Sai T. Reddy
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Sai T. Reddy has a total of 492 co-authors. Here are the top co-authors by number of shared publications:
- Alexander Yermanos
University Medical Center Utrecht Center for Translational Immunology Lundlaan 6 , 3584 EA Utrecht
22 shared publications - Victor Greiff
University of Oslo
22 shared publications Daniel Neumeier
Department of Biosystems Science and Engineering, ETH Zürich, 4056 Basel, Switzerland
19 shared publicationsCédric R. Weber
Alloy Therapeutics (Switzerland) AG, Basel 4058, Switzerland
16 shared publicationsAlexander Yermanos
Swiss Fed Inst Technol, Dept Biosyst Sci & Engn, CH-4057 Basel, Switzerland
14 shared publicationsSimon Friedensohn
Department of Biosystems Science and Engineering, ETH Zürich, 4056 Basel, Switzerland
14 shared publications
Latest Publications
- Open Access
Generative Drug Design in a Loop with dtSFM
Directed evolution consisting of iterative rounds of diversification, selection, and counter-selection, underlies modern protein and antibody engineering, yet small-molecule drug design still advances largely through high-throughput screening and medicinal-chemistry intuition. Transformer softmax attention is mathematically identical to the Boltzmann distribution that governs molecular binding at thermal equilibrium 1 , an isomorphism that prescribes a sequence-native Specificity Foundation Model (SFM) 2 . This framework was recently applied across seven molecular recognition domains 3,4 and scaled into the drug–target SFM (dtSFM), the first to pair a full-scale encoder with a generative decoder 5 . Whether such a model can be driven, iteratively and under selection, to optimize leads rather than sample them once has not been shown. Here we present GenLoop, a closed generative drug design loop that turns single-pass generation into directed evolution of chemistry. dtSFM generates target-conditioned molecules and reranks them by their thermodynamic compatibility score. An orthogonal structural verifier, AlphaFold 3, is used that shares no architecture or training data with dtSFM. Cheminformatics filters enforce developability, and generative evolution is performed on the structurally verified candidates, selecting for predicted binders and counter-selecting against off-target chemistry. Applied across twelve drug targets spanning pharmacologically distinct mechanism classes, GenLoop produced AlphaFold 3-verified designs that reached the structural confidence of the approved drug for five of the twelve targets, with the best designs at interface iPTM 0.93–0.98 and PAE 0.8–2.0 Å, as well as resolving paralog selectivity across nine targets. Two full disease campaigns followed. For the cystic-fibrosis transmembrane conductance regulator, GenLoop designed nine developability-filtered and structurally novel lead candidates (iPTM up to 0.93, interface PAE 2.3 Å) targeting all three orthogonal sites of the approved drug Trikafta. For the GLP-1 receptor family, dtSFM engineered tunable single-, dual-, and triple-receptor incretin designs, yielding 23 central-pocket candidates that are structurally novel at median iPTM 0.89 and interface PAE 1.95 Å. GenLoop with dtSFM brings directed evolution to small molecules through computational-thermodynamic selection; wet-lab validation is the immediate next step.
