >5 days oral pretreatment in vivo with clopidogrel (10 mg/kg per day) or indomethacin (2.4 mg/kg per day). CONCLUSIONS: Platelets facilitate VF during acute ischemia independently of their ability to participate in occlusive thrombosis. Moreover, the effect is unresponsive to antiplatelet drugs commonly used. Labile secretome constituents appear to be responsible. This opens a novel avenue for antiarrhythmic drug research.