Machine Learning for Biologics: Opportunities for Protein Engineering, Developability, and Formulation

Enzymatic CO2 reduction catalyzed by natural and artificial Metalloenzymes

The continuously increasing level of atmospheric CO2 in the atmosphere has led to global warming. Converting CO2 into other carbon compounds could mitigate its atmospheric levels and produce valuable products, as CO2 also serves as a plentiful and inexpensive carbon feedstock. However, the inert nature of CO2 poses a major challenge for its reduction. To meet the challenge, nature has evolved metalloenzymes using transition metal ions like Fe, Ni, Mo, and W, as well as electron-transfer partners for their functions. Mimicking these enzymes, artificial metalloenzymes (ArMs) have been designed using alternative protein scaffolds and various metallocofactors like Ni, Co, Re, Rh, and FeS clusters. Both the catalytic efficiency and the scope of CO2-reduction product of these ArMs have been improved over the past decade. This review first focuses on the natural metalloenzymes that directly reduce CO2 by discussing their structures and active sites, as well as the proposed reaction mechanisms. It then introduces the common strategies for electrochemical, photochemical, or photoelectrochemical utilization of these native enzymes for CO2 reduction and highlights the most recent advancements from the past five years. We also summarize principles of protein design for bio-inspired ArMs, comparing them with native enzymatic systems and outlining challenges and opportunities in enzymatic CO2 reduction.

Protein stability models fail to capture epistatic interactions of double point mutations

There is strong interest in accurate methods for predicting changes in protein stability resulting from amino acid mutations to the protein sequence. Recombinant proteins must often be stabilized to be used as therapeutics or reagents, and destabilizing mutations are implicated in a variety of diseases. Due to increased data availability and improved modeling techniques, recent studies have shown advancements in predicting changes in protein stability when a single point mutation is made. Less focus has been directed toward predicting changes in protein stability when there are two or more mutations, despite the significance of mutation clusters for disease pathways and protein design studies. Here, we analyze the largest available dataset of double point mutation stability and benchmark several widely used protein stability models on this and other datasets. We identify a blind spot in how predictors are typically evaluated on multiple mutations, finding that, contrary to assumptions in the field, current stability models are unable to consistently capture epistatic interactions between double mutations. We observe one notable deviation from this trend, which is that epistasis-aware models provide marginally better predictions on stabilizing double point mutations. We develop an extension of the ThermoMPNN framework for double mutant modeling as well as a novel data augmentation scheme which mitigates some of the limitations in available datasets. Collectively, our findings indicate that current protein stability models fail to capture the nuanced epistatic interactions between concurrent mutations due to several factors, including training dataset limitations and insufficient model sensitivity.

Significance

Protein stability is governed in part by epistatic interactions between energetically coupled residues. Prediction of these couplings represents the next frontier in protein stability modeling. In this work, we benchmark protein stability models on a large dataset of double point mutations and identify previously overlooked limitations in model design and evaluation. We also introduce several new strategies to improve modeling of epistatic couplings between protein point mutations.

Holistic in silico developability assessment of novel classes of small proteins using publicly available sequence-based predictors

The development of novel therapeutic proteins is a lengthy and costly process, with an average attrition rate of 91% (Thomas et al. Clinical Development Success Rates and Contributing Factors 2011-2020, 2021). To increase the probability of success and ensure robust drug supply beyond approval, it is essential to assess the developability profile of new potential drug candidates as early and broadly as possible in development (Jain et al. MAbs, 2023. https://doi.org/10.1016/j.copbio.2011.06.002 ). Predicting these properties in silico is expected to be the next leap in innovation as it would enable significantly reduced development timelines combined with broader screens at lower costs. However, developing predictive algorithms typically requires substantial datasets generated under very defined conditions, a limiting factor especially for new classes of therapeutic proteins that hold immense clinical promise. Here we describe a strategy for assessing the developability of a novel class of small therapeutic Anticalin® proteins using machine learning in conjunction with a knowledge-driven approach. The knowledge-driven approach considers developability attributes such as aggregation propensity, charge variants, immunogenicity, specificity, thermal stability, hydrophobicity, and potential post-translational modifications, to calculate a holistic developability score. Based on sequence-derived descriptors as input parameters we established novel statistical models designed to predict the developability scores for Anticalin proteins. The best models yielded low root mean square errors across the entire dataset and were further validated by removing input data from individual screening campaigns and predicting developability scores for those drug candidates. The adoption of the described workflow will enable significantly streamlined preclinical development of Anticalin drug candidates and could potentially be applied to other therapeutic protein scaffolds.

Machine learning: an advancement in biochemical engineering

One of the most remarkable techniques recently introduced into the field of bioprocess engineering is machine learning. Bioprocess engineering has drawn much attention due to its vast application in different domains like biopharmaceuticals, fossil fuel alternatives, environmental remediation, and food and beverage industry, etc. However, due to their unpredictable mechanisms, they are very often challenging to optimize. Furthermore, biological systems are extremely complicated; hence, machine learning algorithms could potentially be utilized to improve and build new biotechnological processes. Gaining insight into the fundamental mathematical understanding of commonly used machine learning algorithms, including Support Vector Machine, Principal Component Analysis, Partial Least Squares and Reinforcement Learning, the present study aims to discuss various case studies related to the application of machine learning in bioprocess engineering. Recent advancements as well as challenges posed in this area along with their potential solutions are also presented.

Best practices for machine learning in antibody discovery and development

In the past 40 years, therapeutic antibody discovery and development have advanced considerably, with machine learning (ML) offering a promising way to speed up the process by reducing costs and the number of experiments required. Recent progress in ML-guided antibody design and development (D&D) has been hindered by the diversity of data sets and evaluation methods, which makes it difficult to conduct comparisons and assess utility. Establishing standards and guidelines will be crucial for the wider adoption of ML and the advancement of the field. This perspective critically reviews current practices, highlights common pitfalls and proposes method development and evaluation guidelines for various ML-based techniques in therapeutic antibody D&D. Addressing challenges across the ML process, best practices are recommended for each stage to enhance reproducibility and progress.

Piston-driven automated liquid handlers

Laboratory capacities are often limited by time-consuming manual repetitive procedures rather than analysis time itself. While modern instruments are typically equipped with an autosampler, sample preparation often follows manual procedures including many labor-intensive, monotonous tasks. Particularly, for a high number of samples, well plates, and low microliter pipetting, manual preparation is error-prone often requiring repeated experiments. Sampling and sample preparation can account for greater analytical variability than instrument analysis. Repetitive tasks such as liquid handling benefit strongly from technological advances and led to the increasing applications of various automated liquid handlers (ALHs). In this review, we discuss the considerations for ALHs in the microliter range and highlight advantages and challenges when transforming from manual to automated workflows. We strongly focused on differences in liquid handling and outlined advantages due to sensor-controlled pipetting. ALHs can substantially improve costs-effectiveness and laboratory capacity. This is a consequence of increased efficiency, and throughput of laboratories while simultaneously raising data quality. Additionally, ALHs can improve safety, documentation of data, and sustainability. While automation requires careful consideration and resource demanding implementation, we believe it offers numerous advantages and can help to transform modern laboratories.