Open AccessGenerative Drug Design in a Loop with dtSFM
Directed evolution consisting of iterative rounds of diversification, selection, and counter-selection, underlies modern protein and antibody engineering, yet small-molecule drug design still advances largely through high-throughput screening and medicinal-chemistry intuition. Transformer softmax attention is mathematically identical to the Boltzmann distribution that governs molecular binding at thermal equilibrium 1 , an isomorphism that prescribes a sequence-native Specificity Foundation Model (SFM) 2 . This framework was recently applied across seven molecular recognition domains 3,4 and scaled into the drug–target SFM (dtSFM), the first to pair a full-scale encoder with a generative decoder 5 . Whether such a model can be driven, iteratively and under selection, to optimize leads rather than sample them once has not been shown. Here we present GenLoop, a closed generative drug design loop that turns single-pass generation into directed evolution of chemistry. dtSFM generates target-conditioned molecules and reranks them by their thermodynamic compatibility score. An orthogonal structural verifier, AlphaFold 3, is used that shares no architecture or training data with dtSFM. Cheminformatics filters enforce developability, and generative evolution is performed on the structurally verified candidates, selecting for predicted binders and counter-selecting against off-target chemistry. Applied across twelve drug targets spanning pharmacologically distinct mechanism classes, GenLoop produced AlphaFold 3-verified designs that reached the structural confidence of the approved drug for five of the twelve targets, with the best designs at interface iPTM 0.93–0.98 and PAE 0.8–2.0 Å, as well as resolving paralog selectivity across nine targets. Two full disease campaigns followed. For the cystic-fibrosis transmembrane conductance regulator, GenLoop designed nine developability-filtered and structurally novel lead candidates (iPTM up to 0.93, interface PAE 2.3 Å) targeting all three orthogonal sites of the approved drug Trikafta. For the GLP-1 receptor family, dtSFM engineered tunable single-, dual-, and triple-receptor incretin designs, yielding 23 central-pocket candidates that are structurally novel at median iPTM 0.89 and interface PAE 1.95 Å. GenLoop with dtSFM brings directed evolution to small molecules through computational-thermodynamic selection; wet-lab validation is the immediate next step.
- Open Access
A Drug–Target Specificity Foundation Model for Off-target Prediction, Repurposing, and Generative Design
Molecular recognition - which small molecule binds which protein, and with what selectivity - governs the efficacy, safety, and discovery of every therapeutic, yet binding specificity is still determined by experimental screening or by computational methods that first predict three-dimensional structure. Transformer softmax attention is mathematically isomorphic to the Boltzmann distribution governing molecular binding at thermal equilibrium 1 , an identity that prescribes a single sequence-native architecture: the Specificity Foundation Model (SFM), which computes molecular binding compatibility as a thermodynamic quantity directly from sequence 2 . The framework was recently realized as prototype encoders across six molecular-recognition domains 3 . Here we report the small molecule drug-target protein SFM (dtSFM) as the first instance to pair a full-scale encoder with a generative decoder, trained on publicly available data consisting of 714,747 measured drug–protein interactions spanning 522,776 compounds and 22,964 proteins. Throughout, we verify binding predictions with AlphaFold 3 4 as an orthogonal structural verifier that shares no architecture, training data, or representational basis with dtSFM. From this single dtSFM model we demonstrate the three sequence-native applications of drug discovery: off-target prediction, repurposing, and generative design. The dtSFM encoder retrieves a drug’s target, and a target’s drug, at 95% and 89% recall-at-10 in distribution, respectively. In the drug→target direction it screens off-targets at proteome scale, ranking the documented off-targets of clinical kinase inhibitors at a median of 30th out of 4,910 genes - the top 0.6% of the screen - when validated against a chemoproteomic panel 5 . In the target→drug direction it ranks the full 522,776-compound library against three immunology targets, identifying 46 novel candidates that pass AlphaFold-3 structural gating. The dtSFM cross-attentive decoder generates novel molecules for 16 targets, 850 of 1,200 (71%) designed candidates match the AlphaFold 3 structural confidence of the approved drug (iPTM ≥ 0.9 and interface PAE ≤ 1.67 Å), with the best candidates reaching iPTM 0.95–0.99 and interface PAE 0.79–1.37 Å. dtSFM brings computational thermodynamics to every stage where molecular recognition shapes drug discovery; experimental wet-lab validation is the immediate next step.
