Deep learning guided optimization of human antibody against SARS-CoV-2 variants with broad neutralization

Clinical microbiology and artificial intelligence: Different applications, challenges, and future prospects

Conventional clinical microbiological techniques are enhanced by the introduction of artificial intelligence (AI). Comprehensive data processing and analysis enabled the development of curated datasets that has been effectively used in training different AI algorithms. Recently, a number of machine learning (ML) and deep learning (DL) algorithms are developed and evaluated using diverse microbiological datasets. These datasets included spectral analysis (Raman and MALDI-TOF spectroscopy), microscopic images (Gram and acid fast stains), and genomic and protein sequences (whole genome sequencing (WGS) and protein data banks (PDBs)). The primary objective of these algorithms is to minimize the time, effort, and expenses linked to conventional analytical methods. Furthermore, AI algorithms are incorporated with quantitative structure-activity relationship (QSAR) models to predict novel antimicrobial agents that address the continuing surge of antimicrobial resistance. During the COVID-19 pandemic, AI algorithms played a crucial role in vaccine developments and the discovery of new antiviral agents, and introduced potential drug candidates via drug repurposing. However, despite their significant benefits, the implementation of AI encounters various challenges, including ethical considerations, the potential for bias, and errors related to data training. This review seeks to provide an overview of the most recent applications of artificial intelligence in clinical microbiology, with the intention of educating a wider audience of clinical practitioners regarding the current uses of machine learning algorithms and encouraging their implementation. Furthermore, it will discuss the challenges related to the incorporation of AI into clinical microbiology laboratories and examine future opportunities for AI within the realm of infectious disease epidemiology.

Computational electrostatic engineering of nanobodies for enhanced SARS-CoV-2 receptor binding domain recognition

This study presents a novel computational approach for engineering nanobodies (Nbs) for improved interaction with receptor-binding domain (RBD) of the SARS-CoV-2 spike protein. Using Protein Structure Reliability reports, RBD (7VYR_R) was selected and refined for subsequent Nb-RBD interactions. By leveraging electrostatic complementarity (EC) analysis, we engineered and characterized five Electrostatically Complementary Nbs (ECSb1-ECSb5) based on the CeVICA library's SR6c3 Nb. Through targeted modifications in the complementarity-determining regions (CDR) and framework regions (FR), we optimized electrostatic interactions to improve binding affinity and specificity. The engineered Nbs (ECSb3, ECSb4, and ECSb5) demonstrated high binding specificity for AS3, CA1, and CA2 epitopes. Interestingly, ECSb1 and ECSb2 selectively engaged with AS3 and CA1 instead of AS1 and AS2, respectively, due to a preference for residues that conferred superior binding complementarities. Furthermore, ECSbs significantly outperformed SR6c3 Nb in MM/GBSA results, notably, ECSb4 and ECSb3 exhibited superior binding free energies of -182.58 kcal.mol-1 and -119.07 kcal.mol-1, respectively, compared to SR6c3 (-105.50 kcal.mol-1). ECSbs exhibited significantly higher thermostability (100.4-148.3 kcal·mol⁻1) compared to SR6c3 (62.6 kcal·mol⁻1). Similarly, enhanced electrostatic complementarity was also observed for ECSb4-RBD and ECSb3-RBD (0.305 and 0.390, respectively) relative to SR6c3-RBD (0.233). Surface analyses confirmed optimized electrostatic patches and reduced aggregation propensity in the engineered Nb. This integrated EC and structural engineering approach successfully developed engineered Nbs with enhanced binding specificity, increased thermostability, and reduced aggregation, laying the groundwork for novel therapeutic applications targeting the SARS-CoV-2 spike protein.

$\mathcal{S}$ able: bridging the gap in protein structure understanding with an empowering and versatile pre-training paradigm

Protein pre-training has emerged as a transformative approach for solving diverse biological tasks. While many contemporary methods focus on sequence-based language models, recent findings highlight that protein sequences alone are insufficient to capture the extensive information inherent in protein structures. Recognizing the crucial role of protein structure in defining function and interactions, we introduce $\mathcal{S}$able, a versatile pre-training model designed to comprehensively understand protein structures. $\mathcal{S}$able incorporates a novel structural encoding mechanism that enhances inter-atomic information exchange and spatial awareness, combined with robust pre-training strategies and lightweight decoders optimized for specific downstream tasks. This approach enables $\mathcal{S}$able to consistently outperform existing methods in tasks such as generation, classification, and regression, demonstrating its superior capability in protein structure representation. The code and models can be accessed via GitHub repository at https://github.com/baaihealth/Sable.

Dynamics of an SVEIR transmission model with protection awareness and two strains

As of May 2024, the main strains of COVID-19 caused hundreds of millions of infection cases and millions of deaths worldwide. In this study, we consider the COVID-19 epidemics with the main strains in the Chinese mainland. We study complex interactions among hosts, non-pharmaceutical interventions, and vaccinations for the main strains by a differential equation model called SVEIR. The disease transmission model incorporates two strains and protection awareness of the susceptible population. Results of this study show that the protection awareness plays a crucial role against infection of the population, and that the vaccines are effective against the circulation of the earlier strains, but ineffective for emerging strains. By using the next generation matrix method, the basic reproduction number of the SVEIR model is firstly obtained. Our analysis by Hurwitz criterion and LaSalle's invariance principle shows that the disease free-equilibrium point is locally and globally asymptotically stable when the threshold value is below one. The existences of endemic equilibrium points are also established, and the global asymptotic stabilities are analyzed using the Lyapunov function method. Further, the SVEIR model is confirmed to satisfy the principle of competitive exclusion, of which the strain with the larger value of the basic reproduction number is dominant. Numerically, the surveillance data with the Omicron strain and the XBB strain are split by the cubic spline interpolation method. The fitting curves against the surveillance data are plotted using the least-squares method from MATLAB. The results indicate that the XBB strain dominates in this study. Moreover, a global sensitivity analysis of the key parameters is performed by using of PRCC. The numerical simulations imply that combination control strategy positively impacts on the infection scale than what separate control strategy does, and that the earlier time producing protection awareness for the public creates less infection scale, further that the increment of protection awareness also reduces the infection scale. Therefore, the policymakers of the local government are suggested to concern the changes of protection awareness of the public.

Decoding the effects of mutation on protein interactions using machine learning

Accurately predicting mutation-caused binding free energy changes (ΔΔGs) on protein interactions is crucial for understanding how genetic variations affect interactions between proteins and other biomolecules, such as proteins, DNA/RNA, and ligands, which are vital for regulating numerous biological processes. Developing computational approaches with high accuracy and efficiency is critical for elucidating the mechanisms underlying various diseases, identifying potential biomarkers for early diagnosis, and developing targeted therapies. This review provides a comprehensive overview of recent advancements in predicting the impact of mutations on protein interactions across different interaction types, which are central to understanding biological processes and disease mechanisms, including cancer. We summarize recent progress in predictive approaches, including physicochemical-based, machine learning, and deep learning methods, evaluating the strengths and limitations of each. Additionally, we discuss the challenges related to the limitations of mutational data, including biases, data quality, and dataset size, and explore the difficulties in developing accurate prediction tools for mutation-induced effects on protein interactions. Finally, we discuss future directions for advancing these computational tools, highlighting the capabilities of advancing technologies, such as artificial intelligence to drive significant improvements in mutational effects prediction.