AttABseq: an attention-based deep learning prediction method for antigen-antibody binding affinity changes based on protein sequences

The optimization of therapeutic antibodies through traditional techniques, such as candidate screening via hybridoma or phage display, is resource-intensive and time-consuming. In recent years, computational and artificial intelligence-based methods have been actively developed to accelerate and improve the development of therapeutic antibodies. In this study, we developed an end-to-end sequence-based deep learning model, termed AttABseq, for the predictions of the antigen-antibody binding affinity changes connected with antibody mutations. AttABseq is a highly efficient and generic attention-based model by utilizing diverse antigen-antibody complex sequences as the input to predict the binding affinity changes of residue mutations. The assessment on the three benchmark datasets illustrates that AttABseq is 120% more accurate than other sequence-based models in terms of the Pearson correlation coefficient between the predicted and experimental binding affinity changes. Moreover, AttABseq also either outperforms or competes favorably with the structure-based approaches. Furthermore, AttABseq consistently demonstrates robust predictive capabilities across a diverse array of conditions, underscoring its remarkable capacity for generalization across a wide spectrum of antigen-antibody complexes. It imposes no constraints on the quantity of altered residues, rendering it particularly applicable in scenarios where crystallographic structures remain unavailable. The attention-based interpretability analysis indicates that the causal effects of point mutations on antibody-antigen binding affinity changes can be visualized at the residue level, which might assist automated antibody sequence optimization. We believe that AttABseq provides a fiercely competitive answer to therapeutic antibody optimization.

Intensified perfusion culture (IPC) reduced recombinant protein fragmentation

Mammalian cells remain the mainstay of biological production host. In industry, cultivating and harvest strategies are sorted in batch mode (e.g., batch, fed-batch, concentrated fed-batch and intensified fed-batch) and continuous mode (e.g., perfusion). To retrieve greater productivity and better product quality, especially for the sensitive products prone to fragmentation, culture modes with various modifications are innovated (e.g., intensified perfusion culture [IPC]). In our study, we demonstrated that the fragmentation of Fc-fusion product (Molecule A) is time-dependent in traditional fed-batch (TFB) culture. The fragmentation proportion increased from 3.8% to 12.4% for Clone A, 0.8% to 1.7% for Clone B and 0.9% to 2.0% for Clone C from Day 10 to Day 14. By applying a novel bioprocess, IPC, which allows continuous feeding of the fresh medium and constant removal of the spent medium without bleeding cells to maintain a defined constant viable cell density, the fragmentation was reduced to 0.3% while the productivity was increased from 2.96 g/L to 15.51 g/L for Clone A. To validate whether the fragmentation reduction is product-sensitive, plasmids carrying the DNA sequences of two other Fc-fusion molecules (Molecule B and Molecule C) were transfected into the host. The results showed consistent fragmentation reducing effect by using IPC. Furthermore, the cultivation scale was expanded to 50 L and 1000 L. A minimum fragmentation level below 0.1% was observed for Molecule C. Our study revealed the capability of IPC in reducing Fc-fusion protein fragmentation and the reproducibility when scaling up while maintaining high productivity.

Unlocking novel therapies: cyclic peptide design for amyloidogenic targets through synergies of experiments, simulations, and machine learning

Existing therapies for neurodegenerative diseases like Parkinson's and Alzheimer's address only their symptoms and do not prevent disease onset. Common therapeutic agents, such as small molecules and antibodies struggle with insufficient selectivity, stability and bioavailability, leading to poor performance in clinical trials. Peptide-based therapeutics are emerging as promising candidates, with successful applications for cardiovascular diseases and cancers due to their high bioavailability, good efficacy and specificity. In particular, cyclic peptides have a long in vivo stability, while maintaining a robust antibody-like binding affinity. However, the de novo design of cyclic peptides is challenging due to the lack of long-lived druggable pockets of the target polypeptide, absence of exhaustive conformational distributions of the target and/or the binder, unknown binding site, methodological limitations, associated constraints (failed trials, time, money) and the vast combinatorial sequence space. Hence, efficient alignment and cooperation between disciplines, and synergies between experiments and simulations complemented by popular techniques like machine-learning can significantly speed up the therapeutic cyclic-peptide development for neurodegenerative diseases. We review the latest advancements in cyclic peptide design against amyloidogenic targets from a computational perspective in light of recent advancements and potential of machine learning to optimize the design process. We discuss the difficulties encountered when designing novel peptide-based inhibitors and we propose new strategies incorporating experiments, simulations and machine learning to design cyclic peptides to inhibit the toxic propagation of amyloidogenic polypeptides. Importantly, these strategies extend beyond the mere design of cyclic peptides and serve as template for the de novo generation of (bio)materials with programmable properties.

Empirical validation of ProteinMPNN's efficiency in enhancing protein fitness

Introduction: Protein engineering, which aims to improve the properties and functions of proteins, holds great research significance and application value. However, current models that predict the effects of amino acid substitutions often perform poorly when evaluated for precision. Recent research has shown that ProteinMPNN, a large-scale pre-training sequence design model based on protein structure, performs exceptionally well. It is capable of designing mutants with structures similar to the original protein. When applied to the field of protein engineering, the diverse designs for mutation positions generated by this model can be viewed as a more precise mutation range. Methods: We collected three biological experimental datasets and compared the design results of ProteinMPNN for wild-type proteins with the experimental datasets to verify the ability of ProteinMPNN in improving protein fitness. Results: The validation on biological experimental datasets shows that ProteinMPNN has the ability to design mutation types with higher fitness in single and multi-point mutations. We have verified the high accuracy of ProteinMPNN in protein engineering tasks from both positive and negative perspectives. Discussion: Our research indicates that using large-scale pre trained models to design protein mutants provides a new approach for protein engineering, providing strong support for guiding biological experiments and applications in biotechnology.

Bridging artificial intelligence and fucoxanthin for the recovery and quantification from microalgae

Fucoxanthin is a carotenoid that possesses various beneficial medicinal properties for human well-being. However, the current extraction technologies and quantification techniques are still lacking in terms of cost validation, high energy consumption, long extraction time, and low yield production. To date, artificial intelligence (AI) models can assist and improvise the bottleneck of fucoxanthin extraction and quantification process by establishing new technologies and processes which involve big data, digitalization, and automation for efficiency fucoxanthin production. This review highlights the application of AI models such as artificial neural network (ANN) and adaptive neuro fuzzy inference system (ANFIS), capable of learning patterns and relationships from large datasets, capturing non-linearity, and predicting optimal conditions that significantly impact the fucoxanthin extraction yield. On top of that, combining metaheuristic algorithm such as genetic algorithm (GA) can further improve the parameter space and discovery of optimal conditions of ANN and ANFIS models, which results in high R2 accuracy ranging from 98.28% to 99.60% after optimization. Besides, AI models such as support vector machine (SVM), convolutional neural networks (CNNs), and ANN have been leveraged for the quantification of fucoxanthin, either computer vision based on color space of images or regression analysis based on statistical data. The findings are reliable when modeling for the concentration of pigments with high R2 accuracy ranging from 66.0% - 99.2%. This review paper has reviewed the feasibility and potential of AI for the extraction and quantification purposes, which can reduce the cost, accelerate the fucoxanthin yields, and development of fucoxanthin-based products.