Open AccessA Drug–Target Specificity Foundation Model for Off-target Prediction, Repurposing, and Generative Design
Molecular recognition - which small molecule binds which protein, and with what selectivity - governs the efficacy, safety, and discovery of every therapeutic, yet binding specificity is still determined by experimental screening or by computational methods that first predict three-dimensional structure. Transformer softmax attention is mathematically isomorphic to the Boltzmann distribution governing molecular binding at thermal equilibrium 1 , an identity that prescribes a single sequence-native architecture: the Specificity Foundation Model (SFM), which computes molecular binding compatibility as a thermodynamic quantity directly from sequence 2 . The framework was recently realized as prototype encoders across six molecular-recognition domains 3 . Here we report the small molecule drug-target protein SFM (dtSFM) as the first instance to pair a full-scale encoder with a generative decoder, trained on publicly available data consisting of 714,747 measured drug–protein interactions spanning 522,776 compounds and 22,964 proteins. Throughout, we verify binding predictions with AlphaFold 3 4 as an orthogonal structural verifier that shares no architecture, training data, or representational basis with dtSFM. From this single dtSFM model we demonstrate the three sequence-native applications of drug discovery: off-target prediction, repurposing, and generative design. The dtSFM encoder retrieves a drug’s target, and a target’s drug, at 95% and 89% recall-at-10 in distribution, respectively. In the drug→target direction it screens off-targets at proteome scale, ranking the documented off-targets of clinical kinase inhibitors at a median of 30th out of 4,910 genes - the top 0.6% of the screen - when validated against a chemoproteomic panel 5 . In the target→drug direction it ranks the full 522,776-compound library against three immunology targets, identifying 46 novel candidates that pass AlphaFold-3 structural gating. The dtSFM cross-attentive decoder generates novel molecules for 16 targets, 850 of 1,200 (71%) designed candidates match the AlphaFold 3 structural confidence of the approved drug (iPTM ≥ 0.9 and interface PAE ≤ 1.67 Å), with the best candidates reaching iPTM 0.95–0.99 and interface PAE 0.79–1.37 Å. dtSFM brings computational thermodynamics to every stage where molecular recognition shapes drug discovery; experimental wet-lab validation is the immediate next step.
- Open Access
Vibe Coding Specificity Foundation Models
Molecular recognition - the determination of which agent binds which target - governs adaptive immunity, gene regulation, signal transduction, RNA silencing, enzyme catalysis, and the selectivity of therapeutics. Determining binding specificity remains dependent on experimental screening or domain-specific computational tools that do not generalize across binding modalities. Transformer softmax attention is mathematically identical to the Boltzmann distribution governing molecular binding. This identity, together with five conditions of molecular recognition systems, prescribes a single neural network architecture for cross-modal binding prediction: dual sequence encoders, symmetric contrastive learning, and a learned physical temperature. A Specificity Foundation Model (SFM) is an instance of this physics-derived, sequence-to-sequence architecture that maps any agent-target sequence pair to a binding compatibility score, enabling bidirectional retrieval across molecular recognition domains without requiring structural information. The first SFM for antibody-antigen binding demonstrated ~100,000-fold greater data efficiency than comparable vision-language models. Here we report six SFMs across six molecular recognition domains - transcription factor-DNA, enzyme-substrate, peptide-MHC, CRISPR gRNA-off-target genomic DNA, microRNA-mRNA target, and small molecule drug-target protein - using the identical architecture without modification and trained using publicly available data only. Evaluated by cross-modal retrieval from pools of 512 candidates (random baseline 0.2%), in-distribution R@1 ranges from 27.7% to 98.0% across the six domains. mir-SFM retrieves miRNA targets at 98.0% R@1, including the ~80% of validated interactions that seed-matching tools cannot find. mhcSFM achieves 95.4% R@1 on held-out rare HLA alleles absent from training. Applying crisprSFM to CRISPR off-target prediction improves precision to 94.0% compared to 33.2% from Hamming distance alone. All six SFMs were built by a domain expert with no programming experience using vibe coding - natural-language-directed AI coding agents - with numerical claims independently verified by an orthogonal AI auditor. These results establish SFMs as a physics-derived, sequence-native class of model that augments experimental and computational workflows across molecular recognition domains.
