DOTAD: A Database of Therapeutic Antibody Developability

The development of therapeutic antibodies is an important aspect of new drug discovery pipelines. The assessment of an antibody's developability-its suitability for large-scale production and therapeutic use-is a particularly important step in this process. Given that experimental assays to assess antibody developability in large scale are expensive and time-consuming, computational methods have been a more efficient alternative. However, the antibody research community faces significant challenges due to the scarcity of readily accessible data on antibody developability, which is essential for training and validating computational models. To address this gap, DOTAD (Database Of Therapeutic Antibody Developability) has been built as the first database dedicated exclusively to the curation of therapeutic antibody developability information. DOTAD aggregates all available therapeutic antibody sequence data along with various developability metrics from the scientific literature, offering researchers a robust platform for data storage, retrieval, exploration, and downloading. In addition to serving as a comprehensive repository, DOTAD enhances its utility by integrating a web-based interface that features state-of-the-art tools for the assessment of antibody developability. This ensures that users not only have access to critical data but also have the convenience of analyzing and interpreting this information. The DOTAD database represents a valuable resource for the scientific community, facilitating the advancement of therapeutic antibody research. It is freely accessible at http://i.uestc.edu.cn/DOTAD/ , providing an open data platform that supports the continuous growth and evolution of computational methods in the field of antibody development.

Molecular surface descriptors to predict antibody developability: sensitivity to parameters, structure models, and conformational sampling

In silico assessment of antibody developability during early lead candidate selection and optimization is of paramount importance, offering a rapid and material-free screening approach. However, the predictive power and reproducibility of such methods depend heavily on the selection of molecular descriptors, model parameters, accuracy of predicted structure models, and conformational sampling techniques. Here, we present a set of molecular surface descriptors specifically designed for predicting antibody developability. We assess the performance of these descriptors by benchmarking their correlations with an extensive array of experimentally determined biophysical properties, including viscosity, aggregation, hydrophobic interaction chromatography, human pharmacokinetic clearance, heparin retention time, and polyspecificity. Further, we investigate the sensitivity of these surface descriptors to methodological nuances, such as the choice of interior dielectric constant, hydrophobicity scales, structure prediction methods, and the impact of conformational sampling. Notably, we observe systematic shifts in the distribution of surface descriptors depending on the structure prediction method used, driving weak correlations of surface descriptors across structure models. Averaging the descriptor values over conformational distributions from molecular dynamics mitigates the systematic shifts and improves the consistency across different structure prediction methods, albeit with inconsistent improvements in correlations with biophysical data. Based on our benchmarking analysis, we propose six in silico developability risk flags and assess their effectiveness in predicting potential developability issues for a set of case study molecules.

Nonclinical immunogenicity risk assessment for knobs-into-holes bispecific IgG1 antibodies

Bispecific antibodies, including bispecific IgG, are emerging as an important new class of antibody therapeutics. As a result, we, as well as others, have developed engineering strategies designed to facilitate the efficient production of bispecific IgG for clinical development. For example, we have extensively used knobs-into-holes (KIH) mutations to facilitate the heterodimerization of antibody heavy chains and more recently Fab mutations to promote cognate heavy/light chain pairing for efficient in vivo assembly of bispecific IgG in single host cells. A panel of related monospecific and bispecific IgG1 antibodies was constructed and assessed for immunogenicity risk by comparison with benchmark antibodies with known low (Avastin and Herceptin) or high (bococizumab and ATR-107) clinical incidence of anti-drug antibodies. Assay methods used include dendritic cell internalization, T cell proliferation, and T cell epitope identification by in silico prediction and MHC-associated peptide proteomics. Data from each method were considered independently and then together for an overall integrated immunogenicity risk assessment. In toto, these data suggest that the KIH mutations and in vitro assembly of half antibodies do not represent a major risk for immunogenicity of bispecific IgG1, nor do the Fab mutations used for efficient in vivo assembly of bispecifics in single host cells. Comparable or slightly higher immunogenicity risk assessment data were obtained for research-grade preparations of trastuzumab and bevacizumab versus Herceptin and Avastin, respectively. These data provide experimental support for the common practice of using research-grade preparations of IgG1 as surrogates for immunogenicity risk assessment of their corresponding pharmaceutical counterparts.

Deep learning in preclinical antibody drug discovery and development

Antibody drugs have become a key part of biotherapeutics. Patients suffering from various diseases have benefited from antibody therapies. However, its development process is rather long, expensive and risky. To speed up the process, reduce cost and improve success rate, artificial intelligence, especially deep learning methods, have been widely used in all aspects of preclinical antibody drug development, from library generation to hit identification, developability screening, lead selection and optimization. In this review, we systematically summarize antibody encodings, deep learning architectures and models used in preclinical antibody drug discovery and development. We also critically discuss challenges and opportunities, problems and possible solutions, current applications and future directions of deep learning in antibody drug development.

How can we discover developable antibody-based biotherapeutics?

Antibody-based biotherapeutics have emerged as a successful class of pharmaceuticals despite significant challenges and risks to their discovery and development. This review discusses the most frequently encountered hurdles in the research and development (R&D) of antibody-based biotherapeutics and proposes a conceptual framework called biopharmaceutical informatics. Our vision advocates for the syncretic use of computation and experimentation at every stage of biologic drug discovery, considering developability (manufacturability, safety, efficacy, and pharmacology) of potential drug candidates from the earliest stages of the drug discovery phase. The computational advances in recent years allow for more precise formulation of disease concepts, rapid identification, and validation of targets suitable for therapeutic intervention and discovery of potential biotherapeutics that can agonize or antagonize them. Furthermore, computational methods for de novo and epitope-specific antibody design are increasingly being developed, opening novel computationally driven opportunities for biologic drug discovery. Here, we review the opportunities and limitations of emerging computational approaches for optimizing antigens to generate robust immune responses, in silico generation of antibody sequences, discovery of potential antibody binders through virtual screening, assessment of hits, identification of lead drug candidates and their affinity maturation, and optimization for developability. The adoption of biopharmaceutical informatics across all aspects of drug discovery and development cycles should help bring affordable and effective biotherapeutics to patients more quickly.

