AI-accelerated therapeutic antibody development: practical insights DOI Creative Commons
Luca Santuari,

Marianne Bachmann Salvy,

Ioannis Xénarios

et al.

Frontiers in Drug Discovery, Journal Year: 2024, Volume and Issue: 4

Published: Sept. 3, 2024

Antibodies represent the largest class of biotherapeutics thanks to their high target specificity, binding affinity and versatility. Recent breakthroughs in Artificial Intelligence (AI) have enabled information-rich silico representations antibodies, accurate prediction antibody structure from sequence, generation novel antibodies tailored specific characteristics optimize for developability properties. Here we summarize state-of-the-art methods analysis. This valuable resource will serve as a reference application AI analysis sequencing datasets.

Language: Английский

A Comprehensive Review on Phage Therapy and Phage-Based Drug Development DOI Creative Commons
Longzhu Cui, Shinya Watanabe, Kazuhiko Miyanaga

et al.

Antibiotics, Journal Year: 2024, Volume and Issue: 13(9), P. 870 - 870

Published: Sept. 11, 2024

Phage therapy, the use of bacteriophages (phages) to treat bacterial infections, is regaining momentum as a promising weapon against rising threat multidrug-resistant (MDR) bacteria. This comprehensive review explores historical context, modern resurgence phage and phage-facilitated advancements in medical technological fields. It details mechanisms action applications phages treating MDR particularly those associated with biofilms intracellular pathogens. The further highlights innovative uses vaccine development, cancer gene delivery vectors. Despite its targeted efficient approach, therapy faces challenges related stability, immune response, regulatory approval. By examining these areas detail, this underscores immense potential remaining hurdles integrating phage-based therapies into practices.

Language: Английский

Citations

17
Challenges and compromises: Predicting unbound antibody structures with deep learning DOI Creative Commons
Alexander Greenshields‐Watson, Odysseas Vavourakis, Fabian C. Spoendlin

et al.

Current Opinion in Structural Biology, Journal Year: 2025, Volume and Issue: 90, P. 102983 - 102983

Published: Jan. 24, 2025

Language: Английский

Citations

1
The Patent and Literature Antibody Database (PLAbDab): an evolving reference set of functionally diverse, literature-annotated antibody sequences and structures DOI Creative Commons
Brennan Abanades, Tobias Hegelund Olsen, Matthew I. J. Raybould

et al.

Nucleic Acids Research, Journal Year: 2023, Volume and Issue: 52(D1), P. D545 - D551

Published: Nov. 16, 2023

Abstract Antibodies are key proteins of the adaptive immune system, and there exists a large body academic literature patents dedicated to their study concomitant conversion into therapeutics, diagnostics, or reagents. These documents often contain extensive functional characterisations sets antibodies they describe. However, leveraging these heterogeneous reports, for example offer insights properties query interest, is currently challenging as no central repository through which this wide corpus can be mined by sequence structure. Here, we present PLAbDab (the Patent Literature Antibody Database), self-updating containing over 150,000 paired antibody sequences 3D structural models, 65 000 unique. We describe methods used extract, filter, pair, model in PLAbDab, showcase how searched sequence, structure, keyword. uses include annotating with potential antigen information from similar entries, analysing models existing identify modifications that could improve properties, facilitating compilation bespoke datasets sequences/structures bind specific antigen. freely available via Github (https://github.com/oxpig/PLAbDab) searchable webserver (https://opig.stats.ox.ac.uk/webapps/plabdab/).

Language: Английский

Citations

22
Phage display technology and its impact in the discovery of novel protein-based drugs DOI Creative Commons

Catherine J. Hutchings,

Aaron K. Sato

Expert Opinion on Drug Discovery, Journal Year: 2024, Volume and Issue: 19(8), P. 887 - 915

Published: June 18, 2024

Introduction Phage display technology is a well-established versatile in vitro that has been used for over 35 years to identify peptides and antibodies use as reagents therapeutics, well exploring the diversity of alternative scaffolds another option conventional therapeutic antibody discovery. Such successes have responsible spawning range biotechnology companies, many complementary technologies devised expedite drug discovery process resolve bottlenecks workflow.

Language: Английский

Citations

6
Benchmarking antibody clustering methods using sequence, structural, and machine learning similarity measures for antibody discovery applications DOI Creative Commons

Dawid Chomicz,

Jarosław Kończak,

Sonia Wróbel

et al.

Frontiers in Molecular Biosciences, Journal Year: 2024, Volume and Issue: 11

Published: March 28, 2024

Antibodies are proteins produced by our immune system that have been harnessed as biotherapeutics. The discovery of antibody-based therapeutics relies on analyzing large volumes diverse sequences coming from phage display or animal immunizations. Identification suitable therapeutic candidates is achieved grouping the their similarity and subsequent selection a set antibodies for further tests. Such groupings typically created using sequence-similarity measures alone. Maximizing diversity in selected crucial to reducing number tests molecules with near-identical properties. With advances structural modeling machine learning, can now be grouped across other dimensions, such predicted paratopes three-dimensional structures. Here we benchmarked antibody methods clonotype, sequence, paratope prediction, structure embedding information. results were two tasks: binder detection epitope mapping. We demonstrate no method appears outperform others, while mapping, paratope, clusterings top performers. Most importantly, all propose orthogonal groupings, offering more pools when multiple than any single To facilitate exploring different methods, an online tool-CLAP-available at ( clap.naturalantibody.com ) allows users group, contrast, visualize methods.