Formulation Graphs for Mapping Structure-Composition of Battery Electrolytes to Device Performance

Advanced computational methods are being actively sought to address the challenges associated with the discovery and development of new combinatorial materials, such as formulations. A widely adopted approach involves domain-informed high-throughput screening of individual components that can be combined together to form a formulation. This manages to accelerate the discovery of new compounds for a target application but still leaves the process of identifying the right "formulation" from the shortlisted chemical space largely a laboratory experiment-driven process. We report a deep learning model, the Formulation Graph Convolution Network (F-GCN), that can map the structure-composition relationship of the formulation constituents to the property of liquid formulation as a whole. Multiple GCNs are assembled in parallel that featurize formulation constituents domain-intuitively on the fly. The resulting molecular descriptors are scaled based on the respective constituent's molar percentage in the formulation, followed by integration into a combined formulation descriptor that represents the complete formulation to an external learning architecture. The use case of the proposed formulation learning model is demonstrated for battery electrolytes by training and testing it on two exemplary data sets representing electrolyte formulations vs battery performance: one data set is sourced from the literature about Li/Cu half-cells, while the other is obtained by lab experiments related to lithium-iodide full-cell chemistry. The model is shown to predict performance metrics such as Coulombic efficiency (CE) and specific capacity of new electrolyte formulations with the lowest reported errors. The best-performing F-GCN model uses molecular descriptors derived from molecular graphs (GCNs) that are informed with HOMO-LUMO and electric moment properties of the molecules using a knowledge transfer technique.

Evaluating the chaos game representation of proteins for applications in machine learning models: prediction of antibody affinity and specificity as a case study

CONTEXT

Machine learning techniques are becoming increasingly important in the selection and optimization of therapeutic molecules, as well as for the selection of formulation components and the prediction of long-term stability. Compared to first-principle models, machine learning techniques are easier to implement, and can identify correlations that would be hard to describe at a mechanistic level, but strongly rely on high-quality input training data. Here, we evaluate the potential of the "chaos game" representation to provide input data for machine learning models. The chaos game is an algorithm originally developed for the production of fractal structures, and later on applied also to the representation of biological sequences, such as genes and proteins. Our results show that the combination of the chaos game representation with convolutional neural networks results in comparable accuracy to other machine learning approaches, thus indicating that chaos game representations could be a valid alternative to existing featurization strategies for machine learning models of biological sequences.

METHODS

We implement the chaos game in Python 3.8.10, and use it to produce fractal as well as novel expanding representations of protein sequences. We then feed the resulting images to a convolutional neural network, built in Python 3.8.10, using TensorFlow 2.9.1, Keras 2.9.0, and the scikit-learn 1.1.1 packages. We select as case study a recently published dataset for the antibody emibetuzumab, with the objective of co-optimizing antibodies variants with both high affinity and low non-specific binding.

Mutexa: A Computational Ecosystem for Intelligent Protein Engineering

Protein engineering holds immense promise in shaping the future of biomedicine and biotechnology. This Review focuses on our ongoing development of Mutexa, a computational ecosystem designed to enable "intelligent protein engineering". In this vision, researchers will seamlessly acquire sequences of protein variants with desired functions as biocatalysts, therapeutic peptides, and diagnostic proteins through a finely-tuned computational machine, akin to Amazon Alexa's role as a versatile virtual assistant. The technical foundation of Mutexa has been established through the development of a database that combines and relates enzyme structures and their respective functions (e.g., IntEnzyDB), workflow software packages that enable high-throughput protein modeling (e.g., EnzyHTP and LassoHTP), and scoring functions that map the sequence-structure-function relationship of proteins (e.g., EnzyKR and DeepLasso). We will showcase the applications of these tools in benchmarking the convergence conditions of enzyme functional descriptors across mutants, investigating protein electrostatics and cavity distributions in SAM-dependent methyltransferases, and understanding the role of nonelectrostatic dynamic effects in enzyme catalysis. Finally, we will conclude by addressing the future steps and fundamental challenges in our endeavor to develop new Mutexa applications that assist the identification of beneficial mutants in protein engineering.

Accelerating Biocatalysis Discovery with Machine Learning: A Paradigm Shift in Enzyme Engineering, Discovery, and Design

Emerging computational tools promise to revolutionize protein engineering for biocatalytic applications and accelerate the development timelines previously needed to optimize an enzyme to its more efficient variant. For over a decade, the benefits of predictive algorithms have helped scientists and engineers navigate the complexity of functional protein sequence space. More recently, spurred by dramatic advances in underlying computational tools, the promise of faster, cheaper, and more accurate enzyme identification, characterization, and engineering has catapulted terms such as artificial intelligence and machine learning to the must-have vocabulary in the field. This Perspective aims to showcase the current status of applications in pharmaceutical industry and also to discuss and celebrate the innovative approaches in protein science by highlighting their potential in selected recent developments and offering thoughts on future opportunities for biocatalysis. It also critically assesses the technology's limitations, unanswered questions, and unmet challenges.

Artificial intelligence-aided protein engineering: from topological data analysis to deep protein language models

Protein engineering is an emerging field in biotechnology that has the potential to revolutionize various areas, such as antibody design, drug discovery, food security, ecology, and more. However, the mutational space involved is too vast to be handled through experimental means alone. Leveraging accumulative protein databases, machine learning (ML) models, particularly those based on natural language processing (NLP), have considerably expedited protein engineering. Moreover, advances in topological data analysis (TDA) and artificial intelligence-based protein structure prediction, such as AlphaFold2, have made more powerful structure-based ML-assisted protein engineering strategies possible. This review aims to offer a comprehensive, systematic, and indispensable set of methodological components, including TDA and NLP, for protein engineering and to facilitate their future development.