Identifying developability risks for clinical progression of antibodies using high-throughput in vitro and in silico approaches

With the growing significance of antibodies as a therapeutic class, identifying developability risks early during development is of paramount importance. Several high-throughput in vitro assays and in silico approaches have been proposed to de-risk antibodies during early stages of the discovery process. In this review, we have compiled and collectively analyzed published experimental assessments and computational metrics for clinical antibodies. We show that flags assigned based on in vitro measurements of polyspecificity and hydrophobicity are more predictive of clinical progression than their in silico counterparts. Additionally, we assessed the performance of published models for developability predictions on molecules not used during model training. We find that generalization to data outside of those used for training remains a challenge for models. Finally, we highlight the challenges of reproducibility in computed metrics arising from differences in homology modeling, in vitro assessments relying on complex reagents, as well as curation of experimental data often used to assess the utility of high-throughput approaches. We end with a recommendation to enable assay reproducibility by inclusion of controls with disclosed sequences, as well as sharing of structural models to enable the critical assessment and improvement of in silico predictions.

Analytical Workflows to Unlock Predictive Power in Biotherapeutic Developability

PURPOSE

Forming accurate data models that assist the design of developability assays is one area that requires a deep and practical understanding of the problem domain. We aim to incorporate expert knowledge into the model building process by creating new metrics from instrument data and by guiding the choice of input parameters and Machine Learning (ML) techniques.

METHODS

We generated datasets from the biophysical characterisation of 5 monoclonal antibodies (mAbs). We explored combinations of techniques and parameters to uncover the ones that better describe specific molecular liabilities, such as conformational and colloidal instability. We also employed ML algorithms to predict metrics from the dataset.

RESULTS

We found that the combination of Differential Scanning Calorimetry (DSC) and Light Scattering thermal ramps enabled us to identify domain-specific aggregation in mAbs that would be otherwise overlooked by common developability workflows. We also found that the response to different salt concentrations provided information about colloidal stability in agreement with charge distribution models. Finally, we predicted DSC transition temperatures from the dataset, and used the order of importance of different metrics to increase the explainability of the model.

CONCLUSIONS

The new analytical workflows enabled a better description of molecular behaviour and uncovered links between structural properties and molecular liabilities. In the future this new understanding will be coupled with ML algorithms to unlock their predictive power during developability assessment.

High concentration formulation developability approaches and considerations

The growing need for biologics to be administered subcutaneously and ocularly, coupled with certain indications requiring high doses, has resulted in an increase in drug substance (DS) and drug product (DP) protein concentrations. With this increase, more emphasis must be placed on identifying critical physico-chemical liabilities during drug development, including protein aggregation, precipitation, opalescence, particle formation, and high viscosity. Depending on the molecule, liabilities, and administration route, different formulation strategies can be used to overcome these challenges. However, due to the high material requirements, identifying optimal conditions can be slow, costly, and often prevent therapeutics from moving rapidly into the clinic/market. In order to accelerate and derisk development, new experimental and in-silico methods have emerged that can predict high concentration liabilities. Here, we review the challenges in developing high concentration formulations, the advances that have been made in establishing low mass and high-throughput predictive analytics, and advances in in-silico tools and algorithms aimed at identifying risks and understanding high concentration protein behavior.

Prediction of polyreactive and nonspecific single-chain fragment variables through structural biochemical features and protein language-based descriptors

BACKGROUND

Monoclonal antibodies (mAbs) have been used as therapeutic agents, which must overcome many developability issues after the discovery from in vitro display libraries. Especially, polyreactive mAbs can strongly bind to a specific target and weakly bind to off-target proteins, which leads to poor antibody pharmacokinetics in clinical development. Although early assessment of polyreactive mAbs is important in the early discovery stage, experimental assessments are usually time-consuming and expensive. Therefore, computational approaches for predicting the polyreactivity of single-chain fragment variables (scFvs) in the early discovery stage would be promising for reducing experimental efforts.

RESULTS

Here, we made prediction models for the polyreactivity of scFvs with the known polyreactive antibody features and natural language model descriptors. We predicted 19,426 protein structures of scFvs with trRosetta to calculate the polyreactive antibody features and investigated the classifying performance of each factor for polyreactivity. In the known polyreactive features, the net charge of the CDR2 loop, the tryptophan and glycine residues in CDR-H3, and the lengths of the CDR1 and CDR2 loops, importantly contributed to the performance of the models. Additionally, the hydrodynamic features, such as partial specific volume, gyration radius, and isoelectric points of CDR loops and scFvs, were newly added to improve model performance. Finally, we made the prediction model with a robust performance ([Formula: see text]) with an ensemble learning of the top 3 best models.

CONCLUSION

The prediction models for polyreactivity would help assess polyreactive scFvs in the early discovery stage and our approaches would be promising to develop machine learning models with quantitative data from high throughput assays for antibody screening.