Rewiring protein sequence and structure generative models to enhance protein stability prediction

Predicting changes in protein thermostability due to amino acid substitutions is essential for understanding human diseases and engineering useful proteins for clinical and industrial applications. While recent advances in protein generative models, which learn probability distributions over amino acids conditioned on structural or evolutionary sequence contexts, have shown impressive performance in predicting various protein properties without task-specific training, their strong unsupervised prediction ability does not extend to all protein functions. In particular, their potential to improve protein stability prediction remains underexplored. In this work, we present SPURS, a novel deep learning framework that adapts and integrates two general-purpose protein generative models-a protein language model (ESM) and an inverse folding model (ProteinMPNN)-into an effective stability predictor. SPURS employs a lightweight neural network module to rewire per-residue structure representations learned by ProteinMPNN into the attention layers of ESM, thereby informing and enhancing ESM's sequence representation learning. This rewiring strategy enables SPURS to harness evolutionary patterns from both sequence and structure data, where the sequence like-lihood distribution learned by ESM is conditioned on structure priors encoded by ProteinMPNN to predict mutation effects. We steer this integrated framework to a stability prediction model through supervised training on a recently released mega-scale thermostability dataset. Evaluations across 12 benchmark datasets showed that SPURS delivers accurate, rapid, scalable, and generalizable stability predictions, consistently outperforming current state-of-the-art methods. Notably, SPURS demonstrates remarkable versatility in protein stability and function analyses: when combined with a protein language model, it accurately identifies protein functional sites in an unsupervised manner. Additionally, it enhances current low- N protein fitness prediction models by serving as a stability prior model to improve accuracy. These results highlight SPURS as a powerful tool to advance current protein stability prediction and machine learning-guided protein engineering workflows. The source code of SPURS is available at https://github.com/luo-group/SPURS .

Predicting Antibody Affinity Changes upon Mutation Based on Unbound Protein Structures

Antibodies are key proteins in the immune system that can reversibly and non-covalently bind specifically to their corresponding antigens, forming antigen-antibody complexes. They play a crucial role in recognizing foreign or self-antigens during the adaptive immune response. Monoclonal antibodies have emerged as a promising class of biological macromolecule therapeutics with broad market prospects. In the process of antibody drug development, a key engineering challenge is to improve the affinity of candidate antibodies, without experimentally resolved structures of the antigen-antibody complexes as input for computer-aided predictive methods. In this work, we present an approach for predicting the effect of residue mutations on antibody affinity without the structures of the antigen-antibody complexes. The method involves the graph representation of proteins and utilizes a pre-trained encoder. The encoder captures the residue-level microenvironment of the target residue on the antibody along with the antigen context pre- and post-mutation. The encoder inherently possesses the potential to identify paratope residues. In addition, we curated a benchmark dataset specifically for mutations of the antibody. Compared to baseline methods based on complex structures and sequences, our approach achieves superior or comparable average accuracy on benchmark datasets. Additionally, we validate its advantage of not requiring antigen-antibody complex structures as input for predicting the effects of mutations in antibodies against SARS-CoV-2, influenza, and human cytomegalovirus. Our method shows its potential for identifying mutations that improve antibody affinity in practical antibody engineering applications.

Artificial intelligence for modelling infectious disease epidemics

Infectious disease threats to individual and public health are numerous, varied and frequently unexpected. Artificial intelligence (AI) and related technologies, which are already supporting human decision making in economics, medicine and social science, have the potential to transform the scope and power of infectious disease epidemiology. Here we consider the application to infectious disease modelling of AI systems that combine machine learning, computational statistics, information retrieval and data science. We first outline how recent advances in AI can accelerate breakthroughs in answering key epidemiological questions and we discuss specific AI methods that can be applied to routinely collected infectious disease surveillance data. Second, we elaborate on the social context of AI for infectious disease epidemiology, including issues such as explainability, safety, accountability and ethics. Finally, we summarize some limitations of AI applications in this field and provide recommendations for how infectious disease epidemiology can harness most effectively current and future developments in AI.

Learning the language of antibody hypervariability

Protein language models (PLMs) have demonstrated impressive success in modeling proteins. However, general-purpose "foundational" PLMs have limited performance in modeling antibodies due to the latter's hypervariable regions, which do not conform to the evolutionary conservation principles that such models rely on. In this study, we propose a transfer learning framework called Antibody Mutagenesis-Augmented Processing (AbMAP), which fine-tunes foundational models for antibody-sequence inputs by supervising on antibody structure and binding specificity examples. Our learned feature representations accurately predict mutational effects on antigen binding, paratope identification, and other key antibody properties. We experimentally validate AbMAP for antibody optimization by applying it to refine a set of antibodies that bind to a SARS-CoV-2 peptide, and obtain an 82% hit-rate and up to 22-fold increase in binding affinity. AbMAP also unlocks large-scale analyses of immune repertoires, revealing that B-cell receptor repertoires of individuals, while remarkably different in sequence, converge toward similar structural and functional coverage. Importantly, AbMAP's transfer learning approach can be readily adapted to advances in foundational PLMs. We anticipate AbMAP will accelerate the efficient design and modeling of antibodies, expedite the discovery of antibody-based therapeutics, and deepen our understanding of humoral immunity.

PPB-Affinity: Protein-Protein Binding Affinity dataset for AI-based protein drug discovery

Prediction of protein-protein binding (PPB) affinity plays an important role in large-molecular drug discovery. Deep learning (DL) has been adopted to predict the changes of PPB binding affinities upon mutations, but there was a scarcity of studies predicting the PPB affinity itself. The major reason is the paucity of open-source dataset with PPB affinity data. To address this gap, the current study introduced a large comprehensive PPB affinity (PPB-Affinity) dataset. The PPB-Affinity dataset contains key information such as crystal structures of protein-protein complexes (with or without protein mutation patterns), PPB affinity, receptor protein chain, ligand protein chain, etc. To the best of our knowledge, this is the largest publicly available PPB affinity dataset, and we believe it will significantly advance drug discovery by streamlining the screening of potential large-molecule drugs. We also developed a deep-learning benchmark model with this dataset to predict the PPB affinity, providing a foundational comparison for the research community.

A deep learning approach predicting the activity of COVID-19 therapeutics and vaccines against emerging variants

Understanding which viral variants evade neutralization is crucial for improving antibody-based treatments, especially with rapidly evolving viruses like SARS-CoV-2. Yet, conventional assays are labor intensive and cannot capture the full spectrum of variants. We present a deep learning approach to predict changes in neutralizing antibody activity of COVID-19 therapeutics and vaccine-elicited sera/plasma against emerging viral variants. Our approach leverages data of 67,885 unique SARS-CoV-2 Spike sequences and 7,069 in vitro assays. The resulting model accurately predicted fold changes in neutralizing activity (R2 = 0.77) for a test set (N = 980) of data collected up to eight months after the training data. Next, the model was used to predict changes in activity of current therapeutic and vaccine-induced antibodies against emerging SARS-CoV-2 lineages. Consistent with other work, we found significantly reduced activity against newer XBB descendants, notably EG.5, FL.1.5.1, and XBB.1.16; primarily attributed to the F456L spike mutation.

