Hallucinating structure-conditioned antibody libraries for target-specific binders

Abstract

Antibodies are widely developed and used as therapeutics to treat cancer, infectious disease, and inflammation. During development, initial leads routinely undergo additional engineering to increase their target affinity. Experimental methods for affinity maturation are expensive, laborious, and time-consuming and rarely allow the efficient exploration of the relevant design space. Deep learning (DL) models are transforming the field of protein engineering and design. While several DL-based protein design methods have shown promise, the antibody design problem is distinct, and specialized models for antibody design are desirable. Inspired by hallucination frameworks that leverage accurate structure prediction DL models, we propose the FvHallucinator for designing antibody sequences, especially the CDR loops, conditioned on an antibody structure. Such a strategy generates targeted CDR libraries that retain the conformation of the binder and thereby the mode of binding to the epitope on the antigen. On a benchmark set of 60 antibodies, FvHallucinator generates sequences resembling natural CDRs and recapitulates perplexity of canonical CDR clusters. Furthermore, the FvHallucinator designs amino acid substitutions at the VH-VL interface that are enriched in human antibody repertoires and therapeutic antibodies. We propose a pipeline that screens FvHallucinator designs to obtain a library enriched in binders for an antigen of interest. We apply this pipeline to the CDR H3 of the Trastuzumab-HER2 complex to generate in silico designs predicted to improve upon the binding affinity and interfacial properties of the original antibody. Thus, the FvHallucinator pipeline enables generation of inexpensive, diverse, and targeted antibody libraries enriched in binders for antibody affinity maturation.

Strategies to improve the therapeutic efficacy of mesenchymal stem cell‐derived extracellular vesicle (MSC-EV): a promising cell-free therapy for liver disease

“…AI can be used to mine large amounts of bioinformatics data, including genomic, proteomic, and transcriptomic data. High-throughput screening techniques, molecular docking, and simulation techniques, as well as machine learning and model prediction, are used to screen, identify, predict, and characterize peptide sequences, protein expression patterns, and signaling pathways with organ-targeting properties, thus identifying potential targeting peptides ( Lin et al, 2022 ; Mahajan et al, 2022 ). However, it should be noted that although AI can help screen peptide sequences with potential targeting properties, the final synthesizability requires further experimental validation and optimization.…”

Toward enhancement of antibody thermostability and affinity by computational design in the absence of antigen

“…In the antibody design space, similar methods have utilized DeepAb for designing libraries of putative binders or optimizing the framework residues. 15 Beyond repurposing of structure predictors, several other modeling strategies have shown promise for antibody design. Protein language models are a class of self-supervised machine learning models that learn directly from large databases of protein sequences.…”

Enhancement of antibody thermostability and affinity by computational design in the absence of antigen

“…Subsequent methods have utilized structure predictors as differentiable protein designers [14], directly optimizing the input sequence according to an objective function – e.g., adopting a particular fold, hosting a scaffold, or binding another protein. In the antibody design space, similar methods have utilized DeepAb for designing libraries of putative binders or optimizing the framework residues [15]. Beyond repurposing of structure predictors, several other modeling strategies have shown promise for antibody design.…”