The use of predictive models to develop chromatography-based purification processes

Chromatography is the workhorse of biopharmaceutical downstream processing because it can selectively enrich a target product while removing impurities from complex feed streams. This is achieved by exploiting differences in molecular properties, such as size, charge and hydrophobicity (alone or in different combinations). Accordingly, many parameters must be tested during process development in order to maximize product purity and recovery, including resin and ligand types, conductivity, pH, gradient profiles, and the sequence of separation operations. The number of possible experimental conditions quickly becomes unmanageable. Although the range of suitable conditions can be narrowed based on experience, the time and cost of the work remain high even when using high-throughput laboratory automation. In contrast, chromatography modeling using inexpensive, parallelized computer hardware can provide expert knowledge, predicting conditions that achieve high purity and efficient recovery. The prediction of suitable conditions in silico reduces the number of empirical tests required and provides in-depth process understanding, which is recommended by regulatory authorities. In this article, we discuss the benefits and specific challenges of chromatography modeling. We describe the experimental characterization of chromatography devices and settings prior to modeling, such as the determination of column porosity. We also consider the challenges that must be overcome when models are set up and calibrated, including the cross-validation and verification of data-driven and hybrid (combined data-driven and mechanistic) models. This review will therefore support researchers intending to establish a chromatography modeling workflow in their laboratory.

Comparison of hydrophobicity scales for predicting biophysical properties of antibodies

While antibody-based therapeutics have grown to be one of the major classes of novel medicines, some antibody development candidates face significant challenges regarding expression levels, solubility, as well as stability and aggregation, under physiological and storage conditions. A major determinant of those properties is surface hydrophobicity, which promotes unspecific interactions and has repeatedly proven problematic in the development of novel antibody-based drugs. Multiple computational methods have been devised for in-silico prediction of antibody hydrophobicity, often using hydrophobicity scales to assign values to each amino acid. Those approaches are usually validated by their ability to rank potential therapeutic antibodies in terms of their experimental hydrophobicity. However, there is significant diversity both in the hydrophobicity scales and in the experimental methods, and consequently in the performance of in-silico methods to predict experimental results. In this work, we investigate hydrophobicity of monoclonal antibodies using hydrophobicity scales. We implement several scoring schemes based on the solvent-accessibility and the assigned hydrophobicity values, and compare the different scores and scales based on their ability to predict retention times from hydrophobic interaction chromatography. We provide an overview of the strengths and weaknesses of several commonly employed hydrophobicity scales, thereby improving the understanding of hydrophobicity in antibody development. Furthermore, we test several datasets, both publicly available and proprietary, and find that the diversity of the dataset affects the performance of hydrophobicity scores. We expect that this work will provide valuable guidelines for the optimization of biophysical properties in future drug discovery campaigns.

Machine-designed biotherapeutics: opportunities, feasibility and advantages of deep learning in computational antibody discovery

Antibodies are versatile molecular binders with an established and growing role as therapeutics. Computational approaches to developing and designing these molecules are being increasingly used to complement traditional lab-based processes. Nowadays, in silico methods fill multiple elements of the discovery stage, such as characterizing antibody-antigen interactions and identifying developability liabilities. Recently, computational methods tackling such problems have begun to follow machine learning paradigms, in many cases deep learning specifically. This paradigm shift offers improvements in established areas such as structure or binding prediction and opens up new possibilities such as language-based modeling of antibody repertoires or machine-learning-based generation of novel sequences. In this review, we critically examine the recent developments in (deep) machine learning approaches to therapeutic antibody design with implications for fully computational antibody design.

Co-optimization of therapeutic antibody affinity and specificity using machine learning models that generalize to novel mutational space

Therapeutic antibody development requires selection and engineering of molecules with high affinity and other drug-like biophysical properties. Co-optimization of multiple antibody properties remains a difficult and time-consuming process that impedes drug development. Here we evaluate the use of machine learning to simplify antibody co-optimization for a clinical-stage antibody (emibetuzumab) that displays high levels of both on-target (antigen) and off-target (non-specific) binding. We mutate sites in the antibody complementarity-determining regions, sort the antibody libraries for high and low levels of affinity and non-specific binding, and deep sequence the enriched libraries. Interestingly, machine learning models trained on datasets with binary labels enable predictions of continuous metrics that are strongly correlated with antibody affinity and non-specific binding. These models illustrate strong tradeoffs between these two properties, as increases in affinity along the co-optimal (Pareto) frontier require progressive reductions in specificity. Notably, models trained with deep learning features enable prediction of novel antibody mutations that co-optimize affinity and specificity beyond what is possible for the original antibody library. These findings demonstrate the power of machine learning models to greatly expand the exploration of novel antibody sequence space and accelerate the development of highly potent, drug-like antibodies.

DeepSCM: An efficient convolutional neural network surrogate model for the screening of therapeutic antibody viscosity

Predicting high concentration antibody viscosity is essential for developing subcutaneous administration. Computer simulations provide promising tools to reach this aim. One such model is the spatial charge map (SCM) proposed by Agrawal and coworkers (mAbs. 2015, 8(1):43-48). SCM applies molecular dynamics simulations to calculate a score for the screening of antibody viscosity at high concentrations. However, molecular dynamics simulations are computationally costly and require structural information, a significant application bottleneck. In this work, high throughput computing was performed to calculate the SCM scores for 6596 nonredundant antibody variable regions. A convolutional neural network surrogate model, DeepSCM, requiring only sequence information, was then developed based on this dataset. The linear correlation coefficient of the DeepSCM and SCM scores achieved 0.9 on the test set (N = 1320). The DeepSCM model was applied to screen the viscosity of 38 therapeutic antibodies that SCM correctly classified and resulted in only one misclassification. The DeepSCM model will facilitate high concentration antibody viscosity screening. The code and parameters are freely available at https://github.com/Lailabcode/DeepSCM.