Production of a monoclonal antibody targeting the SARS-CoV-2 Omicron spike protein and analysis of SARS-CoV-2 Omicron mutations related to monoclonal antibody resistance

SARS-CoV-2 mutations have resulted in the emergence of multiple concerning variants, with Omicron being the dominant strain presently. Therefore, we developed a monoclonal antibody (mAb) against the spike (S) protein of SARS-CoV-2 Omicron for therapeutic applications. We established the 1E3H12 mAb, recognizing the receptor binding domain (RBD) of the Omicron S protein, and found that the 1E3H12 mAb can efficiently recognize the Omicron S protein with weak affinity to the Alpha, Beta, and Mu variants, but not to the parental strain and Delta variant. Based on in vitro assays, the mAb demonstrated neutralizing activity against Omicron BA.1, BA.4/5, BQ.1.1, and XBB. A humanized antibody was further produced and proved to have neutralizing activity. To verify the potential limitations of the 1E3H12 mAb due to viral escape of SARS-CoV-2 Omicron variants, we analyzed the emergence of variants by whole genome deep sequencing after serial passage in cell culture. The results showed a few unique S protein mutations in the genome associated with resistance to the mAb. These findings suggest that this antibody not only contributes to the therapeutic arsenal against COVID-19 but also addresses the ongoing challenge of antibody resistance among the evolving subvariants of SARS-CoV-2 Omicron.

Single-cell spatiotemporal analysis reveals alveolar dendritic cell-T cell immunity hubs defending against pulmonary infection

How immune cells are spatiotemporally coordinated in the lung to effectively monitor, respond to, and resolve infection and inflammation in primed form needs to be fully illustrated. Here we apply immunocartography, a high-resolution technique that integrates spatial and single-cell RNA sequencing (scRNA-seq) through deconvolution and co-localization analyses, to the SARS-CoV-2-infected Syrian hamster model. We generate a comprehensive transcriptome map of the whole process of pulmonary infection from physiological condition, infection initiation, severe pneumonia to natural recovery at organ scale and single-cell resolution, with 142,965 cells and 45 lung lobes from 25 hamsters at 5 time points. Integrative analysis identifies that alveolar dendritic cell-T cell immunity hubs, where Ccr7+Ido1+ dendritic cells, Cd160+Cd8+ T cells, and Tnfrsf4+Cd4+ T cells physiologically co-localize, rapidly expand during SARS-CoV-2 infection, eliminate SARS-CoV-2 with the aid of Slamf9+ macrophages, and then restore to physiological levels after viral clearance. We verify the presence of these cell subpopulations in the immunity hubs in normal and SARS-CoV-2-infected hACE2 mouse models, as well as in publicly available human scRNA-seq datasets, demonstrating the potential broad relevance of our findings in lung immunity.

From chaos to clarity: The scientometric breakthrough in COVID-19 research

BACKGROUND

The COVID-19 pandemic paralyzed the world for over three years, generating unprecedented social changes in recent human history.

AIMS

We aimed to scientometrically summarize a global and temporal overview of publications on COVID-19 in the two worst years of the pandemic and its progression in early 2022, after the start of vaccination.

METHODS

Using the Web of Science database, this review covered the period from late 2019 to March 2022 and included all publications identified using the following terms: "SARS-CoV-2", "COVID-19", "Coronavirus Disease 19", and "2019-nCoV". We retrieved 268,904 publications, with evident global spreading, demonstrating that the pandemic triggered worldwide scientific research efforts.

RESULTS

Within the dataset, 195 countries have published about Covid-19. In initial publications, a solid trend in genotyping, sequencing, and detection of the virus was evident; however, in the development of the pandemic, new knowledge and research focus gained relevance, with continental solid trends, revealed by the keywords sustainability (eastern Europe); material sciences (Asia); public and mental health (Africa); information sciences (western Europe); education (Latin America). It identified high-impact research, mainly on diagnosis and vaccines, but also equally essential topics for returning life to the new normal, such as mental health, education, and remote work. The world experienced a highly transmissible infection that proved how fragile we are regarding organization and society.

CONCLUSIONS

It is necessary to learn from such an event and establish a protocol of actions and measures to be taken and avoided in a health emergency, aiming to act differently from the chaos experienced during the pandemic. Following the One Health approach, humanity must be aware of the need for more sustainable attitudes, given the inseparability of human beings from the environment.

Pretrainable geometric graph neural network for antibody affinity maturation

Increasing the binding affinity of an antibody to its target antigen is a crucial task in antibody therapeutics development. This paper presents a pretrainable geometric graph neural network, GearBind, and explores its potential in in silico affinity maturation. Leveraging multi-relational graph construction, multi-level geometric message passing and contrastive pretraining on mass-scale, unlabeled protein structural data, GearBind outperforms previous state-of-the-art approaches on SKEMPI and an independent test set. A powerful ensemble model based on GearBind is then derived and used to successfully enhance the binding of two antibodies with distinct formats and target antigens. ELISA EC50 values of the designed antibody mutants are decreased by up to 17 fold, and KD values by up to 6.1 fold. These promising results underscore the utility of geometric deep learning and effective pretraining in macromolecule interaction modeling tasks.

Novel antibody language model accelerates IgG screening and design for broad-spectrum antiviral therapy

Therapeutic antibodies have become one of the most influential therapeutics in modern medicine to fight against infectious pathogens, cancer, and many other diseases. However, experimental screening for highly efficacious targeting antibodies is labor-intensive and of high cost, which is exacerbated by evolving antigen targets under selective pressure such as fast-mutating viral variants. As a proof-of-concept, we developed a machine learning-assisted antibody generation pipeline AbGen that greatly accelerates the screening and re-design of immunoglobulins G (IgGs) against a broad spectrum of SARS-CoV-2 coronavirus variant strains. Our AbGen centers around a novel antibody language model (AbLM) that is pretrained on 12 million generic protein domain sequences and fine-tuned on 4,000+ paired VH-VL sequences, with IgG-specific CDR-masking and VH-VL cross-attention. AbLM provides a latent space of IgG sequence embeddings for AbGen, including (a) landscapes of IgGs' activities in neutralizing the wild-type virus are analyzed through structure prediction for IgG and IgG-antigen (viral protein spike's receptor binding domain, RBD) interactions; and (b) landscapes of IgGs' susceptibility in neutralizing variant viruses are predicted through Gaussian process regression, despite that as few as 14 clinical antibodies' responses to variants of concern are available. The AbGen pipeline was applied to over 1300 IgG sequences we collected from RBD-binding B cells of convalescent patients. With experimental validations, AbGen efficiently prioritized IgG candidates against a broad spectrum of viral variants (wildtype, Delta, and Omicron), preventing the infection of host cells in vitro and hACE2 transgenic mice in vivo. Compared to other existing protein language models that require 10-100 times more model parameters, AbLM improved the precision from around 50% to 75% to predict IgGs with low variant susceptibility. Furthermore, AbGen enables structure-based computational protein redesign for selected IgG clones with single amino acid substitutions at the RBD-binding interface that doubled the IgG blockade efficacy for one of the severe, therapy-resistant strains - Delta (B.1.617). Our work expedites applications of artificial intelligence in antibody screen and re-design combining data-driven protein language models and Kriging for antibody sequence analysis and activity prediction, in synergy with physics-driven protein docking and design for antibody-antigen interface analyses and functional optimization.