Open AccessVibe Coding Specificity Foundation Models
Molecular recognition - the determination of which agent binds which target - governs adaptive immunity, gene regulation, signal transduction, RNA silencing, enzyme catalysis, and the selectivity of therapeutics. Determining binding specificity remains dependent on experimental screening or domain-specific computational tools that do not generalize across binding modalities. Transformer softmax attention is mathematically identical to the Boltzmann distribution governing molecular binding. This identity, together with five conditions of molecular recognition systems, prescribes a single neural network architecture for cross-modal binding prediction: dual sequence encoders, symmetric contrastive learning, and a learned physical temperature. A Specificity Foundation Model (SFM) is an instance of this physics-derived, sequence-to-sequence architecture that maps any agent-target sequence pair to a binding compatibility score, enabling bidirectional retrieval across molecular recognition domains without requiring structural information. The first SFM for antibody-antigen binding demonstrated ~100,000-fold greater data efficiency than comparable vision-language models. Here we report six SFMs across six molecular recognition domains - transcription factor-DNA, enzyme-substrate, peptide-MHC, CRISPR gRNA-off-target genomic DNA, microRNA-mRNA target, and small molecule drug-target protein - using the identical architecture without modification and trained using publicly available data only. Evaluated by cross-modal retrieval from pools of 512 candidates (random baseline 0.2%), in-distribution R@1 ranges from 27.7% to 98.0% across the six domains. mir-SFM retrieves miRNA targets at 98.0% R@1, including the ~80% of validated interactions that seed-matching tools cannot find. mhcSFM achieves 95.4% R@1 on held-out rare HLA alleles absent from training. Applying crisprSFM to CRISPR off-target prediction improves precision to 94.0% compared to 33.2% from Hamming distance alone. All six SFMs were built by a domain expert with no programming experience using vibe coding - natural-language-directed AI coding agents - with numerical claims independently verified by an orthogonal AI auditor. These results establish SFMs as a physics-derived, sequence-native class of model that augments experimental and computational workflows across molecular recognition domains.
- Open Access
- 23 April 2026
- Published by Elsevier in Cell Reports Methods
- Vol. 6 (6) , 101417
Dissecting serum polyclonal antibody escape to SARS-CoV-2 variants by deep mutational learning
- 23 April 2026
- Published by Elsevier in Cell Reports Methods
- Vol. 6 (6) , 101417
Open AccessDissecting serum polyclonal antibody escape to SARS-CoV-2 variants by deep mutational learning
- Open Access
Methods for Molecular Recognition Computing
The softmax attention mechanism in transformer architectures (Vaswani et al., 2017) is mathematically identical to the Boltzmann distribution governing molecular binding at thermal equilibrium (Boltzmann, 1877). Luce's Choice Axiom (1959) establishes this function - which we term the convergence equation - as the unique function satisfying five axioms of competitive selection: positivity, normalization, unrestricted domain, rank preservation, and independence of irrelevant alternatives. We show that five additional architecture conditions - discrete intermolecular contacts, bilinear energy decomposition, finite competitor pools, thermal equilibrium, and stochastic selection - are satisfied by at least ten biological molecular recognition systems and together prescribe a complete neural architecture: dual encoders, cross-attention, InfoNCE contrastive training, symmetric loss, learned temperature, and cross-attentive decoder. We term this architecture a Specificity Foundation Model (SFM) and specify it for antibody-antigen, TCR-peptide-MHC, transcription factor-DNA, microRNA-mRNA, enzyme-substrate, CRISPR guide RNA-DNA, drug-target, peptide-MHC, receptor-ligand, and RNA-binding protein-RNA recognition. The first implementation (CALM; Lee et al., 2026) achieves antibody-antigen retrieval from approximately 4,000 training pairs with ~100,000-fold greater data efficiency than comparable contrastive architectures trained without the physics derivation. We classify this as Level 3 architecture-physics alignment and derive three further theoretical results: an exponential scaling law for retrieval accuracy as a function of training data diversity (the MRC scaling law), a two-parameter affinity calibration framework connecting contrastive scores to binding free energies, and a hybrid recursive learning framework for cross-modal reinforcement learning with orthogonal verification. The failure conditions of the framework are analyzed in terms of the validity of equilibrium thermodynamics for molecular binding and the convergence properties of gradient-based parameter estimation.