SSH2.0: A Better Tool for Predicting the Hydrophobic Interaction Risk of Monoclonal Antibody

Therapeutic antibodies play a crucial role in the treatment of various diseases. However, the success rate of antibody drug development is low partially because of unfavourable biophysical properties of antibody drug candidates such as the high aggregation tendency, which is mainly driven by hydrophobic interactions of antibody molecules. Therefore, early screening of the risk of hydrophobic interaction of antibody drug candidates is crucial. Experimental screening is laborious, time-consuming, and costly, warranting the development of efficient and high-throughput computational tools for prediction of hydrophobic interactions of therapeutic antibodies. In the present study, 131 antibodies with hydrophobic interaction experiment data were used to train a new support vector machine-based ensemble model, termed SSH2.0, to predict the hydrophobic interactions of antibodies. Feature selection was performed against CKSAAGP by using the graph-based algorithm MRMD2.0. Based on the antibody sequence, SSH2.0 achieved the sensitivity and accuracy of 100.00 and 83.97%, respectively. This approach eliminates the need of three-dimensional structure of antibodies and enables rapid screening of therapeutic antibody candidates in the early developmental stage, thereby saving time and cost. In addition, a web server was constructed that is freely available at http://i.uestc.edu.cn/SSH2/.

Alterations of the Gut Microbiome in Recurrent Malignant Gliomas Patients Received Bevacizumab and Temozolomide Combination Treatment and Temozolomide Monotherapy

This case-control study explored compositions of gut microbiome in recurrent malignant gliomas patients who had received bevacizumab and Temozolomide combination treatment and Temozolomide monotherapy. We investigated gut microbiota communities in feces of 29 recurrent malignant gliomas patients received combination treatment with bevacizumab and Temozolomide (Group 1) and monotherapy with Temozolomide alone (Group 2). We took advantage of the high-throughput Illumina Miseq sequencing technology by targeting the third and fourth hypervariable (V3-V4) regions of the 16S ribosomal RNA (rRNA) gene. We found that the structures and richness of the fecal microbiota in Group 1 were different from Group 2 with LEfSe analysis. The fecal microbiota in both Group 1 and Group 2 were mainly composed by Firmicutes, Proteobacteria, Bacteroidetes and Actinobacteria. However, Group 1 patients had higher relative abundance of Firmicutes, Bacteroidetes, Actinobacteria and lower relative abundance of Bacteroidetes and Cyanobacteria in their fecal microbiota than that in Group 2 patients. To evaluate bevacizumab involved post-treatment state of the fecal microbiota profile, we used random forest predictive model and ensembled decision trees with an AUC of 0.54. This study confirmed that the gut microbiota was different in recurrent malignant gliomas patients received the combination therapy of bevacizumab and Temozolomide compared with Temozolomide monotherapy. Our discover can help better understand the influence of bevacizumab related treatment on recurrent malignant gliomas patients. Therefore, this finding may also support the potentially therapeutic options for recurrent malignant gliomas patients such as fecal microbiota transplant.

SUPPLEMENTARY INFORMATION

The online version contains supplementary material available at 10.1007/s12088-021-00962-2.

Current advances in biopharmaceutical informatics: guidelines, impact and challenges in the computational developability assessment of antibody therapeutics

Therapeutic monoclonal antibodies and their derivatives are key components of clinical pipelines in the global biopharmaceutical industry. The availability of large datasets of antibody sequences, structures, and biophysical properties is increasingly enabling the development of predictive models and computational tools for the "developability assessment" of antibody drug candidates. Here, we provide an overview of the antibody informatics tools applicable to the prediction of developability issues such as stability, aggregation, immunogenicity, and chemical degradation. We further evaluate the opportunities and challenges of using biopharmaceutical informatics for drug discovery and optimization. Finally, we discuss the potential of developability guidelines based on in silico metrics that can be used for the assessment of antibody stability and manufacturability.

Explicit solvation thermodynamics in ionic solution: extending grid inhomogeneous solvation theory to solvation free energy of salt-water mixtures

Hydration thermodynamics play a fundamental role in fields ranging from the pharmaceutical industry to environmental research. Numerous methods exist to predict solvation thermodynamics of compounds ranging from small molecules to large biomolecules. Arguably the most precise methods are those based on molecular dynamics (MD) simulations in explicit solvent. One theory that has seen increased use is inhomogeneous solvation theory (IST). However, while many applications require accurate description of salt-water mixtures, no implementation of IST is currently able to estimate solvation properties involving more than one solvent species. Here, we present an extension to grid inhomogeneous solvation theory (GIST) that can take salt contributions into account. At the example of carbazole in 1 M NaCl solution, we compute the solvation energy as well as first and second order entropies. While the effect of the first order ion entropy is small, both the water-water and water-ion entropies contribute strongly. We show that the water-ion entropies are efficiently approximated using the Kirkwood superposition approximation. However, this approach cannot be applied to the water-water entropy. Furthermore, we test the quantitative validity of our method by computing salting-out coefficients and comparing them to experimental data. We find a good correlation to experimental salting-out constants, while the absolute values are overpredicted due to the approximate second order entropy. Since ions are frequently used in MD, either to neutralize the system or as a part of the investigated process, our method greatly extends the applicability of GIST. The use-cases range from biopharmaceuticals, where many assays require high salt concentrations, to environmental research, where solubility in sea water is important to model the fate of organic substances.