De novo generation of SARS-CoV-2 antibody CDRH3 with a pre-trained generative large language model

Artificial Intelligence (AI) techniques have made great advances in assisting antibody design. However, antibody design still heavily relies on isolating antigen-specific antibodies from serum, which is a resource-intensive and time-consuming process. To address this issue, we propose a Pre-trained Antibody generative large Language Model (PALM-H3) for the de novo generation of artificial antibodies heavy chain complementarity-determining region 3 (CDRH3) with desired antigen-binding specificity, reducing the reliance on natural antibodies. We also build a high-precision model antigen-antibody binder (A2binder) that pairs antigen epitope sequences with antibody sequences to predict binding specificity and affinity. PALM-H3-generated antibodies exhibit binding ability to SARS-CoV-2 antigens, including the emerging XBB variant, as confirmed through in-silico analysis and in-vitro assays. The in-vitro assays validate that PALM-H3-generated antibodies achieve high binding affinity and potent neutralization capability against spike proteins of SARS-CoV-2 wild-type, Alpha, Delta, and the emerging XBB variant. Meanwhile, A2binder demonstrates exceptional predictive performance on binding specificity for various epitopes and variants. Furthermore, by incorporating the attention mechanism inherent in the Roformer architecture into the PALM-H3 model, we improve its interpretability, providing crucial insights into the fundamental principles of antibody design.

Locuaz: an in silico platform for protein binders optimization

MOTIVATION

Engineering high-affinity binders targeting specific antigenic determinants remains a challenging and often daunting task, requiring extensive experimental screening. Computational methods have the potential to accelerate this process, reducing costs and time, but only if they demonstrate broad applicability and efficiency in exploring mutations, evaluating affinity, and pruning unproductive mutation paths.

RESULTS

In response to these challenges, we introduce a new computational platform for optimizing protein binders towards their targets. The platform is organized as a series of modules, performing mutation selection and application, molecular dynamics simulations to sample conformations around interaction poses, and mutation prioritization using suitable scoring functions. Notably, the platform supports parallel exploration of different mutation streams, enabling in silico high-throughput screening on High Performance Computing (HPC) systems. Furthermore, the platform is highly customizable, allowing users to implement their own protocols.

AVAILABILITY AND IMPLEMENTATION

The source code is available at https://github.com/pgbarletta/locuaz and documentation is at https://locuaz.readthedocs.io/. The data underlying this article are available at https://github.com/pgbarletta/suppl_info_locuaz.

Machine-learning-assisted high-throughput identification of potent and stable neutralizing antibodies against all four dengue virus serotypes

Several computational methods have been developed to identify neutralizing antibodies (NAbs) covering four dengue virus serotypes (DENV-1 to DENV-4); however, limitations of the dataset and the resulting performance remain. Here, we developed a new computational framework to predict potent and stable NAbs against DENV-1 to DENV-4 using only antibody (CDR-H3) and epitope sequences as input. Specifically, our proposed computational framework employed sequence-based ML and molecular dynamic simulation (MD) methods to achieve more accurate identification. First, we built a novel dataset (n = 1108) by compiling the interactions of CDR-H3 and epitope sequences with the half maximum inhibitory concentration (IC50) values, which represent neutralizing activities. Second, we achieved an accurately predictive ML model that showed high AUC values of 0.879 and 0.885 by tenfold cross-validation and independent tests, respectively. Finally, our computational framework could be applied to filter approximately 2.5 million unseen antibodies into two final candidates that showed strong and stable binding to all four serotypes. In addition, the most potent and stable candidate (1B3B9_V21) was evaluated for its development potential as a therapeutic agent by molecular docking and MD simulations. This study provides an antibody computational approach to facilitate the high-throughput identification of NAbs and accelerate the development of therapeutic antibodies.

Cmai: Predicting Antigen-Antibody Interactions from Massive Sequencing Data

The interaction between antigens and antibodies (B cell receptors, BCRs) is the key step underlying the function of the humoral immune system in various biological contexts. The capability to profile the landscape of antigen-binding affinity of a vast number of BCRs will provide a powerful tool to reveal novel insights at unprecedented levels and will yield powerful tools for translational development. However, current experimental approaches for profiling antibody-antigen interactions are costly and time-consuming, and can only achieve low-to-mid throughput. On the other hand, bioinformatics tools in the field of antibody informatics mostly focus on optimization of antibodies given known binding antigens, which is a very different research question and of limited scope. In this work, we developed an innovative Artificial Intelligence tool, Cmai, to address the prediction of the binding between antibodies and antigens that can be scaled to high-throughput sequencing data. Cmai achieved an AUROC of 0.91 in our validation cohort. We devised a biomarker metric based on the output from Cmai applied to high-throughput BCR sequencing data. We found that, during immune-related adverse events (irAEs) caused by immune-checkpoint inhibitor (ICI) treatment, the humoral immunity is preferentially responsive to intracellular antigens from the organs affected by the irAEs. In contrast, extracellular antigens on malignant tumor cells are inducing B cell infiltrations, and the infiltrating B cells have a greater tendency to co-localize with tumor cells expressing these antigens. We further found that the abundance of tumor antigen-targeting antibodies is predictive of ICI treatment response. Overall, Cmai and our biomarker approach filled in a gap that is not addressed by current antibody optimization works nor works such as AlphaFold3 that predict the structures of complexes of proteins that are known to bind.

Equivariant Line Graph Neural Network for Protein-Ligand Binding Affinity Prediction

Binding affinity prediction of three-dimensional (3D) protein-ligand complexes is critical for drug repositioning and virtual drug screening. Existing approaches usually transform a 3D protein-ligand complex to a two-dimensional (2D) graph, and then use graph neural networks (GNNs) to predict its binding affinity. However, the node and edge features of the 2D graph are extracted based on invariant local coordinate systems of the 3D complex. As a result, these approaches can not fully learn the global information of the complex, such as the physical symmetry and the topological information of bonds. To address these issues, we propose a novel Equivariant Line Graph Network (ELGN) for binding affinity prediction of 3D protein-ligand complexes. The proposed ELGN firstly adds a super node to the 3D complex, and then builds a line graph based on the 3D complex. After that, ELGN uses a new E(3)-equivariant network layer to pass the messages between nodes and edges based on the global coordinate system of the 3D complex. Experimental results on two real datasets demonstrate the effectiveness of ELGN over several state-of-the-art baselines.

DDAffinity: predicting the changes in binding affinity of multiple point mutations using protein 3D structure

MOTIVATION

Mutations are the crucial driving force for biological evolution as they can disrupt protein stability and protein-protein interactions which have notable impacts on protein structure, function, and expression. However, existing computational methods for protein mutation effects prediction are generally limited to single point mutations with global dependencies, and do not systematically take into account the local and global synergistic epistasis inherent in multiple point mutations.