Open AccessMethods for Molecular Recognition Computing
The softmax attention mechanism in transformer architectures (Vaswani et al., 2017) is mathematically identical to the Boltzmann distribution governing molecular binding at thermal equilibrium (Boltzmann, 1877). Luce's Choice Axiom (1959) establishes this function - which we term the convergence equation - as the unique function satisfying five axioms of competitive selection: positivity, normalization, unrestricted domain, rank preservation, and independence of irrelevant alternatives. We show that five additional architecture conditions - discrete intermolecular contacts, bilinear energy decomposition, finite competitor pools, thermal equilibrium, and stochastic selection - are satisfied by at least ten biological molecular recognition systems and together prescribe a complete neural architecture: dual encoders, cross-attention, InfoNCE contrastive training, symmetric loss, learned temperature, and cross-attentive decoder. We term this architecture a Specificity Foundation Model (SFM) and specify it for antibody-antigen, TCR-peptide-MHC, transcription factor-DNA, microRNA-mRNA, enzyme-substrate, CRISPR guide RNA-DNA, drug-target, peptide-MHC, receptor-ligand, and RNA-binding protein-RNA recognition. The first implementation (CALM; Lee et al., 2026) achieves antibody-antigen retrieval from approximately 4,000 training pairs with ~100,000-fold greater data efficiency than comparable contrastive architectures trained without the physics derivation. We classify this as Level 3 architecture-physics alignment and derive three further theoretical results: an exponential scaling law for retrieval accuracy as a function of training data diversity (the MRC scaling law), a two-parameter affinity calibration framework connecting contrastive scores to binding free energies, and a hybrid recursive learning framework for cross-modal reinforcement learning with orthogonal verification. The failure conditions of the framework are analyzed in terms of the validity of equilibrium thermodynamics for molecular binding and the convergence properties of gradient-based parameter estimation.
- Open Access
Contrastive learning for antibody-antigen sequence-to-specificity prediction
Predicting which antibodies bind to which antigens directly from primary amino acid sequences remains a major challenge, as no current method can reliably determine this specificity at both a repertoire and proteome scale. Structure-based protein design frameworks can propose antibody binders to specified antigenic epitopes, but they do not solve the “sequence-to-specificity” task of mapping antibodies to cognate epitopes, and vice versa. Here, we introduce CALM (Cross-attention Adaptive Immune Receptor–Antigen Language Model), a dual-encoder plus cross-attentive decoder architecture that treats antibody–antigen recognition as molecular translation. Using contrastive learning, antigen and antibody encoders learn a shared embedding space that aligns cognate epitope–paratope binding pairs. CALM-1.0 is trained and evaluated on 4,138 curated antibody-antigen pairs obtained from the PDB-derived structural antibody database (SAbDab). On a leakage-controlled test split drawn from sequence clusters at 80% identity and unseen during training, CALM-1.0 achieves a mean top-1 retrieval (R@1) of 7%, with consistent performance across both directions (Ab→Ag and Ag→Ab). CALM establishes a foundation for bidirectional antibody-antigen sequence-to-specificity prediction with the potential to unify retrieval and generative design.
Open AccessContrastive learning for antibody-antigen sequence-to-specificity prediction
Predicting which antibodies bind to which antigens directly from primary amino acid sequences remains a major challenge, as no current method can reliably determine this specificity at both a repertoire and proteome scale. Structure-based protein design frameworks can propose antibody binders to specified antigenic epitopes, but they do not solve the “sequence-to-specificity” task of mapping antibodies to cognate epitopes, and vice versa. Here, we introduce CALM (Cross-attention Adaptive Immune Receptor–Antigen Language Model), a dual-encoder plus cross-attentive decoder architecture that treats antibody–antigen recognition as molecular translation. Using contrastive learning, antigen and antibody encoders learn a shared embedding space that aligns cognate epitope–paratope binding pairs. CALM-1.0 is trained and evaluated on 4,138 curated antibody-antigen pairs obtained from the PDB-derived structural antibody database (SAbDab). On a leakage-controlled test split drawn from sequence clusters at 80% identity and unseen during training, CALM-1.0 achieves a mean top-1 retrieval (R@1) of 7%, with consistent performance across both directions (Ab→Ag and Ag→Ab). CALM establishes a foundation for bidirectional antibody-antigen sequence-to-specificity prediction with the potential to unify retrieval and generative design.