Assessment of Therapeutic Antibody Developability by Combinations of In Vitro and In Silico Methods

Although antibodies have become the fastest-growing class of therapeutics on the market, it is still challenging to develop them for therapeutic applications, which often require these molecules to withstand stresses that are not present in vivo. We define developability as the likelihood of an antibody candidate with suitable functionality to be developed into a manufacturable, stable, safe, and effective drug that can be formulated to high concentrations while retaining a long shelf life. The implementation of reliable developability assessments from the early stages of antibody discovery enables flagging and deselection of potentially problematic candidates, while focussing available resources on the development of the most promising ones. Currently, however, thorough developability assessment requires multiple in vitro assays, which makes it labor intensive and time consuming to implement at early stages. Furthermore, accurate in vitro analysis at the early stage is compromised by the high number of potential candidates that are often prepared at low quantities and purity. Recent improvements in the performance of computational predictors of developability potential are beginning to change this scenario. Many computational methods only require the knowledge of the amino acid sequences and can be used to identify possible developability issues or to rank available candidates according to a range of biophysical properties. Here, we describe how the implementation of in silico tools into antibody discovery pipelines is increasingly offering time- and cost-effective alternatives to in vitro experimental screening, thus streamlining the drug development process. We discuss in particular the biophysical and biochemical properties that underpin developability potential and their trade-offs, review various in vitro assays to measure such properties or parameters that are predictive of developability, and give an overview of the growing number of in silico tools available to predict properties important for antibody development, including the CamSol method developed in our laboratory.

A computational method for predicting the aggregation propensity of IgG1 and IgG4(P) mAbs in common storage buffers

The propensity for some monoclonal antibodies (mAbs) to aggregate at physiological and manufacturing pH values can prevent their use as therapeutic molecules or delay time to market. Consequently, developability assessments are essential to select optimum candidates, or inform on mitigation strategies to avoid potential late-stage failures. These studies are typically performed in a range of buffer solutions because factors such as pH can dramatically alter the aggregation propensity of the test mAbs (up to 100-fold in extreme cases). A computational method capable of robustly predicting the aggregation propensity at the pH values of common storage buffers would have substantial value. Here, we describe a mAb aggregation prediction tool (MAPT) that builds on our previously published isotype-dependent, charge-based model of aggregation. We show that the addition of a homology model-derived hydrophobicity descriptor to our electrostatic aggregation model enabled the generation of a robust mAb developability indicator. To contextualize our aggregation scoring system, we analyzed 97 clinical-stage therapeutic mAbs. To further validate our approach, we focused on six mAbs (infliximab, tocilizumab, rituximab, CNTO607, MEDI1912 and MEDI1912_STT) which have been reported to cover a large range of aggregation propensities. The different aggregation propensities of the case study molecules at neutral and slightly acidic pH were correctly predicted, verifying the utility of our computational method.

Predicting antibody binders and generating synthetic antibodies using deep learning

The antibody drug field has continually sought improvements to methods for candidate discovery and engineering. Historically, most such methods have been laboratory-based, but informatics methods have recently started to make an impact. Deep learning, a subfield of machine learning, is rapidly gaining prominence in the biomedical research. Recent advances in microfluidics technologies and next-generation sequencing have not only revolutionized therapeutic antibody discovery, but also contributed to a vast amount of antibody repertoire sequencing data, providing opportunities for deep learning-based applications. Previously, we used microfluidics, yeast display, and deep sequencing to generate a panel of binder and non-binder antibody sequences to the cancer immunotherapy targets PD-1 and CTLA-4. Here we encoded the antibody light and heavy chain complementarity-determining regions (CDR3s) into antibody images, then built and trained convolutional neural network models to classify binders and non-binders. To improve model interpretability, we performed in silico mutagenesis to identify CDR3 residues that were important for binder classification. We further built generative deep learning models using generative adversarial network models to produce synthetic antibodies against PD-1 and CTLA-4. Our models generated variable length CDR3 sequences that resemble real sequences. Overall, our study demonstrates that deep learning methods can be leveraged to mine and learn patterns in antibody sequences, offering insights into antibody engineering, optimization, and discovery.

Deep Learning in Therapeutic Antibody Development

Deep learning applied to antibody development is in its adolescence. Low data volumes and biological platform differences make it challenging to develop supervised models that can predict antibody behavior in actual commercial development steps. But successes in modeling general protein behaviors and early antibody models give indications of what is possible for antibodies in general, particularly since antibodies share a common fold. Meanwhile, new methods of data collection and the development of unsupervised and self-supervised deep learning methods like generative models and masked language models give the promise of rich and deep data sets and deep learning architectures for better supervised model development. Together, these move the industry toward improved developability , lower costs, and broader access of biotherapeutics .

The Therapeutic Antibody Profiler for Computational Developability Assessment

The need to consider an antibody's "developability" (immunogenicity, solubility, specificity, stability, manufacturability, and storability) is now well understood in therapeutic antibody design. Predicting these properties rapidly and inexpensively is critical to industrial workflows, to avoid devoting resources to non-productive candidates. Here, we describe a high-throughput computational developability assessment tool, the Therapeutic Antibody Profiler (TAP), which assesses the physicochemical "druglikeness" of an antibody candidate. Input variable domain sequences are converted to three-dimensional structural models, and then five developability-linked molecular surface descriptors are calculated and compared to advanced-stage clinical therapeutics. Values at the extremes of/outside of the distributions seen in therapeutics imply an increased risk of developability issues. Therefore, TAP, starting only from sequence information, provides a route to rapidly identifying drug candidate antibodies that are likely to have poor developability. Our web application ( opig.stats.ox.ac.uk/webapps/tap ) profiles input antibody sequences against a continually updated reference set of clinical therapeutics.