RESULTS

To this end, we propose a novel spatial and sequential message passing neural network, named DDAffinity, to predict the changes in binding affinity caused by multiple point mutations based on protein 3D structures. Specifically, instead of being on the whole protein, we perform message passing on the k-nearest neighbor residue graphs to extract pocket features of the protein 3D structures. Furthermore, to learn global topological features, a two-step additive Gaussian noising strategy during training is applied to blur out local details of protein geometry. We evaluate DDAffinity on benchmark datasets and external validation datasets. Overall, the predictive performance of DDAffinity is significantly improved compared with state-of-the-art baselines on multiple point mutations, including end-to-end and pre-training based methods. The ablation studies indicate the reasonable design of all components of DDAffinity. In addition, applications in nonredundant blind testing, predicting mutation effects of SARS-CoV-2 RBD variants, and optimizing human antibody against SARS-CoV-2 illustrate the effectiveness of DDAffinity.

AVAILABILITY AND IMPLEMENTATION

DDAffinity is available at https://github.com/ak422/DDAffinity.

Accurate prediction of CDR-H3 loop structures of antibodies with deep learning

Accurate prediction of the structurally diverse complementarity determining region heavy chain 3 (CDR-H3) loop structure remains a primary and long-standing challenge for antibody modeling. Here, we present the H3-OPT toolkit for predicting the 3D structures of monoclonal antibodies and nanobodies. H3-OPT combines the strengths of AlphaFold2 with a pre-trained protein language model and provides a 2.24 Å average RMSDCα between predicted and experimentally determined CDR-H3 loops, thus outperforming other current computational methods in our non-redundant high-quality dataset. The model was validated by experimentally solving three structures of anti-VEGF nanobodies predicted by H3-OPT. We examined the potential applications of H3-OPT through analyzing antibody surface properties and antibody-antigen interactions. This structural prediction tool can be used to optimize antibody-antigen binding and engineer therapeutic antibodies with biophysical properties for specialized drug administration route.

Antibody design using deep learning: from sequence and structure design to affinity maturation

Deep learning has achieved impressive results in various fields such as computer vision and natural language processing, making it a powerful tool in biology. Its applications now encompass cellular image classification, genomic studies and drug discovery. While drug development traditionally focused deep learning applications on small molecules, recent innovations have incorporated it in the discovery and development of biological molecules, particularly antibodies. Researchers have devised novel techniques to streamline antibody development, combining in vitro and in silico methods. In particular, computational power expedites lead candidate generation, scaling and potential antibody development against complex antigens. This survey highlights significant advancements in protein design and optimization, specifically focusing on antibodies. This includes various aspects such as design, folding, antibody-antigen interactions docking and affinity maturation.

Protein Engineering with Lightweight Graph Denoising Neural Networks

Protein engineering faces challenges in finding optimal mutants from a massive pool of candidate mutants. In this study, we introduce a deep-learning-based data-efficient fitness prediction tool to steer protein engineering. Our methodology establishes a lightweight graph neural network scheme for protein structures, which efficiently analyzes the microenvironment of amino acids in wild-type proteins and reconstructs the distribution of the amino acid sequences that are more likely to pass natural selection. This distribution serves as a general guidance for scoring proteins toward arbitrary properties on any order of mutations. Our proposed solution undergoes extensive wet-lab experimental validation spanning diverse physicochemical properties of various proteins, including fluorescence intensity, antigen-antibody affinity, thermostability, and DNA cleavage activity. More than 40% of ProtLGN-designed single-site mutants outperform their wild-type counterparts across all studied proteins and targeted properties. More importantly, our model can bypass the negative epistatic effect to combine single mutation sites and form deep mutants with up to seven mutation sites in a single round, whose physicochemical properties are significantly improved. This observation provides compelling evidence of the structure-based model's potential to guide deep mutations in protein engineering. Overall, our approach emerges as a versatile tool for protein engineering, benefiting both the computational and bioengineering communities.

Leveraging conformal prediction to annotate enzyme function space with limited false positives

Machine learning (ML) is increasingly being used to guide biological discovery in biomedicine such as prioritizing promising small molecules in drug discovery. In those applications, ML models are used to predict the properties of biological systems, and researchers use these predictions to prioritize candidates as new biological hypotheses for downstream experimental validations. However, when applied to unseen situations, these models can be overconfident and produce a large number of false positives. One solution to address this issue is to quantify the model's prediction uncertainty and provide a set of hypotheses with a controlled false discovery rate (FDR) pre-specified by researchers. We propose CPEC, an ML framework for FDR-controlled biological discovery. We demonstrate its effectiveness using enzyme function annotation as a case study, simulating the discovery process of identifying the functions of less-characterized enzymes. CPEC integrates a deep learning model with a statistical tool known as conformal prediction, providing accurate and FDR-controlled function predictions for a given protein enzyme. Conformal prediction provides rigorous statistical guarantees to the predictive model and ensures that the expected FDR will not exceed a user-specified level with high probability. Evaluation experiments show that CPEC achieves reliable FDR control, better or comparable prediction performance at a lower FDR than existing methods, and accurate predictions for enzymes under-represented in the training data. We expect CPEC to be a useful tool for biological discovery applications where a high yield rate in validation experiments is desired but the experimental budget is limited.

Structure-Based Optimization of One Neutralizing Antibody against SARS-CoV-2 Variants Bearing the L452R Mutation

Neutralizing antibodies (nAbs) play an important role against SARS-CoV-2 infections. Previously, we have reported one potent receptor binding domain (RBD)-binding nAb Ab08 against the SARS-CoV-2 prototype and a panel of variants, but Ab08 showed much less efficacy against the variants harboring the L452R mutation. To overcome the antibody escape caused by the L452R mutation, we generated several structure-based Ab08 derivatives. One derivative, Ab08-K99E, displayed the mostly enhanced neutralizing potency against the Delta pseudovirus bearing the L452R mutation compared to the Ab08 and other derivatives. Ab08-K99E also showed improved neutralizing effects against the prototype, Omicron BA.1, and Omicron BA.4/5 pseudoviruses. In addition, compared to the original Ab08, Ab08-K99E exhibited high binding properties and affinities to the RBDs of the prototype, Delta, and Omicron BA.4/5 variants. Altogether, our findings report an optimized nAb, Ab08-K99E, against SARS-CoV-2 variants and demonstrate structure-based optimization as an effective way for antibody development against pathogens.