- Open Access
- 19 February 2026
- Published by Springer Nature in Nature Chemical Biology
- Vol. 22 (10) , 1654-1665
Drug-controlled CAR T cells through the regulation of cell–cell interactions
Chimeric antigen receptor (CAR) T cell therapy is constrained by on-target, off-tumor toxicities and cellular exhaustion because of chronic antigen exposure. CARs incorporating small-molecule controlled on- and off-switches can enhance both safety and therapeutic efficacy but their design is limited by the scarcity of nonimmunogenic protein elements responsive to nonimmunosuppressive, clinically approved drugs with favorable pharmacodynamics. Here we combine rational design and library-based optimization of a protein–protein interaction (PPI) of human origin to develop venetoclax-controlled drug-regulated off-switch PPI (DROP)-CARs. DROP-CARs enable dose-dependent release of the tumor-targeting scFv and consequent reduction in T cell binding to the tumor cell. Additionally, we present proof of concept for a dual DROP-CAR controlled by different small molecules, as well as for logic-gated synthetic receptors enabling STAT3 signaling. We demonstrate in vitro and in vivo function of DROP-CAR T cells and conclude that the approach holds promise for clinical application.
- 19 February 2026
- Published by Springer Nature in Nature Chemical Biology
- Vol. 22 (10) , 1654-1665
Open AccessDrug-controlled CAR T cells through the regulation of cell–cell interactions
Chimeric antigen receptor (CAR) T cell therapy is constrained by on-target, off-tumor toxicities and cellular exhaustion because of chronic antigen exposure. CARs incorporating small-molecule controlled on- and off-switches can enhance both safety and therapeutic efficacy but their design is limited by the scarcity of nonimmunogenic protein elements responsive to nonimmunosuppressive, clinically approved drugs with favorable pharmacodynamics. Here we combine rational design and library-based optimization of a protein–protein interaction (PPI) of human origin to develop venetoclax-controlled drug-regulated off-switch PPI (DROP)-CARs. DROP-CARs enable dose-dependent release of the tumor-targeting scFv and consequent reduction in T cell binding to the tumor cell. Additionally, we present proof of concept for a dual DROP-CAR controlled by different small molecules, as well as for logic-gated synthetic receptors enabling STAT3 signaling. We demonstrate in vitro and in vivo function of DROP-CAR T cells and conclude that the approach holds promise for clinical application.
- Open Access
Computational Convergence of Adaptive Immunity and Artificial Intelligence
The adaptive immune system and modern artificial intelligence (AI) have independently converged on identical computational strategies for solving the recognition and generalization problem: learning to identify which inputs should be associated with which outputs, such that performance extends to novel instances never seen during training. We demonstrate that four major AI architectural innovations can be derived from immunological first principles. Two represent exact mathematical equivalences: the softmax function in transformer attention is the Boltzmann distribution governing antibody-antigen binding; the InfoNCE loss in contrastive learning is the negative log of clonal selection probability. Two represent strategic convergences: the pre-training/fine-tuning/RLHF training hierarchy parallels the germline/somatic hypermutation/T follicular helper stages of antibody affinity maturation; retrieval-augmented generation mirrors the long-lived plasma cell and memory B cell dual memory system. Critically, these derivations run from immunology to AI, not the reverse—the mathematical structures of transformer attention and contrastive learning could have been derived from immunological first principles before their empirical discovery in AI. The framework generates testable predictions for immune specificity foundation models and suggests that certain computational strategies may be fundamental to any system solving recognition and generalization under resource constraints. Modern AI has independently converged on computational strategies that the adaptive immune system discovered through 500 million years of evolution. One-sentence summary: The mathematics of transformer attention and contrastive learning can be derived from the biophysics of immune recognition.