Predicting and Interpreting Protein Developability Via Transfer of Convolutional Sequence Representation

Engineered proteins have emerged as novel diagnostics, therapeutics, and catalysts. Often, poor protein developability─quantified by expression, solubility, and stability─hinders utility. The ability to predict protein developability from amino acid sequence would reduce the experimental burden when selecting candidates. Recent advances in screening technologies enabled a high-throughput (HT) developability dataset for 105 of 1020 possible variants of protein ligand scaffold Gp2. In this work, we evaluate the ability of neural networks to learn a developability representation from a HT dataset and transfer this knowledge to predict recombinant expression beyond observed sequences. The model convolves learned amino acid properties to predict expression levels 44% closer to the experimental variance compared to a non-embedded control. Analysis of learned amino acid embeddings highlights the uniqueness of cysteine, the importance of hydrophobicity and charge, and the unimportance of aromaticity, when aiming to improve the developability of small proteins. We identify clusters of similar sequences with increased recombinant expression through nonlinear dimensionality reduction and we explore the inferred expression landscape via nested sampling. The analysis enables the first direct visualization of the fitness landscape and highlights the existence of evolutionary bottlenecks in sequence space giving rise to competing subpopulations of sequences with different developability. The work advances applied protein engineering efforts by predicting and interpreting protein scaffold expression from a limited dataset. Furthermore, our statistical mechanical treatment of the problem advances foundational efforts to characterize the structure of the protein fitness landscape and the amino acid characteristics that influence protein developability.

Position-Specific Enrichment Ratio Matrix scores predict antibody variant properties from deep sequencing data

MOTIVATION

Deep sequencing of antibody and related protein libraries after phage or yeast-surface display sorting is widely used to identify variants with increased affinity, specificity, and/or improvements in key biophysical properties. Conventional approaches for identifying optimal variants typically use the frequencies of observation in enriched libraries or the corresponding enrichment ratios. However, these approaches disregard the vast majority of deep sequencing data and often fail to identify the best variants in the libraries.

RESULTS

Here, we present a method, Position-Specific Enrichment Ratio Matrix (PSERM) scoring, that uses entire deep sequencing datasets from pre- and post-selections to score each observed protein variant. The PSERM scores are the sum of the site-specific enrichment ratios observed at each mutated position. We find that PSERM scores are much more reproducible and correlate more strongly with experimentally measured properties than frequencies or enrichment ratios, including for multiple antibody properties (affinity and non-specific binding) for a clinical-stage antibody (emibetuzumab). We expect that this method will be broadly applicable to diverse protein engineering campaigns.

AVAILABILITY AND IMPLEMENTATION

All deep sequencing datasets and code to perform the analyses presented within are available via https://github.com/Tessier-Lab-UMich/PSERM_paper.

Data-driven prediction models for forecasting multistep ahead profiles of mammalian cell culture toward bioprocess digital twins

Recently, the advancement in process analytical technology and artificial intelligence (AI) has enabled the generation of enormous culture data sets from biomanufacturing processes that produce various recombinant therapeutic proteins (RTPs), such as monoclonal antibodies (mAbs). Thus, now it is very important to exploit them for the enhanced reliability, efficiency, and consistency of the RTP-producing culture processes and for the reduced incipient or abrupt faults. It is achievable by AI-based data-driven models (DDMs), which allow us to correlate biological and process conditions and cell culture states. In this work, we provide practical guidelines for choosing the best combination of model elements to design and implement successful DDMs for given hypothetical in-line data sets during mAb-producing Chinese hamster ovary cell culture, as such enabling us to forecast dynamic behaviors of culture performance such as viable cell density, mAb titer as well as glucose, lactate and ammonia concentrations. To do so, we created DDMs that balance computational load with model accuracy and reliability by identifying the best combination of multistep ahead forecasting strategies, input features, and AI algorithms, which is potentially applicable to implementation of interactive DDM within bioprocess digital twins. We believe this systematic study can help bioprocess engineers start developing predictive DDMs with their own data sets and learn how their cell cultures behave in near future, thereby rendering proactive decision possible.

Integrated Multimodal Omics and Dietary Approaches for the Management of Neurodegeneration

Neurodegenerative diseases, such as Alzheimer's disease and Parkinson's disease, are caused by a combination of multiple events that damage neuronal function. A well-characterized biomarker of neurodegeneration is the accumulation of proteinaceous aggregates in the brain. However, the gradually worsening symptoms of neurodegenerative diseases are unlikely to be solely due to the result of a mutation in a single gene, but rather a multi-step process involving epigenetic changes. Recently, it has been suggested that a fraction of epigenetic alternations may be correlated to neurodegeneration in the brain. Unlike DNA mutations, epigenetic alterations are reversible, and therefore raise the possibilities for therapeutic intervention, including dietary modifications. Additionally, reactive oxygen species may contribute to the pathogenesis of Alzheimer's disease and Parkinson's disease through epigenetic alternation. Given that the antioxidant properties of plant-derived phytochemicals are likely to exhibit pleiotropic effects against ROS-mediated epigenetic alternation, dietary intervention may be promising for the management of neurodegeneration in these diseases. In this review, the state-of-the-art applications using single-cell multimodal omics approaches, including epigenetics, and dietary approaches for the identification of novel biomarkers and therapeutic approaches for the treatment of neurodegenerative diseases are discussed.

Review and perspective on bioinformatics tools using machine learning and deep learning for predicting antiviral peptides

Viruses constitute a constant threat to global health and have caused millions of human and animal deaths throughout human history. Despite advances in the discovery of antiviral compounds that help fight these pathogens, finding a solution to this problem continues to be a task that consumes time and financial resources. Currently, artificial intelligence (AI) has revolutionized many areas of the biological sciences, making it possible to decipher patterns in amino acid sequences that encode different functions and activities. Within the field of AI, machine learning, and deep learning algorithms have been used to discover antimicrobial peptides. Due to their effectiveness and specificity, antimicrobial peptides (AMPs) hold excellent promise for treating various infections caused by pathogens. Antiviral peptides (AVPs) are a specific type of AMPs that have activity against certain viruses. Unlike the research focused on the development of tools and methods for the prediction of antimicrobial peptides, those related to the prediction of AVPs are still scarce. Given the significance of AVPs as potential pharmaceutical options for human and animal health and the ongoing AI revolution, we have reviewed and summarized the current machine learning and deep learning-based tools and methods available for predicting these types of peptides.

Position-Specific Enrichment Ratio Matrix scores predict antibody variant properties from deep sequencing data

Motivation

Deep sequencing of antibody and related protein libraries after phage or yeast-surface display sorting is widely used to identify variants with increased affinity, specificity and/or improvements in key biophysical properties. Conventional approaches for identifying optimal variants typically use the frequencies of observation in enriched libraries or the corresponding enrichment ratios. However, these approaches disregard the vast majority of deep sequencing data and often fail to identify the best variants in the libraries.

Results

Here, we present a method, Position-Specific Enrichment Ratio Matrix (PSERM) scoring, that uses entire deep sequencing datasets from pre- and post-selections to score each observed protein variant. The PSERM scores are the sum of the site-specific enrichment ratios observed at each mutated position. We find that PSERM scores are much more reproducible and correlate more strongly with experimentally measured properties than frequencies or enrichment ratios, including for multiple antibody properties (affinity and non-specific binding) for a clinical-stage antibody (emibetuzumab). We expect that this method will be broadly applicable to diverse protein engineering campaigns.