Benchmarking the accuracy of structure-based binding affinity predictors on Spike-ACE2 deep mutational interaction set

Since the start of COVID-19 pandemic, a huge effort has been devoted to understanding the Spike (SARS-CoV-2)-ACE2 recognition mechanism. To this end, two deep mutational scanning studies traced the impact of all possible mutations across receptor binding domain (RBD) of Spike and catalytic domain of human ACE2. By concentrating on the interface mutations of these experimental data, we benchmarked six commonly used structure-based binding affinity predictors (FoldX, EvoEF1, MutaBind2, SSIPe, HADDOCK, and UEP). These predictors were selected based on their user-friendliness, accessibility, and speed. As a result of our benchmarking efforts, we observed that none of the methods could generate a meaningful correlation with the experimental binding data. The best correlation is achieved by FoldX (R = -0.51). When we simplified the prediction problem to a binary classification, that is, whether a mutation is enriching or depleting the binding, we showed that the highest accuracy is achieved by FoldX with a 64% success rate. Surprisingly, on this set, simple energetic scoring functions performed significantly better than the ones using extra evolutionary-based terms, as in Mutabind and SSIPe. Furthermore, we demonstrated that recent AI approaches, mmCSM-PPI and TopNetTree, yielded comparable performances to the force field-based techniques. These observations suggest plenty of room to improve the binding affinity predictors in guessing the variant-induced binding profile changes of a host-pathogen system, such as Spike-ACE2. To aid such improvements we provide our benchmarking data at https://github.com/CSB-KaracaLab/RBD-ACE2-MutBench with the option to visualize our mutant models at https://rbd-ace2-mutbench.github.io/.

PLMSearch: Protein language model powers accurate and fast sequence search for remote homology

Homologous protein search is one of the most commonly used methods for protein annotation and analysis. Compared to structure search, detecting distant evolutionary relationships from sequences alone remains challenging. Here we propose PLMSearch (Protein Language Model), a homologous protein search method with only sequences as input. PLMSearch uses deep representations from a pre-trained protein language model and trains the similarity prediction model with a large number of real structure similarity. This enables PLMSearch to capture the remote homology information concealed behind the sequences. Extensive experimental results show that PLMSearch can search millions of query-target protein pairs in seconds like MMseqs2 while increasing the sensitivity by more than threefold, and is comparable to state-of-the-art structure search methods. In particular, unlike traditional sequence search methods, PLMSearch can recall most remote homology pairs with dissimilar sequences but similar structures. PLMSearch is freely available at https://dmiip.sjtu.edu.cn/PLMSearch .

Integrating physics in deep learning algorithms: a force field as a PyTorch module

MOTIVATION

Deep learning algorithms applied to structural biology often struggle to converge to meaningful solutions when limited data is available, since they are required to learn complex physical rules from examples. State-of-the-art force-fields, however, cannot interface with deep learning algorithms due to their implementation.

RESULTS

We present MadraX, a forcefield implemented as a differentiable PyTorch module, able to interact with deep learning algorithms in an end-to-end fashion.

AVAILABILITY AND IMPLEMENTATION

MadraX documentation, together with tutorials and installation guide, is available at madrax.readthedocs.io.

Leveraging Artificial Intelligence to Expedite Antibody Design and Enhance Antibody-Antigen Interactions

This perspective sheds light on the transformative impact of recent computational advancements in the field of protein therapeutics, with a particular focus on the design and development of antibodies. Cutting-edge computational methods have revolutionized our understanding of protein-protein interactions (PPIs), enhancing the efficacy of protein therapeutics in preclinical and clinical settings. Central to these advancements is the application of machine learning and deep learning, which offers unprecedented insights into the intricate mechanisms of PPIs and facilitates precise control over protein functions. Despite these advancements, the complex structural nuances of antibodies pose ongoing challenges in their design and optimization. Our review provides a comprehensive exploration of the latest deep learning approaches, including language models and diffusion techniques, and their role in surmounting these challenges. We also present a critical analysis of these methods, offering insights to drive further progress in this rapidly evolving field. The paper includes practical recommendations for the application of these computational techniques, supplemented with independent benchmark studies. These studies focus on key performance metrics such as accuracy and the ease of program execution, providing a valuable resource for researchers engaged in antibody design and development. Through this detailed perspective, we aim to contribute to the advancement of antibody design, equipping researchers with the tools and knowledge to navigate the complexities of this field.

Efficient evolution of human antibodies from general protein language models

Natural evolution must explore a vast landscape of possible sequences for desirable yet rare mutations, suggesting that learning from natural evolutionary strategies could guide artificial evolution. Here we report that general protein language models can efficiently evolve human antibodies by suggesting mutations that are evolutionarily plausible, despite providing the model with no information about the target antigen, binding specificity or protein structure. We performed language-model-guided affinity maturation of seven antibodies, screening 20 or fewer variants of each antibody across only two rounds of laboratory evolution, and improved the binding affinities of four clinically relevant, highly mature antibodies up to sevenfold and three unmatured antibodies up to 160-fold, with many designs also demonstrating favorable thermostability and viral neutralization activity against Ebola and severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) pseudoviruses. The same models that improve antibody binding also guide efficient evolution across diverse protein families and selection pressures, including antibiotic resistance and enzyme activity, suggesting that these results generalize to many settings.

Advancing Antibody Engineering through Synthetic Evolution and Machine Learning

Abs are versatile molecules with the potential to achieve exceptional binding to target Ags, while also possessing biophysical properties suitable for therapeutic drug development. Protein display and directed evolution systems have transformed synthetic Ab discovery, engineering, and optimization, vastly expanding the number of Ab clones able to be experimentally screened for binding. Moreover, the burgeoning integration of high-throughput screening, deep sequencing, and machine learning has further augmented in vitro Ab optimization, promising to accelerate the design process and massively expand the Ab sequence space interrogated. In this Brief Review, we discuss the experimental and computational tools employed in synthetic Ab engineering and optimization. We also explore the therapeutic challenges posed by developing Abs for infectious diseases, and the prospects for leveraging machine learning-guided protein engineering to prospectively design Abs resistant to viral escape.

Structural insights into broadly neutralizing antibodies elicited by hybrid immunity against SARS-CoV-2

ABSTRACTIncreasing spread by SARS-CoV-2 Omicron variants challenges existing vaccines and broadly reactive neutralizing antibodies (bNAbs) against COVID-19. Here we determine the diversity, potency, breadth and structural insights of bNAbs derived from memory B cells of BNT162b2-vaccinee after homogeneous Omicron BA.1 breakthrough infection. The infection activates diverse memory B cell clonotypes for generating potent class I/II and III bNAbs with new epitopes mapped to the receptor-binding domain (RBD). The top eight bNAbs neutralize wildtype and BA.1 potently but display divergent IgH/IgL sequences and neuralization profiles against other variants of concern (VOCs). Two of them (P2D9 and P3E6) belonging to class III NAbs display comparable potency against BA.4/BA.5, although structural analysis reveals distinct modes of action. P3E6 neutralizes all variants tested through a unique bivalent interaction with two RBDs. Our findings provide new insights into hybrid immunity on BNT162b2-induced diverse memory B cells in response to Omicron breakthrough infection for generating diverse bNAbs with distinct structural basis.

Passive antibody therapy in emerging infectious diseases

The epidemic of corona virus disease 2019 (COVID-19) caused by severe acute respiratory syndrome Coronavirus 2 and its variants of concern (VOCs) has been ongoing for over 3 years. Antibody therapies encompassing convalescent plasma, hyperimmunoglobulin, and neutralizing monoclonal antibodies (mAbs) applied in passive immunotherapy have yielded positive outcomes and played a crucial role in the early COVID-19 treatment. In this review, the development path, action mechanism, clinical research results, challenges, and safety profile associated with the use of COVID-19 convalescent plasma, hyperimmunoglobulin, and mAbs were summarized. In addition, the prospects of applying antibody therapy against VOCs was assessed, offering insights into the coping strategies for facing new infectious disease outbreaks.

