Low-data interpretable deep learning prediction of antibody viscosity using a biophysically meaningful representation

DeepSP: Deep learning-based spatial properties to predict monoclonal antibody stability

Therapeutic antibody development faces challenges due to high viscosities and aggregation tendencies. The spatial charge map (SCM) and spatial aggregation propensity (SAP) are computational techniques that aid in predicting viscosity and aggregation, respectively. These methods rely on structural data derived from molecular dynamics (MD) simulations, which are computationally demanding. DeepSCM, a deep learning surrogate model based on sequence information to predict SCM, was recently developed to screen high-concentration antibody viscosity. This study further utilized a dataset of 20,530 antibody sequences to train a convolutional neural network deep learning surrogate model called Deep Spatial Properties (DeepSP). DeepSP directly predicts SAP and SCM scores in different domains of antibody variable regions based solely on their sequences without performing MD simulations. The linear correlation coefficient between DeepSP scores and MD-derived scores for 30 properties achieved values between 0.76 and 0.96 with an average of 0.87. DeepSP descriptors were employed as features to build machine learning models to predict the aggregation rate of 21 antibodies, and the performance is similar to the results obtained from the previous study using MD simulations. This result demonstrates that the DeepSP approach significantly reduces the computational time required compared to MD simulations. The DeepSP model enables the rapid generation of 30 structural properties that can also be used as features in other research to train machine learning models for predicting various antibody stability using sequences only. DeepSP is freely available as an online tool via https://deepspwebapp.onrender.com and the codes and parameters are freely available at https://github.com/Lailabcode/DeepSP.

Sequence-Based Viscosity Prediction for Rapid Antibody Engineering

Through machine learning, identifying correlations between amino acid sequences of antibodies and their observed characteristics, we developed an internal viscosity prediction model to empower the rapid engineering of therapeutic antibody candidates. For a highly viscous anti-IL-13 monoclonal antibody, we used a structure-based rational design strategy to generate a list of variants that were hypothesized to mitigate viscosity. Our viscosity prediction tool was then used as a screen to cull virtually engineered variants with a probability of high viscosity while advancing those with a probability of low viscosity to production and testing. By combining the rational design engineering strategy with the in silico viscosity prediction screening step, we were able to efficiently improve the highly viscous anti-IL-13 candidate, successfully decreasing the viscosity at 150 mg/mL from 34 cP to 13 cP in a panel of 16 variants.

ProtT5 and random forests-based viscosity prediction method for therapeutic mAbs

Viscosity is a key characteristic of therapeutic antibodies for subcutaneous administration which requires low volume and high concentration formulations. It would be highly beneficial to accurately predict the viscosity of newly developed therapeutic antibodies in the early stages of development. In this work, a ProtT5-XL-UniRef50 (ProtT5) and Random Forests (RF)-based prediction method was proposed for accurately predicting the viscosity of monoclonal antibodies, with only corresponding sequences needed. Starting from the given heavy and light chain V-region sequences, corresponding features were first extracted from the ProtT5 pretrained model. Kernel principal analysis (Kernel-PCA) was then used for reducing the extracted 2048-D (1024-D for each sequence) feature vector to a reasonable level for efficient training of the RF-regressor. Then, the RF model was constructed on 40 commercially available therapeutic antibodies and tested with 3-folds cross-validation. Test results show that the model could reproduce the viscosity value at a high level (Pearson correlation coefficient (PCC) = 0.928). Performance on classifying high (>30 cP) and low (<30 cP) viscosity is much more satisfactory, the Accuracy (ACC) and the area under precision-recall curve (AUC) of the classification model from validation tests are 0.975 and 1.000, respectively. Compared to 5 existing state-of-the-art viscosity prediction methods, the proposed method performs best which would facilitate high concentration antibody viscosity high-throughput screening.

Recent advances in generative biology for biotherapeutic discovery

Generative biology combines artificial intelligence (AI), advanced life sciences technologies, and automation to revolutionize the process of designing novel biomolecules with prescribed properties, giving drug discoverers the ability to escape the limitations of biology during the design of next-generation protein therapeutics. Significant hurdles remain, namely: (i) the inherently complex nature of drug discovery, (ii) the bewildering number of promising computational and experimental techniques that have emerged in the past several years, and (iii) the limited availability of relevant protein sequence-function data for drug-like molecules. There is a need to focus on computational methods that will be most practically effective for protein drug discovery and on building experimental platforms to generate the data most appropriate for these methods. Here, we discuss recent advances in computational and experimental life sciences that are most crucial for impacting the pace and success of protein drug discovery.

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.

Reduction of monoclonal antibody viscosity using interpretable machine learning

Early identification of antibody candidates with drug-like properties is essential for simplifying the development of safe and effective antibody therapeutics. For subcutaneous administration, it is important to identify candidates with low self-association to enable their formulation at high concentration while maintaining low viscosity, opalescence, and aggregation. Here, we report an interpretable machine learning model for predicting antibody (IgG1) variants with low viscosity using only the sequences of their variable (Fv) regions. Our model was trained on antibody viscosity data (>100 mg/mL mAb concentration) obtained at a common formulation pH (pH 5.2), and it identifies three key Fv features of antibodies linked to viscosity, namely their isoelectric points, hydrophobic patch sizes, and numbers of negatively charged patches. Of the three features, most predicted antibodies at risk for high viscosity, including antibodies with diverse antibody germlines in our study (79 mAbs) as well as clinical-stage IgG1s (94 mAbs), are those with low Fv isoelectric points (Fv pIs < 6.3). Our model identifies viscous antibodies with relatively high accuracy not only in our training and test sets, but also for previously reported data. Importantly, we show that the interpretable nature of the model enables the design of mutations that significantly reduce antibody viscosity, which we confirmed experimentally. We expect that this approach can be readily integrated into the drug development process to reduce the need for experimental viscosity screening and improve the identification of antibody candidates with drug-like properties.

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.

Development of in silico models to predict viscosity and mouse clearance using a comprehensive analytical data set collected on 83 scaffold-consistent monoclonal antibodies

Biologic drug discovery pipelines are designed to deliver protein therapeutics that have exquisite functional potency and selectivity while also manifesting biophysical characteristics suitable for manufacturing, storage, and convenient administration to patients. The ability to use computational methods to predict biophysical properties from protein sequence, potentially in combination with high throughput assays, could decrease timelines and increase the success rates for therapeutic developability engineering by eliminating lengthy and expensive cycles of recombinant protein production and testing. To support development of high-quality predictive models for antibody developability, we designed a sequence-diverse panel of 83 effector functionless IgG1 antibodies displaying a range of biophysical properties, produced and formulated each protein under standard platform conditions, and collected a comprehensive package of analytical data, including in vitro assays and in vivo mouse pharmacokinetics. We used this robust training data set to build machine learning classifier models that can predict complex protein behavior from these data and features derived from predicted and/or experimental structures. Our models predict with 87% accuracy whether viscosity at 150 mg/mL is above or below a threshold of 15 centipoise (cP) and with 75% accuracy whether the area under the plasma drug concentration-time curve (AUC0-672 h) in normal mouse is above or below a threshold of 3.9 × 106 h x ng/mL.