Availability

All deep sequencing datasets and code to do the analyses presented within are available via GitHub.

Contact

Peter Tessier, ptessier@umich.edu.

Supplementary information

Supplementary data are available at Bioinformatics online.

The Evolution of Commercial Antibody Formulations

It has been nearly four decades since the first therapeutic monoclonal antibodies were approved and made available for widespread human use. Herein, US and EU approved antibody formulations are reviewed, and their nature and compositions are evaluated over time. From 1986 through Jan 2023, significant formulation trends have occurred and to represent this, 165 commercial antibody therapeutic formulations were binned into 5 different periods of time. Overall, we have observed the following: 1) The average formulation pH has decreased in recent years by over 0.5 units along with a decrease in variability that is largely driven by non-high concentration liquid in vial presentations for IV administration, 2) The use of certain excipients and buffers such as histidine, sucrose, metal chelators, arginine and methionine has become significantly more common, whereas formulations that contain phosphate, salt, no sugar or no surfactant have fallen out of favor, 3) Overall formulation space has increasingly become more homogenous and has converged in terms of formulation pH and excipient preferences regardless of formulation concentration, drug product presentation, and route of administration, 4) The average calculated isoelectric point (pI) has decreased 0.26 units, and 5) Overall, the average formulation pH and calculated pI for all commercial antibodies surveyed was 6.0 and 8.4, respectively. These trends and formulation convergence may be driven by multiple factors such as advancements in high-throughput computational and analytical technologies, the increased emphasis and understanding of certain developability attributes and formulation principles during lead selection and formulation development, and the adoption of low-risk development platform approaches.

Integrating machine learning and electrochemistry to develop a glucose biosensor assembled with Ganoderma applanatum lectin

Fungal lectins have enormous biotechnological potential, but limited knowledge about their biochemical and biophysical features prevents their proper use. Herein, we report an innovative alternative to use Ganoderma applanatum lectin (GAL) as a glucose biorecognition element, after identifying the ideal electroanalytical conditions by machine learning studies performed with a homologous agglutinin from the same macrofungus. The research revealed that GAL has moderate resistance to pH (4-8) and temperature (20-60 °C) variations, but its hemagglutinating activity (376.5 HU mg^-1 GAL at 20 °C) was better conserved under physiological conditions. Integrating electrochemical data and semi-empirical molecular modeling, biocompatible and electrostatically favorable conditions were found to immobilize the lectin on Prussian blue-modified glassy carbon electrode, after thermal activation of the metal-complex film. The glucose dose-response relationship obtained with the developed biosensor, defined as GAL/ta-PB/GCE, showed a typical Hill equation correlation, suggesting electrodic interactions represented by a sigmoidal mathematical function. GAL/ta-PB/GCE achieved remarkable electroanalytical performance, with emphasis on the detection limit (10.2 pM) and sensitivity (0.012 µA µM^-1cm^-2). The biosensor was successfully used to quantify glucose in pharmaceutical formulations, reiterating that the association of theoretical and experimental information drives important advances in bioelectrochemical studies.

The Dawn of a New Era: Targeting the "Undruggables" with Antibody-Based Therapeutics

The high selectivity and affinity of antibodies toward their antigens have made them a highly valuable tool in disease therapy, diagnosis, and basic research. A plethora of chemical and genetic approaches have been devised to make antibodies accessible to more "undruggable" targets and equipped with new functions of illustrating or regulating biological processes more precisely. In this Review, in addition to introducing how naked antibodies and various antibody conjugates (such as antibody-drug conjugates, antibody-oligonucleotide conjugates, antibody-enzyme conjugates, etc.) work in therapeutic applications, special attention has been paid to how chemistry tools have helped to optimize the therapeutic outcome (i.e., with enhanced efficacy and reduced side effects) or facilitate the multifunctionalization of antibodies, with a focus on emerging fields such as targeted protein degradation, real-time live-cell imaging, catalytic labeling or decaging with spatiotemporal control as well as the engagement of antibodies inside cells. With advances in modern chemistry and biotechnology, well-designed antibodies and their derivatives via size miniaturization or multifunctionalization together with efficient delivery systems have emerged, which have gradually improved our understanding of important biological processes and paved the way to pursue novel targets for potential treatments of various diseases.

Accelerating therapeutic protein design with computational approaches toward the clinical stage

Therapeutic protein, represented by antibodies, is of increasing interest in human medicine. However, clinical translation of therapeutic protein is still largely hindered by different aspects of developability, including affinity and selectivity, stability and aggregation prevention, solubility and viscosity reduction, and deimmunization. Conventional optimization of the developability with widely used methods, like display technologies and library screening approaches, is a time and cost-intensive endeavor, and the efficiency in finding suitable solutions is still not enough to meet clinical needs. In recent years, the accelerated advancement of computational methodologies has ushered in a transformative era in the field of therapeutic protein design. Owing to their remarkable capabilities in feature extraction and modeling, the integration of cutting-edge computational strategies with conventional techniques presents a promising avenue to accelerate the progression of therapeutic protein design and optimization toward clinical implementation. Here, we compared the differences between therapeutic protein and small molecules in developability and provided an overview of the computational approaches applicable to the design or optimization of therapeutic protein in several developability issues.

Advancing oral delivery of biologics: Machine learning predicts peptide stability in the gastrointestinal tract

The oral delivery of peptide therapeutics could facilitate precision treatment of numerous gastrointestinal (GI) and systemic diseases with simple administration for patients. However, the vast majority of licensed peptide drugs are currently administered parenterally due to prohibitive peptide instability in the GI tract. As such, the development of GI-stable peptides is receiving considerable investment. This study provides researchers with the first tool to predict the GI stability of peptide therapeutics based solely on the amino acid sequence. Both unsupervised and supervised machine learning techniques were trained on literature-extracted data describing peptide stability in simulated gastric and small intestinal fluid (SGF and SIF). Based on 109 peptide incubations, classification models for SGF and SIF were developed. The best models utilized k-Nearest Neighbor (for SGF) and XGBoost (for SIF) algorithms, with accuracies of 75.1% (SGF) and 69.3% (SIF), and f1 scores of 84.5% (SGF) and 73.4% (SIF) under 5-fold cross-validation. Feature importance analysis demonstrated that peptides' lipophilicity, rigidity, and size were key determinants of stability. These models are now available to those working on the development of oral peptide therapeutics.