Structural basis for broad neutralization of human antibody against Omicron sublineages and evasion by XBB variant

IMPORTANCE

The ongoing COVID-19 pandemic has been characterized by the emergence of new SARS-CoV-2 variants including the highly transmissible Omicron XBB sublineages, which have shown significant resistance to neutralizing antibodies (nAbs). This resistance has led to decreased vaccine effectiveness and therefore result in breakthrough infections and reinfections, which continuously threaten public health. To date, almost all available therapeutic nAbs, including those authorized under Emergency Use Authorization nAbs that were previously clinically useful against early strains, have recently been found to be ineffective against newly emerging variants. In this study, we provide a comprehensive structural basis about how the Class 3 nAbs, including 1G11 in this study and noted LY-CoV1404, are evaded by the newly emerged SARS-CoV-2 variants.

Enhancing antibody affinity through experimental sampling of non-deleterious CDR mutations predicted by machine learning

The application of machine learning (ML) models to optimize antibody affinity to an antigen is gaining prominence. Unfortunately, the small and biased nature of the publicly available antibody-antigen interaction datasets makes it challenging to build an ML model that can accurately predict binding affinity changes due to mutations (ΔΔG). Recognizing these inherent limitations, we reformulated the problem to ask whether an ML model capable of classifying deleterious vs non-deleterious mutations can guide antibody affinity maturation in a practical setting. To test this hypothesis, we developed a Random Forest classifier (Antibody Random Forest Classifier or AbRFC) with expert-guided features and integrated it into a computational-experimental workflow. AbRFC effectively predicted non-deleterious mutations on an in-house validation dataset that is free of biases seen in the publicly available training datasets. Furthermore, experimental screening of a limited number of predictions from the model (<10^2 designs) identified affinity-enhancing mutations in two unrelated SARS-CoV-2 antibodies, resulting in constructs with up to 1000-fold increased binding to the SARS-COV-2 RBD. Our findings indicate that accurate prediction and screening of non-deleterious mutations using machine learning offers a powerful approach to improving antibody affinity.

Machine Learning-Guided Protein Engineering

Recent progress in engineering highly promising biocatalysts has increasingly involved machine learning methods. These methods leverage existing experimental and simulation data to aid in the discovery and annotation of promising enzymes, as well as in suggesting beneficial mutations for improving known targets. The field of machine learning for protein engineering is gathering steam, driven by recent success stories and notable progress in other areas. It already encompasses ambitious tasks such as understanding and predicting protein structure and function, catalytic efficiency, enantioselectivity, protein dynamics, stability, solubility, aggregation, and more. Nonetheless, the field is still evolving, with many challenges to overcome and questions to address. In this Perspective, we provide an overview of ongoing trends in this domain, highlight recent case studies, and examine the current limitations of machine learning-based methods. We emphasize the crucial importance of thorough experimental validation of emerging models before their use for rational protein design. We present our opinions on the fundamental problems and outline the potential directions for future research.

Modeling the competitive transmission of the Omicron strain and Delta strain of COVID-19

Since November 2021, there have been cases of COVID-19's Omicron strain spreading in competition with Delta strains in many parts of the world. To explore how these two strains developed in this competitive spread, a new compartmentalized model was established. First, we analyzed the fundamental properties of the model, obtained the expression of the basic reproduction number, proved the local and global asymptotic stability of the disease-free equilibrium. Then by means of the cubic spline interpolation method, we obtained the data of new Omicron and Delta cases in the United States of new cases starting from December 8, 2021, to February 12, 2022. Using the weighted nonlinear least squares estimation method, we fitted six time series (cumulative confirmed cases, cumulative deaths, new cases, new deaths, new Omicron cases, and new Delta cases), got estimates of the unknown parameters, and obtained an approximation of the basic reproduction number in the United States during this time period as R 0 ≈ 1.5165 . Finally, each control strategy was evaluated by cost-effectiveness analysis to obtain the optimal control strategy under different perspectives. The results not only show the competitive transmission characteristics of the new strain and existing strain, but also provide scientific suggestions for effectively controlling the spread of these strains.

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.

Assessment of neutralization susceptibility of Omicron subvariants XBB.1.5 and BQ.1.1 against broad-spectrum neutralizing antibodies through epitopes mapping

The emergence of new variants of the SARS-CoV-2 virus has posed a significant challenge in developing broadly neutralizing antibodies (nAbs) with guaranteed therapeutic potential. Some nAbs, such as Sotrovimab, have exhibited varying levels of efficacy against different variants, while others, such as Bebtelovimab and Bamlanivimab-etesevimab are ineffective against specific variants, including BQ.1.1 and XBB. This highlights the urgent need for developing broadly active monoclonal antibodies (mAbs) providing prophylactic and therapeutic benefits to high-risk patients, especially in the face of the risk of reinfection from new variants. Here, we aimed to investigate the feasibility of redirecting existing mAbs against new variants of SARS-CoV-2, as well as to understand how BQ.1.1 and XBB.1.5 can evade broadly neutralizing mAbs. By mapping epitopes and escape sites, we discovered that the new variants evade multiple mAbs, including FDA-approved Bebtelovimab, which showed resilience against other Omicron variants. Our approach, which included simulations, endpoint free energy calculation, and shape complementarity analysis, revealed the possibility of identifying mAbs that are effective against both BQ.1.1 and XBB.1.5. We identified two broad-spectrum mAbs, R200-1F9 and R207-2F11, as potential candidates with increased binding affinity to XBB.1.5 and BQ.1.1 compared to the reference (Wu01) strain. Additionally, we propose that these mAbs do not interfere with Angiotensin Converting Enzyme 2 (ACE2) and bind to conserved epitopes on the receptor binding domain of Spike that are not-overlapping, potentially providing a solution to neutralize these new variants either independently or as part of a combination (cocktail) treatment.

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.

Deep-learning-enabled protein-protein interaction analysis for prediction of SARS-CoV-2 infectivity and variant evolution

Host-pathogen interactions and pathogen evolution are underpinned by protein-protein interactions between viral and host proteins. An understanding of how viral variants affect protein-protein binding is important for predicting viral-host interactions, such as the emergence of new pathogenic SARS-CoV-2 variants. Here we propose an artificial intelligence-based framework called UniBind, in which proteins are represented as a graph at the residue and atom levels. UniBind integrates protein three-dimensional structure and binding affinity and is capable of multi-task learning for heterogeneous biological data integration. In systematic tests on benchmark datasets and further experimental validation, UniBind effectively and scalably predicted the effects of SARS-CoV-2 spike protein variants on their binding affinities to the human ACE2 receptor, as well as to SARS-CoV-2 neutralizing monoclonal antibodies. Furthermore, in a cross-species analysis, UniBind could be applied to predict host susceptibility to SARS-CoV-2 variants and to predict future viral variant evolutionary trends. This in silico approach has the potential to serve as an early warning system for problematic emerging SARS-CoV-2 variants, as well as to facilitate research on protein-protein interactions in general.

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.