Stress-dependent flexibility of a full-length human monoclonal antibody: Insights from molecular dynamics to support biopharmaceutical development

After several decades of advancements in drug discovery, product development of biopharmaceuticals remains a time- and resource-consuming endeavor. One of the main reasons is associated to the lack of fundamental understanding of conformational dynamics of such biologic entities, and how they respond to various stresses encountered during manufacturing. In this work, we have studied the conformational dynamics of human IgG₁κ b12 monoclonal antibody (mAb) using molecular dynamics simulations. The hundreds of nanoseconds long trajectories reveal that b12 mAb is highly flexible. Its variable domains show greater conformational fluctuations than the constant domains. Additionally, it collapses towards a more globular shape in response to thermal stress, leading to decrease in the total solvent exposed surface area and radius of gyration. This behavior is more pronounced for the deglycosylated b12 mAb, and it appears to correlate with increase in inter-domain contacts between specific regions of the antibody. Conformational fluctuations also cause temporary formation and disruption of hydrophobic and charged patches on the antibody surface, which is particularly important for the prediction of CMC properties during development phases of antibody-based biotherapeutics. The insights gained through these simulations may help the development of biologic drugs, especially with regards to manufacturing processes where antibodies may undergo significant thermal stress.

Pragmatic mAb lead molecule engineering from a developability perspective

As the number of antibody drugs being approved and marketed increases, our knowledge of what makes potential drug candidates a successful product has increased tremendously. One of the critical parameters that have become clear in the field is the importance of mAb "developability." Efforts are being increasingly focused on simultaneously selecting molecules that exhibit both desirable biological potencies and manufacturability attributes. In the current study mutations to improve the developability profile of a problematic antibody that inconsistently precipitates in a batch scale-dependent fashion using a standard platform purification process are described. Initial bioinformatic analysis showed the molecule has no obvious sequence or structural liabilities that might lead it to precipitate. Subsequent analysis of the molecule revealed the presence of two unusual positively charged mutations on the light chain at the interface of VH and VL domains, which were hypothesized to be the primary contributor to molecule precipitation during process development. To investigate this hypothesis, straightforward reversion to the germline of these residues was carried out. The resulting mutants have improved expression titers and recovered stability within a forced precipitation assay, without any change to biological activity. Given the time pressures of drug development in industry, process optimization of the lead molecule was carried out in parallel to the "retrospective" mutagenesis approach. Bespoke process optimization for large-scale manufacturing was successful. However, we propose that such context-dependent sequence liabilities should be included in the arsenal of in silico developability screening early in development; particularly since this specific issue can be efficiently mitigated without the requirement for extensive screening of lead molecule variants.

Applications of Machine and Deep Learning in Adaptive Immunity

Adaptive immunity is mediated by lymphocyte B and T cells, which respectively express a vast and diverse repertoire of B cell and T cell receptors and, in conjunction with peptide antigen presentation through major histocompatibility complexes (MHCs), can recognize and respond to pathogens and diseased cells. In recent years, advances in deep sequencing have led to a massive increase in the amount of adaptive immune receptor repertoire data; additionally, proteomics techniques have led to a wealth of data on peptide-MHC presentation. These large-scale data sets are now making it possible to train machine and deep learning models, which can be used to identify complex and high-dimensional patterns in immune repertoires. This article introduces adaptive immune repertoires and machine and deep learning related to biological sequence data and then summarizes the many applications in this field, which span from predicting the immunological status of a host to the antigen specificity of individual receptors and the engineering of immunotherapeutics.

Predicting Antibody Developability Profiles Through Early Stage Discovery Screening

Monoclonal antibodies play an increasingly important role for the development of new drugs across multiple therapy areas. The term 'developability' encompasses the feasibility of molecules to successfully progress from discovery to development via evaluation of their physicochemical properties. These properties include the tendency for self-interaction and aggregation, thermal stability, colloidal stability, and optimization of their properties through sequence engineering. Selection of the best antibody molecule based on biological function, efficacy, safety, and developability allows for a streamlined and successful CMC phase. An efficient and practical high-throughput developability workflow (100 s-1,000 s of molecules) implemented during early antibody generation and screening is crucial to select the best lead candidates. This involves careful assessment of critical developability parameters, combined with binding affinity and biological properties evaluation using small amounts of purified material (<1 mg), as well as an efficient data management and database system. Herein, a panel of 152 various human or humanized monoclonal antibodies was analyzed in biophysical property assays. Correlations between assays for different sets of properties were established. We demonstrated in two case studies that physicochemical properties and key assay endpoints correlate with key downstream process parameters. The workflow allows the elimination of antibodies with suboptimal properties and a rank ordering of molecules for further evaluation early in the candidate selection process. This enables any further engineering for problematic sequence attributes without affecting program timelines.

Developments and opportunities in continuous biopharmaceutical manufacturing

Today's biologics manufacturing practices incur high costs to the drug makers, which can contribute to high prices for patients. Timely investment in the development and implementation of continuous biomanufacturing can increase the production of consistent-quality drugs at a lower cost and a faster pace, to meet growing demand. Efficient use of equipment, manufacturing footprint, and labor also offer the potential to improve drug accessibility. Although technological efforts enabling continuous biomanufacturing have commenced, challenges remain in the integration, monitoring, and control of traditionally segmented unit operations. Here, we discuss recent developments supporting the implementation of continuous biomanufacturing, along with their benefits.