High-Yield Production of Extracellular Vesicle Subpopulations with Constant Quality Using Batch-Refeed Cultures

The conventional manufacturing of extracellular vesicles (EVs) is characterized by low yields and batch-to-batch variability, hampering fundamental research on EVs and their practical applications. Perfusion operations have huge potential to address these limitations and increase the productivity and quality of EVs. In this study, perfusion cultures are simulated with batch-refeed systems and their productivity is compared with that achieved using batch cultures. It is shown that a shift from batch to batch-refeed system can increase the space-time yields of a target EV subpopulation characterized by CD81 and CD63 biomarkers by threefold. Moreover, it is demonstrated that the method facilitates the consistent production of the target EVs from cells maintained under constant conditions for 13 days. These results indicate that the use of perfusion cultures is a promising strategy to increase the manufacturing yield of EVs and control the production of specific EV subpopulations with constant quality attributes, thereby improving reproducibility.

Graph Neural Tree: A novel and interpretable deep learning-based framework for accurate molecular property predictions

Determining various properties of molecules is a critical step in drug discovery. Recently, with the improvement of large heterogeneous datasets and the development of deep learning approaches, more and more scientists have turned their attention to neural network-based virtual preliminary screening to reduce the time and monetary cost of drug discovery. However, the poor interpretability of deep learning masks causality, so models' conclusions are often beyond the comprehension of human users, which reduces the credibility of the model and makes it difficult for chemists to further narrow the huge chemical space based on models' results. Thus, this study develops a novel framework consisting of Graph Neural Networks for feature extraction, Curriculum-Based Learning Strategies for optimization, and a Learning Binary Neural Tree (LBNT) for prediction, to improve the performance of neural networks and reveal their decision-making process to chemists. The framework encodes molecular graph data with graph neural networks (GNNs), then retrains the encoder with curriculum-based learning strategies to reduce uncertainty and improve accuracy, and finally uses LBNT as the predictor, which joint retrains with the encoder after independently training, for prediction and visualization. The framework is validated on the public datasets and compared to single GNNs with normal training strategies as well as GNN encoders with common machine learning predictors instead of the LBNT predictor. The result reveals that the proposed framework enhances the point prediction accuracy of the completely trained GNN and reduces its uncertainty through curriculum-based learning, and further improves the accuracy by combining LBNT. Besides, compared with common machine learning tools, the LBNT predictor generally has the best performance because of joint retraining with the GNN encoder. The decision-making process of LBNT is also better and easier to explain than that of other models.

AF4 and PEG Precipitation as Predictive Assays for Antibody Self-Association

Monoclonal antibodies (mAbs) are often formulated as high-protein-concentration solutions, which in some cases can exhibit physical stability issues such as high viscosity and opalescence. To ensure that mAb-based drugs can meet their manufacturing, stability, and delivery requirements, it is advantageous to screen for and select mAbs during discovery that are not prone to such behaviors. It has been recently shown that both these macroscopic properties can be predicted to a certain extent from the diffusion interaction parameter (k_D), which is a measure of self-association under dilute conditions.¹ However, k_D can be challenging to measure at the early stage of discovery, where a relatively large amount of a high-purity material, which is required by traditional methods, is often not available. In this study, we demonstrate asymmetric field-flow fractionation (AF4) as a tool to measure self-association and therefore identify antibodies with problematic issues at high concentrations. The principle lies on the ability to concentrate the sample close to the membrane during the injection mode, which can reach formulation-relevant concentrations (>100 mg/mL).² By analyzing a well-characterized library of commercial antibodies, we show that the measured retention time of the antibodies allows us to pinpoint molecules that exhibit issues at high concentrations. Remarkably, our AF4 assay requires very little (30 μg) sample under dilute conditions and does not need extensive sample purification. Furthermore, we show that a polyethylene glycol (PEG) precipitation assay provides results consistent with AF4 and moreover can further differentiate molecules with issues of opalescence or high viscosity. Overall, our results delineate a two-step strategy for the identification of problematic variants at high concentrations, with AF4 for early developability screening, followed by a PEG assay to validate the problematic molecules and further discriminate between opalescence or high-viscosity issues. This two-step antibody selection strategy enables us to select antibodies early in the discovery process, which are compatible with high-concentration formulations.

Persistent spectral theory-guided protein engineering

While protein engineering, which iteratively optimizes protein fitness by screening the gigantic mutational space, is constrained by experimental capacity, various machine learning models have substantially expedited protein engineering. Three-dimensional protein structures promise further advantages, but their intricate geometric complexity hinders their applications in deep mutational screening. Persistent homology, an established algebraic topology tool for protein structural complexity reduction, fails to capture the homotopic shape evolution during the filtration of a given data. This work introduces a Topology-offered protein Fitness (TopFit) framework to complement protein sequence and structure embeddings. Equipped with an ensemble regression strategy, TopFit integrates the persistent spectral theory, a new topological Laplacian, and two auxiliary sequence embeddings to capture mutation-induced topological invariant, shape evolution, and sequence disparity in the protein fitness landscape. The performance of TopFit is assessed by 34 benchmark datasets with 128,634 variants, involving a vast variety of protein structure acquisition modalities and training set size variations.

Surface-Induced Protein Aggregation and Particle Formation in Biologics: Current Understanding of Mechanisms, Detection and Mitigation Strategies

Protein stability against aggregation is a major quality concern for the production of safe and effective biopharmaceuticals. Amongst the different drivers of protein aggregation, increasing evidence indicates that interactions between proteins and interfaces represent a major risk factor for the formation of protein aggregates in aqueous solutions. Potentially harmful surfaces relevant to biologics manufacturing and storage include air-water and silicone oil-water interfaces as well as materials from different processing units, storage containers, and delivery devices. The impact of some of these surfaces, for instance originating from impurities, can be difficult to predict and control. Moreover, aggregate formation may additionally be complicated by the simultaneous presence of interfacial, hydrodynamic and mechanical stresses, whose contributions may be difficult to deconvolute. As a consequence, it remains difficult to identify the key chemical and physical determinants and define appropriate analytical methods to monitor and predict protein instability at these interfaces. In this review, we first discuss the main mechanisms of surface-induced protein aggregation. We then review the types of contact materials identified as potentially harmful or detected as potential triggers of proteinaceous particle formation in formulations and discuss proposed mitigation strategies. Finally, we present current methods to probe surface-induced instabilities, which represent a starting point towards assays that can be implemented in early-stage screening and formulation development of biologics.

The evolutionary and functional significance of germline immunoglobulin gene variation

The recombination between immunoglobulin (IG) gene segments determines an individual's naïve antibody repertoire and, consequently, (auto)antigen recognition. Emerging evidence suggests that mammalian IG germline variation impacts humoral immune responses associated with vaccination, infection, and autoimmunity - from the molecular level of epitope specificity, up to profound changes in the architecture of antibody repertoires. These links between IG germline variants and immunophenotype raise the question on the evolutionary causes and consequences of diversity within IG loci. We discuss why the extreme diversity in IG loci remains a mystery, why resolving this is important for the design of more effective vaccines and therapeutics, and how recent evidence from multiple lines of inquiry may help us do so.