Broad Sarbecovirus Neutralizing Antibodies Obtained by Computational Design and Synthetic Library Screening

Members of the Sarbecovirus subgenus of Coronaviridae have twice caused deadly threats to humans. There is increasing concern about the rapid mutation of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), which has evolved into multiple generations of epidemic variants in 3 years. Broad neutralizing antibodies are of great importance for pandemic preparedness against SARS-CoV-2 variants and divergent zoonotic sarbecoviruses. Here, we analyzed the structural conservation of the receptor-binding domain (RBD) from representative sarbecoviruses and chose S2H97, a previously reported RBD antibody with ideal breadth and resistance to escape, as a template for computational design to enhance the neutralization activity and spectrum. A total of 35 designs were purified for evaluation. The neutralizing activity of a large proportion of these designs against multiple variants was increased from several to hundreds of times. Molecular dynamics simulation suggested that extra interface contacts and enhanced intermolecular interactions between the RBD and the designed antibodies are established. After light and heavy chain reconstitution, AI-1028, with five complementarity determining regions optimized, showed the best neutralizing activity across all tested sarbecoviruses, including SARS-CoV, multiple SARS-CoV-2 variants, and bat-derived viruses. AI-1028 recognized the same cryptic RBD epitope as the parental prototype antibody. In addition to computational design, chemically synthesized nanobody libraries are also a precious resource for rapid antibody development. By applying distinct RBDs as baits for reciprocal screening, we identified two novel nanobodies with broad activities. These findings provide potential pan-sarbecovirus neutralizing drugs and highlight new pathways to rapidly optimize therapeutic candidates when novel SARS-CoV-2 escape variants or new zoonotic coronaviruses emerge. IMPORTANCE The subgenus Sarbecovirus includes human SARS-CoV, SARS-CoV-2, and hundreds of genetically related bat viruses. The continuous evolution of SARS-CoV-2 has led to the striking evasion of neutralizing antibody (NAb) drugs and convalescent plasma. Antibodies with broad activity across sarbecoviruses would be helpful to combat current SARS-CoV-2 mutations and longer term animal virus spillovers. The study of pan-sarbecovirus NAbs described here is significant for the following reasons. First, we established a structure-based computational pipeline to design and optimize NAbs to obtain more potent and broader neutralizing activity across multiple sarbecoviruses. Second, we screened and identified nanobodies from a highly diversified synthetic library with a broad neutralizing spectrum using an elaborate screening strategy. These methodologies provide guidance for the rapid development of antibody therapeutics against emerging pathogens with highly variable characteristics.

Machine learning optimization of candidate antibody yields highly diverse sub-nanomolar affinity antibody libraries

Therapeutic antibodies are an important and rapidly growing drug modality. However, the design and discovery of early-stage antibody therapeutics remain a time and cost-intensive endeavor. Here we present an end-to-end Bayesian, language model-based method for designing large and diverse libraries of high-affinity single-chain variable fragments (scFvs) that are then empirically measured. In a head-to-head comparison with a directed evolution approach, we show that the best scFv generated from our method represents a 28.7-fold improvement in binding over the best scFv from the directed evolution. Additionally, 99% of designed scFvs in our most successful library are improvements over the initial candidate scFv. By comparing a library's predicted success to actual measurements, we demonstrate our method's ability to explore tradeoffs between library success and diversity. Results of our work highlight the significant impact machine learning models can have on scFv development. We expect our method to be broadly applicable and provide value to other protein engineering tasks.

An optimized thermodynamics integration protocol for identifying beneficial mutations in antibody design

Accurate identification of beneficial mutations is central to antibody design. Many knowledge-based (KB) computational approaches have been developed to predict beneficial mutations, but their accuracy leaves room for improvement. Thermodynamic integration (TI) is an alchemical free energy algorithm that offers an alternative technique for identifying beneficial mutations, but its performance has not been evaluated. In this study, we developed an efficient TI protocol with high accuracy for predicting binding free energy changes of antibody mutations. The improved TI method outperforms KB methods at identifying both beneficial and deleterious mutations. We observed that KB methods have higher accuracies in predicting deleterious mutations than beneficial mutations. A pipeline using KB methods to efficiently exclude deleterious mutations and TI to accurately identify beneficial mutations was developed for high-throughput mutation scanning. The pipeline was applied to optimize the binding affinity of a broadly sarbecovirus neutralizing antibody 10-40 against the circulating severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) omicron variant. Three identified beneficial mutations show strong synergy and improve both binding affinity and neutralization potency of antibody 10-40. Molecular dynamics simulation revealed that the three mutations improve the binding affinity of antibody 10-40 through the stabilization of an altered binding mode with increased polar and hydrophobic interactions. Above all, this study presents an accurate and efficient TI-based approach for optimizing antibodies and other biomolecules.

Recent advances in predicting and modeling protein-protein interactions

Protein-protein interactions (PPIs) drive biological processes, and disruption of PPIs can cause disease. With recent breakthroughs in structure prediction and a deluge of genomic sequence data, computational methods to predict PPIs and model spatial structures of protein complexes are now approaching the accuracy of experimental approaches for permanent interactions and show promise for elucidating transient interactions. As we describe here, the key to this success is rich evolutionary information deciphered from thousands of homologous sequences that coevolve in interacting partners. This covariation signal, revealed by sophisticated statistical and machine learning (ML) algorithms, predicts physiological interactions. Accurate artificial intelligence (AI)-based modeling of protein structures promises to provide accurate 3D models of PPIs at a proteome-wide scale.

Computational and artificial intelligence-based methods for antibody development

Due to their high target specificity and binding affinity, therapeutic antibodies are currently the largest class of biotherapeutics. The traditional largely empirical antibody development process is, while mature and robust, cumbersome and has significant limitations. Substantial recent advances in computational and artificial intelligence (AI) technologies are now starting to overcome many of these limitations and are increasingly integrated into development pipelines. Here, we provide an overview of AI methods relevant for antibody development, including databases, computational predictors of antibody properties and structure, and computational antibody design methods with an emphasis on machine learning (ML) models, and the design of complementarity-determining region (CDR) loops, antibody structural components critical for binding.

BERT2DAb: a pre-trained model for antibody representation based on amino acid sequences and 2D-structure

Prior research has generated a vast amount of antibody sequences, which has allowed the pre-training of language models on amino acid sequences to improve the efficiency of antibody screening and optimization. However, compared to those for proteins, there are fewer pre-trained language models available for antibody sequences. Additionally, existing pre-trained models solely rely on embedding representations using amino acids or k-mers, which do not explicitly take into account the role of secondary structure features. Here, we present a new pre-trained model called BERT2DAb. This model incorporates secondary structure information based on self-attention to learn representations of antibody sequences. Our model achieves state-of-the-art performance on three downstream tasks, including two antigen-antibody binding classification tasks (precision: 85.15%/94.86%; recall:87.41%/86.15%) and one antigen-antibody complex mutation binding free energy prediction task (Pearson correlation coefficient: 0.77). Moreover, we propose a novel method to analyze the relationship between attention weights and contact states of pairs of subsequences in tertiary structures. This enhances the interpretability of BERT2DAb. Overall, our model demonstrates strong potential for improving antibody screening and design through downstream applications.