Machine Learning Feature Selection for Predicting High Concentration Therapeutic Antibody Aggregation

Protein aggregation can hinder the development, safety and efficacy of therapeutic antibody-based drugs. Developing a predictive model that evaluates aggregation behaviors during early stage development is therefore desirable. Machine learning is a widely used tool to train models that predict data with different attributes. However, most machine learning techniques require more data than is typically available in antibody development. In this work, we describe a rational feature selection framework to develop accurate models with a small number of features. We applied this framework to predict aggregation behaviors of 21 approved monospecific monoclonal antibodies at high concentration (150 mg/mL), yielding a correlation coefficient of 0.71 on validation tests with only two features using a linear model. The nearest neighbors and support vector regression models further improved the performance, which have correlation coefficients of 0.86 and 0.80, respectively. This framework can be extended to train other models that predict different physical properties.

Biochemical patterns of antibody polyreactivity revealed through a bioinformatics-based analysis of CDR loops

Antibodies are critical components of adaptive immunity, binding with high affinity to pathogenic epitopes. Antibodies undergo rigorous selection to achieve this high affinity, yet some maintain an additional basal level of low affinity, broad reactivity to diverse epitopes, a phenomenon termed ‘polyreactivity’. While polyreactivity has been observed in antibodies isolated from various immunological niches, the biophysical properties that allow for promiscuity in a protein selected for high-affinity binding to a single target remain unclear. Using a database of over 1000 polyreactive and non-polyreactive antibody sequences, we created a bioinformatic pipeline to isolate key determinants of polyreactivity. These determinants, which include an increase in inter-loop crosstalk and a propensity for a neutral binding surface, are sufficient to generate a classifier able to identify polyreactive antibodies with over 75% accuracy. The framework from which this classifier was built is generalizable, and represents a powerful, automated pipeline for future immune repertoire analysis.

To defend itself against bacteria and viruses, the body depends on a group of proteins known as antibodies. Each subset of antibodies undergoes a rigorous training regimen to ensure it recognizes a single epitope well – that is, one specific region on the surface of foreign, harmful organisms.

Most antibodies stick extremely tightly to their one unique epitope, but some can also weakly bind to molecules that are vastly different from their main trained targets. This feature – known as polyreactivity – can in some cases help the immune system fight against multiple strains of viruses. On the other hand, when antibodies are designed in the laboratory to treat diseases, this characteristic can sometimes lead to the failure of pre-clinical trials. Yet it is currently unclear why some antibodies are polyreactive when others are not.

To investigate this question, Boughter et al. compared over 1,000 polyreactive and non-polyreactive antibody sequences from a large database, revealing differences in the physical properties of the region of the antibodies that attaches to epitopes. Using these defining features, Boughter et al. went on to design a new piece of freely available, automated software that could predict which antibodies would be polyreactive more than 75% of the time.

Such software could ultimately help to guide the design of antibody-based treatments, while bypassing the need for costly laboratory tests.

Computational approaches to therapeutic antibody design: established methods and emerging trends

Antibodies are proteins that recognize the molecular surfaces of potentially noxious molecules to mount an adaptive immune response or, in the case of autoimmune diseases, molecules that are part of healthy cells and tissues. Due to their binding versatility, antibodies are currently the largest class of biotherapeutics, with five monoclonal antibodies ranked in the top 10 blockbuster drugs. Computational advances in protein modelling and design can have a tangible impact on antibody-based therapeutic development. Antibody-specific computational protocols currently benefit from an increasing volume of data provided by next generation sequencing and application to related drug modalities based on traditional antibodies, such as nanobodies. Here we present a structured overview of available databases, methods and emerging trends in computational antibody analysis and contextualize them towards the engineering of candidate antibody therapeutics.

Physicochemical Rules for Identifying Monoclonal Antibodies with Drug-like Specificity

The ability of antibodies to recognize their target antigens with high specificity is fundamental to their natural function. Nevertheless, therapeutic antibodies display variable and difficult-to-predict levels of nonspecific and self-interactions that can lead to various drug development challenges, including antibody aggregation, abnormally high viscosity, and rapid antibody clearance. Here we report a method for predicting the overall specificity of antibodies in terms of their relative risk for displaying high levels of nonspecific or self-interactions at physiological conditions. We find that individual and combined sets of chemical rules that limit the maximum and minimum numbers of certain solvent-exposed amino acids in antibody variable regions are strong predictors of specificity for large panels of preclinical and clinical-stage antibodies. We also demonstrate how the chemical rules can be used to identify sites that mediate nonspecific interactions in suboptimal antibodies and guide the design of targeted sublibraries that yield variants with high antibody specificity. These findings can be readily used to improve the selection and engineering of antibodies with drug-like specificity.

SSH: A Tool for Predicting Hydrophobic Interaction of Monoclonal Antibodies Using Sequences

Therapeutic antibodies are one of the most important parts of the pharmaceutical industry. They are widely used in treating various diseases such as autoimmune diseases, cancer, inflammation, and infectious diseases. Their development process however is often brought to a standstill or takes a longer time and is then more expensive due to their hydrophobicity problems. Hydrophobic interactions can cause problems on half-life, drug administration, and immunogenicity at all stages of antibody drug development. Some of the most widely accepted and used technologies for determining the hydrophobic interactions of antibodies include standup monolayer adsorption chromatography (SMAC), salt-gradient affinity-capture self-interaction nanoparticle spectroscopy (SGAC-SINS), and hydrophobic interaction chromatography (HIC). However, to measure SMAC, SGAC-SINS, and HIC for hundreds of antibody drug candidates is time-consuming and costly. To save time and money, a predictor called SSH is developed. Based on the antibody's sequence only, it can predict the hydrophobic interactions of monoclonal antibodies (mAbs). Using the leave-one-out crossvalidation, SSH achieved 91.226% accuracy, 96.396% sensitivity or recall, 84.196% specificity, 87.754% precision, 0.828 Mathew correlation coefficient (MCC), 0.919 f-score, and 0.961 area under the receiver operating characteristic (ROC) curve (AUC).