Approaches to expand the conventional toolbox for discovery and selection of antibodies with drug-like physicochemical properties

Antibody drugs should exhibit not only high-binding affinity for their target antigens but also favorable physicochemical drug-like properties. Such drug-like biophysical properties are essential for the successful development of antibody drug products. The traditional approaches used in antibody drug development require significant experimentation to produce, optimize, and characterize many candidates. Therefore, it is attractive to integrate new methods that can optimize the process of selecting antibodies with both desired target-binding and drug-like biophysical properties. Here, we summarize a selection of techniques that can complement the conventional toolbox used to de-risk antibody drug development. These techniques can be integrated at different stages of the antibody development process to reduce the frequency of physicochemical liabilities in antibody libraries during initial discovery and to co-optimize multiple antibody features during early-stage antibody engineering and affinity maturation. Moreover, we highlight biophysical and computational approaches that can be used to predict physical degradation pathways relevant for long-term storage and in-use stability to reduce the need for extensive experimentation.

Assessing developability early in the discovery process for novel biologics

Beyond potency, a good developability profile is a key attribute of a biological drug. Selecting and screening for such attributes early in the drug development process can save resources and avoid costly late-stage failures. Here, we review some of the most important developability properties that can be assessed early on for biologics. These include the influence of the source of the biologic, its biophysical and pharmacokinetic properties, and how well it can be expressed recombinantly. We furthermore present in silico, in vitro, and in vivo methods and techniques that can be exploited at different stages of the discovery process to identify molecules with liabilities and thereby facilitate the selection of the most optimal drug leads. Finally, we reflect on the most relevant developability parameters for injectable versus orally delivered biologics and provide an outlook toward what general trends are expected to rise in the development of biologics.

Blueprint for antibody biologics developability

Large-molecule antibody biologics have revolutionized medicine owing to their superior target specificity, pharmacokinetic and pharmacodynamic properties, safety and toxicity profiles, and amenability to versatile engineering. In this review, we focus on preclinical antibody developability, including its definition, scope, and key activities from hit to lead optimization and selection. This includes generation, computational and in silico approaches, molecular engineering, production, analytical and biophysical characterization, stability and forced degradation studies, and process and formulation assessments. More recently, it is apparent these activities not only affect lead selection and manufacturability, but ultimately correlate with clinical progression and success. Emerging developability workflows and strategies are explored as part of a blueprint for developability success that includes an overview of the four major molecular properties that affect all developability outcomes: 1) conformational, 2) chemical, 3) colloidal, and 4) other interactions. We also examine risk assessment and mitigation strategies that increase the likelihood of success for moving the right candidate into the clinic.

Unconstrained generation of synthetic antibody-antigen structures to guide machine learning methodology for antibody specificity prediction

Machine learning (ML) is a key technology for accurate prediction of antibody-antigen binding. Two orthogonal problems hinder the application of ML to antibody-specificity prediction and the benchmarking thereof: the lack of a unified ML formalization of immunological antibody-specificity prediction problems and the unavailability of large-scale synthetic datasets to benchmark real-world relevant ML methods and dataset design. Here we developed the Absolut! software suite that enables parameter-based unconstrained generation of synthetic lattice-based three-dimensional antibody-antigen-binding structures with ground-truth access to conformational paratope, epitope and affinity. We formalized common immunological antibody-specificity prediction problems as ML tasks and confirmed that for both sequence- and structure-based tasks, accuracy-based rankings of ML methods trained on experimental data hold for ML methods trained on Absolut!-generated data. The Absolut! framework has the potential to enable real-world relevant development and benchmarking of ML strategies for biotherapeutics design.

Non-specificity as the sticky problem in therapeutic antibody development

Antibodies are highly potent therapeutic scaffolds with more than a hundred different products approved on the market. Successful development of antibody-based drugs requires a trade-off between high target specificity and target binding affinity. In order to better understand this problem, we here review non-specific interactions and explore their fundamental physicochemical origins. We discuss the role of surface patches - clusters of surface-exposed amino acid residues with similar physicochemical properties - as inducers of non-specific interactions. These patches collectively drive interactions including dipole-dipole, π-stacking and hydrophobic interactions to complementary moieties. We elucidate links between these supramolecular assembly processes and macroscopic development issues, such as decreased physical stability and poor in vivo half-life. Finally, we highlight challenges and opportunities for optimizing protein binding specificity and minimizing non-specificity for future generations of therapeutics.

Developability assessment at early-stage discovery to enable development of antibody-derived therapeutics

Developability refers to the likelihood that an antibody candidate will become a manufacturable, safe and efficacious drug. Although the safety and efficacy of a drug candidate will be well considered by sponsors and regulatory agencies, developability in the narrow sense can be defined as the likelihood that an antibody candidate will go smoothly through the chemistry, manufacturing and control (CMC) process at a reasonable cost and within a reasonable timeline. Developability in this sense is the focus of this review. To lower the risk that an antibody candidate with poor developability will move to the CMC stage, the candidate's developability-related properties should be screened, assessed and optimized as early as possible. Assessment of developability at the early discovery stage should be performed in a rapid and high-throughput manner while consuming small amounts of testing materials. In addition to monoclonal antibodies, bispecific antibodies, multispecific antibodies and antibody-drug conjugates, as the derivatives of monoclonal antibodies, should also be assessed for developability. Moreover, we propose that the criterion of developability is relative: expected clinical indication, and the dosage and administration route of the antibody could affect this criterion. We also recommend a general screening process during the early discovery stage of antibody-derived therapeutics. With the advance of artificial intelligence-aided prediction of protein structures and features, computational tools can be used to predict, screen and optimize the developability of antibody candidates and greatly reduce the risk of moving a suboptimal candidate to the development stage.

CLADE 2.0: Evolution-Driven Cluster Learning-Assisted Directed Evolution

Directed evolution, a revolutionary biotechnology in protein engineering, optimizes protein fitness by searching an astronomical mutational space via expensive experiments. The cluster learning-assisted directed evolution (CLADE) efficiently explores the mutational space via a combination of unsupervised hierarchical clustering and supervised learning. However, the initial-stage sampling in CLADE treats all clusters equally despite many clusters containing a large portion of non-functional mutations. Recent statistical and deep learning tools enable evolutionary density modeling to access protein fitness in an unsupervised manner. In this work, we construct an ensemble of multiple evolutionary scores to guide the initial sampling in CLADE. The resulting evolutionary score-enhanced CLADE, called CLADE 2.0, efficiently selects a training set within a small informative space using the evolution-driven clustering sampling. CLADE 2.0 is validated by using two benchmark libraries both having 160,000 sequences from four-site mutational combinations. Extensive computational experiments and comparisons with existing cutting-edge methods indicate that CLADE 2.0 is a new state-of-art tool for machine learning-assisted directed evolution.