Language: Английский

Citations

4
T-cell receptor structures and predictive models reveal comparable alpha and beta chain structural diversity despite differing genetic complexity DOI Creative Commons

Nele P. Quast,

Brennan Abanades, Bora Guloglu

et al.

bioRxiv (Cold Spring Harbor Laboratory), Journal Year: 2024, Volume and Issue: unknown

Published: May 21, 2024

Abstract T-cell receptor (TCR) structures are currently under-utilised in early-stage drug discovery and repertoire-scale informatics. Here, we leverage a large dataset of solved TCR from Immunocore to evaluate the current state-of-the-art for structure prediction, identify which regions remain challenging model. Through clustering analyses training TCR-specific model capable large-scale find that alpha chain VJ-recombined loop (CDRA3) is as structurally diverse correspondingly difficult predict beta VDJ-recombined (CDRB3). This differentiates variable domain loops genetically analogous antibody supports conjecture both chains deterministic antigen specificity. We hypothesise larger number joining genes compared compensates lack diversity gene segment. Overall, our study demonstrates valuable structure-function relationships can lie despite their simpler junctions. also provide over 1.5M predicted enable repertoire structural analysis elucidate strategies towards improving accuracy future predictors.

Language: Английский

Citations

4
Contrastive Learning Enables Epitope Overlap Predictions for Targeted Antibody Discovery DOI Creative Commons
Clinton Holt, Alexis K. Janke, Parastoo Amlashi

et al.

bioRxiv (Cold Spring Harbor Laboratory), Journal Year: 2025, Volume and Issue: unknown

Published: March 3, 2025

Computational epitope prediction remains an unmet need for therapeutic antibody development. We present three complementary approaches predicting relationships from amino acid sequences. First, we analyze ∼18 million pairs targeting ∼250 protein families and establish that a threshold of >70% CDRH3 sequence identity among antibodies sharing both heavy light chain V-genes reliably predicts overlapping-epitope pairs. Next, develop supervised contrastive fine-tuning framework large language models which results in embeddings better correlate with information than those pre-trained models. Applying this learning approach to SARS-CoV-2 receptor binding domain antibodies, achieve 82.7% balanced accuracy distinguishing same-epitope versus different-epitope demonstrate the ability predict relative levels structural overlap on functional bins (Spearman ρ = 0.25). Finally, create AbLang-PDB, generalized model broad range families. AbLang-PDB achieves five-fold improvement average precision compared sequence-based methods, effectively amount ( 0.81). In discovery campaign searching HIV-1 broadly neutralizing 8ANC195, 70% computationally selected candidates demonstrated specificity, 50% showing competitive 8ANC195. Together, computational presented here provide powerful tools epitope-targeted discovery, while demonstrating efficacy improving epitope-representation.

Language: Английский

Citations

0
T-cell receptor structures and predictive models reveal comparable alpha and beta chain structural diversity despite differing genetic complexity DOI Creative Commons

Nele P. Quast,

Brennan Abanades, Bora Guloglu

et al.

Communications Biology, Journal Year: 2025, Volume and Issue: 8(1)

Published: March 4, 2025

Language: Английский

Citations

0
Assessing AF2’s ability to predict structural ensembles of proteins DOI Creative Commons

Jakob R. Riccabona,

Fabian C. Spoendlin,

Anna-Lena M. Fischer

et al.

bioRxiv (Cold Spring Harbor Laboratory), Journal Year: 2024, Volume and Issue: unknown

Published: April 17, 2024

Abstract Recent breakthroughs in protein structure prediction have enhanced the precision and speed at which configurations can be determined, setting new benchmarks for accuracy efficiency field. However, fundamental mechanisms of biological processes a molecular level are often connected to conformational changes proteins. Molecular dynamics (MD) simulations serve as crucial tool capturing space proteins, providing valuable insights into their structural fluctuations. scope MD is limited by accessible timescales computational resources available, posing challenges comprehensively exploring behaviors. Recently emerging approaches focused on expanding capability AlphaFold2 (AF2) predict substates structures manipulating input multiple sequence alignment (MSA). These operate under assumption that MSA also contains information about heterogeneity structures. Here, we benchmark performance various workflows adapted AF2 ensemble focusing subsampling implemented ColabFold compare obtained with ensembles from NMR. As test cases, chose four proteins namely bovine pancreatic inhibitor (BPTI), thrombin two antigen binding fragments (antibody Fv nanobody), reliable experimentally validated (X-ray and/or NMR) was available. Thus, provide an overview levels currently achieved machine learning (ML) based generation. In three out find variations fall within predicted ensembles. Nevertheless, significant minima free energy surfaces remain undetected. This study highlights possibilities pitfalls when generating thus may guide development future tools while informing upon results available applications.

Language: Английский

Citations

1
Assessing AF2’s ability to predict structural ensembles of proteins DOI

Jakob R. Riccabona,

Fabian C. Spoendlin,

Anna-Lena M. Fischer

et al.

Structure, Journal Year: 2024, Volume and Issue: unknown

Published: Sept. 1, 2024

Language: Английский

Citations

1