How repertoire data are changing antibody science

Antibodies are vital proteins of the immune system that recognize potentially harmful molecules and initiate their removal. Mammals can efficiently create vast numbers of antibodies with different sequences capable of binding to any antigen with high affinity and specificity. Because they can be developed to bind to many disease agents, antibodies can be used as therapeutics. In an organism, after antigen exposure, antibodies specific to that antigen are enriched through clonal selection, expansion, and somatic hypermutation. The antibodies present in an organism therefore report on its immune status, describe its innate ability to deal with harmful substances, and reveal how it has previously responded. Next-generation sequencing technologies are being increasingly used to query the antibody, or B-cell receptor (BCR), sequence repertoire, and the amount of BCR data in public repositories is growing. The Observed Antibody Space database, for example, currently contains over a billion sequences from 68 different studies. Repertoires are available that represent both the naive state (i.e. antigen-inexperienced) and that after immunization. This wealth of data has created opportunities to learn more about our immune system. In this review, we discuss the many ways in which BCR repertoire data have been or could be exploited. We highlight its utility for providing insights into how the naive immune repertoire is generated and how it responds to antigens. We also consider how structural information can be used to enhance these data and may lead to more accurate depictions of the sequence space and to applications in the discovery of new therapeutics.

Charge and hydrophobicity are key features in sequence-trained machine learning models for predicting the biophysical properties of clinical-stage antibodies

Improved understanding of properties that mediate protein solubility and resistance to aggregation are important for developing biopharmaceuticals, and more generally in biotechnology and synthetic biology. Recent acquisition of large datasets for antibody biophysical properties enables the search for predictive models. In this report, machine learning methods are used to derive models for 12 biophysical properties. A physicochemical perspective is maintained in analysing the models, leading to the observation that models cluster largely according to charge (cross-interaction measurements) and hydrophobicity (self-interaction methods). These two properties also overlap in some cases, for example in a new interpretation of variation in hydrophobic interaction chromatography. Since the models are developed from differences of antibody variable loops, the next stage is to extend models to more diverse protein sets.

AVAILABILITY

The web application for the sequence-based algorithms are available on the protein-sol webserver, at https://protein-sol.manchester.ac.uk/abpred, with models and virtualisation software available at https://protein-sol.manchester.ac.uk/software.

Five computational developability guidelines for therapeutic antibody profiling

Immunogenicity, instability, self-association, high viscosity, polyspecificity, or poor expression can all preclude an antibody from becoming a therapeutic. Early identification of these negative characteristics is essential. Akin to the Lipinski guidelines, which measure druglikeness in small molecules, our Therapeutic Antibody Profiler highlights antibodies that possess characteristics that are rare/unseen in clinical-stage mAb therapeutics. The only required input is the variable domain sequence. We show examples where our approach would have advised against manufacturing antibodies that were found to aggregate or have poor expression.

Therapeutic mAbs must not only bind to their target but must also be free from “developability issues” such as poor stability or high levels of aggregation. While small-molecule drug discovery benefits from Lipinski’s rule of five to guide the selection of molecules with appropriate biophysical properties, there is currently no in silico analog for antibody design. Here, we model the variable domain structures of a large set of post-phase-I clinical-stage antibody therapeutics (CSTs) and calculate in silico metrics to estimate their typical properties. In each case, we contextualize the CST distribution against a snapshot of the human antibody gene repertoire. We describe guideline values for five metrics thought to be implicated in poor developability: the total length of the complementarity-determining regions (CDRs), the extent and magnitude of surface hydrophobicity, positive charge and negative charge in the CDRs, and asymmetry in the net heavy- and light-chain surface charges. The guideline cutoffs for each property were derived from the values seen in CSTs, and a flagging system is proposed to identify nonconforming candidates. On two mAb drug discovery sets, we were able to selectively highlight sequences with developability issues. We make available the Therapeutic Antibody Profiler (TAP), a computational tool that builds downloadable homology models of variable domain sequences, tests them against our five developability guidelines, and reports potential sequence liabilities and canonical forms. TAP is freely available at opig.stats.ox.ac.uk/webapps/sabdab-sabpred/TAP.php.

AggScore: Prediction of aggregation-prone regions in proteins based on the distribution of surface patches

Protein aggregation is a phenomenon that has attracted considerable attention within the pharmaceutical industry from both a developability standpoint (to ensure stability of protein formulations) and from a research perspective for neurodegenerative diseases. Experimental identification of aggregation behavior in proteins can be expensive; and hence, the development of accurate computational approaches is crucial. The existing methods for predicting protein aggregation rely mostly on the primary sequence and are typically trained on amyloid-like proteins. However, the training bias toward beta amyloid peptides may worsen prediction accuracy of such models when applied to larger protein systems. Here, we present a novel algorithm to identify aggregation-prone regions in proteins termed "AggScore" that is based entirely on three-dimensional structure input. The method uses the distribution of hydrophobic and electrostatic patches on the surface of the protein, factoring in the intensity and relative orientation of the respective surface patches into an aggregation propensity function that has been trained on a benchmark set of 31 adnectin proteins. AggScore can accurately identify aggregation-prone regions in several well-studied proteins and also reliably predict changes in aggregation behavior upon residue mutation. The method is agnostic to an amyloid-specific aggregation context and thus may be applied to globular proteins, small peptides and antibodies.