Protein scaffolds in human clinics

Fundamental clinical areas such as drug delivery and regenerative medicine require biocompatible materials as mechanically stable scaffolds or as nanoscale drug carriers. Among the wide set of emerging biomaterials, polypeptides offer enticing properties over alternative polymers, including full biocompatibility, biodegradability, precise interactivity, structural stability and conformational and functional versatility, all of them tunable by conventional protein engineering. However, proteins from non-human sources elicit immunotoxicities that might bottleneck further development and narrow their clinical applicability. In this context, selecting human proteins or developing humanized protein versions as building blocks is a strict demand to design non-immunogenic protein materials. We review here the expanding catalogue of human or humanized proteins tailored to execute different levels of scaffolding functions and how they can be engineered as self-assembling materials in form of oligomers, polymers or complex networks. In particular, we emphasize those that are under clinical development, revising their fields of applicability and how they have been adapted to offer, apart from mere mechanical support, highly refined functions and precise molecular interactions.

Structure-based rational design of hydroxysteroid dehydrogenases for improving and diversifying steroid synthesis

A group of steroidogenic enzymes, hydroxysteroid dehydrogenases are involved in steroid metabolism which is very important in the cell: signaling, growth, reproduction, and energy homeostasis. The enzymes show an inherent function in the interconversion of ketosteroids and hydroxysteroids in a position- and stereospecific manner on the steroid nucleus and side-chains. However, the biocatalysis of steroids reaction is a vital and demanding, yet challenging, task to produce the desired enantiopure products with non-natural substrates or non-natural cofactors, and/or in non-physiological conditions. This has driven the use of protein design strategies to improve their inherent biosynthetic efficiency or activate their silent catalytic ability. In this review, the innate features and catalytic characteristics of enzymes based on sequence-structure-function relationships of steroidogenic enzymes are reviewed. Combining structure information and catalytic mechanisms, progress in protein redesign to stimulate potential function, for example, substrate specificity, cofactor dependence, and catalytic stability are discussed.

Conformational Entropy as a Potential Liability of Computationally Designed Antibodies

In silico antibody discovery is emerging as a viable alternative to traditional in vivo and in vitro approaches. Many challenges, however, remain open to enabling the properties of designed antibodies to match those produced by the immune system. A major question concerns the structural features of computer-designed complementarity determining regions (CDRs), including the role of conformational entropy in determining the stability and binding affinity of the designed antibodies. To address this problem, we used enhanced-sampling molecular dynamics simulations to compare the free energy landscapes of single-domain antibodies (sdAbs) designed using structure-based (DesAb-HSA-D3) and sequence-based approaches (DesAbO), with that of a nanobody derived from llama immunization (Nb10). Our results indicate that the CDR3 of DesAbO is more conformationally heterogeneous than those of both DesAb-HSA-D3 and Nb10, and the CDR3 of DesAb-HSA-D3 is slightly more dynamic than that of Nb10, which is the original scaffold used for the design of DesAb-HSA-D3. These differences underline the challenges in the rational design of antibodies by revealing the presence of conformational substates likely to have different binding properties and to generate a high entropic cost upon binding.

Holistic Process Models: A Bayesian Predictive Ensemble Method for Single and Coupled Unit Operation Models

The coupling of individual models in terms of end-to-end calculations for unit operations in manufacturing processes is a challenging task. We present a probability distribution-based approach for the combined outcomes of parametric and non-parametric models. With this so-called Bayesian predictive ensemble, the statistical moments such as mean value and standard deviation can be accurately computed without any further approximation. It is shown that the ensemble of different model predictions leads to an uninformed prior distribution, which can be transformed into a predictive posterior distribution using Bayesian inference and numerical Markov Chain Monte Carlo calculations. We demonstrate the advantages of our method using several numerical examples. Our approach is not restricted to certain unit operations, and can also be used for the more robust interpretation and assessment of model predictions in general.

A Descriptor set for Quantitative Structure-Property Relationship Prediction in Biologics

In addition to attaining the desired binding to their targets, a crucial aspect in the development of biotherapeutics is 'developability', which includes several desirable properties such as high solubility, low viscosity and aggregation, physico-chemical stability and low immunogenicity. The lack of any of these properties can lead to significant obstacles in advancing them to clinic; thus in silico methods capable of raising warning flags in earlier stages of development are highly beneficial. We have developed a computational framework based on a large and diverse set of protein specific descriptors ideal for making liability predictions using a machine-learning approach. This set offers a high degree of feature diversity classifiable by sequence, structure and surface patches. We assess the sensitivity and applicability of these descriptors in four dedicated case studies that are believed to be representative of biophysical characterizations commonly employed during the development process. In addition to data sets obtained from public sources, we have validated the descriptors on novel experimental data sets in order to address antibody developability and to generate prospective predictions on Adnectins. The results demonstrate that the descriptors are well suited to assist in the improvement of properties of systems that exhibit poor solubility or aggregation.

Rapid Characterization and Quantification of Extracellular Vesicles by Fluorescence-Based Microfluidic Diffusion Sizing

Extracellular vesicles (EVs) are emerging as promising diagnostic and therapeutic tools for a variety of diseases. The characterization of EVs requires a series of orthogonal techniques that are overall time- and material-consuming. Here, a microfluidic device is presented that exploits the combination of diffusion sizing and multiwavelength fluorescence detection to simultaneously provide information on EV size, concentration, and composition. The latter is achieved with the nonspecific staining of lipids and proteins combined with the specific staining of EV markers such as EV-associated tetraspanins via antibodies. The device can be operated as a single-step immunoassay thanks to the integrated separation and quantification of free and EV-bound fluorophores. This microfluidic technique is capable of detecting and quantifying components associated to EV subtypes and impurities and thus to measure EV purity in a time scale of minutes, requiring less than 5 µL of sample and minimal sample handling before the analysis. Moreover, the analysis is performed directly in solution without immobilization steps. Therefore, this method can accelerate screening of EV samples and aid the evaluation of sample reproducibility, representing an important complementary tool to the current array of biophysical methods for EV characterization, particularly valuable for instance for bioprocess development.

Advancing pharmacy and healthcare with virtual digital technologies

Digitalisation of the healthcare sector promises to revolutionise patient healthcare globally. From the different technologies, virtual tools including artificial intelligence, blockchain, virtual, and augmented reality, to name but a few, are providing significant benefits to patients and the pharmaceutical sector alike, ranging from improving access to clinicians and medicines, as well as improving real-time diagnoses and treatments. Indeed, it is envisioned that such technologies will communicate together in real-time, as well as with their physical counterparts, to create a large-scale, cyber healthcare system. Despite the significant benefits that virtual-based digital health technologies can bring to patient care, a number of challenges still remain, ranging from data security to acceptance within the healthcare sector. This review provides a timely account of the benefits and challenges of virtual health interventions, as well an outlook on how such technologies can be transitioned from research-focused towards real-world healthcare and pharmaceutical applications to transform treatment pathways for patients worldwide.

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.