Open AccessComputational Convergence of Adaptive Immunity and Artificial Intelligence
The adaptive immune system and modern artificial intelligence (AI) have independently converged on identical computational strategies for solving the recognition and generalization problem: learning to identify which inputs should be associated with which outputs, such that performance extends to novel instances never seen during training. We demonstrate that four major AI architectural innovations can be derived from immunological first principles. Two represent exact mathematical equivalences: the softmax function in transformer attention is the Boltzmann distribution governing antibody-antigen binding; the InfoNCE loss in contrastive learning is the negative log of clonal selection probability. Two represent strategic convergences: the pre-training/fine-tuning/RLHF training hierarchy parallels the germline/somatic hypermutation/T follicular helper stages of antibody affinity maturation; retrieval-augmented generation mirrors the long-lived plasma cell and memory B cell dual memory system. Critically, these derivations run from immunology to AI, not the reverse—the mathematical structures of transformer attention and contrastive learning could have been derived from immunological first principles before their empirical discovery in AI. The framework generates testable predictions for immune specificity foundation models and suggests that certain computational strategies may be fundamental to any system solving recognition and generalization under resource constraints. Modern AI has independently converged on computational strategies that the adaptive immune system discovered through 500 million years of evolution. One-sentence summary: The mathematics of transformer attention and contrastive learning can be derived from the biophysics of immune recognition.
- Open Access
A yeast mating-based platform enables the generation and screening of ultra large antibody libraries
Yeast display is a powerful platform for antibody engineering, offering eukaryotic protein folding and compatibility with fluorescence-activated sorting (FACS). However, conventional yeast display faces limitations in terms of library size and display heterogeneity due to plasmid-based expression. Here, we present a yeast display platform that combines genomic library integration and high-efficiency mating to generate ultra-diverse antibody libraries approaching the diversity of phage-libraries (>10 11 ). We engineered two yeast strains, LFYa and LFYalpha, featuring genomically integrated landing pads for high-efficiency library insertion. To enable compatibility with deep sequencing, we integrated the recombinase BxB1, which links heavy (HC) and light chain (LC) information onto a single chromosome. Using our platform, we constructed synthetic antibody libraries with HC and LC diversities of 4.1×10 7 and 1.7×10 7 variants, respectively and with an improved mating protocol, we achieved a combinatorial library diversity exceeding 10 11 . We demonstrated that genomic integration yields uniform surface display. Screening this antibody library against multiple antigens resulted in the discovery of binders with affinities in the single-digit nanomolar to picomolar range, demonstrating the platform’s utility for the discovery of antibodies with therapeutically relevant affinities. This work establishes a robust and deep sequencing-compatible yeast display system that overcomes key limitations of previous mating-based platforms.
Open AccessA yeast mating-based platform enables the generation and screening of ultra large antibody libraries
Yeast display is a powerful platform for antibody engineering, offering eukaryotic protein folding and compatibility with fluorescence-activated sorting (FACS). However, conventional yeast display faces limitations in terms of library size and display heterogeneity due to plasmid-based expression. Here, we present a yeast display platform that combines genomic library integration and high-efficiency mating to generate ultra-diverse antibody libraries approaching the diversity of phage-libraries (>10 11 ). We engineered two yeast strains, LFYa and LFYalpha, featuring genomically integrated landing pads for high-efficiency library insertion. To enable compatibility with deep sequencing, we integrated the recombinase BxB1, which links heavy (HC) and light chain (LC) information onto a single chromosome. Using our platform, we constructed synthetic antibody libraries with HC and LC diversities of 4.1×10 7 and 1.7×10 7 variants, respectively and with an improved mating protocol, we achieved a combinatorial library diversity exceeding 10 11 . We demonstrated that genomic integration yields uniform surface display. Screening this antibody library against multiple antigens resulted in the discovery of binders with affinities in the single-digit nanomolar to picomolar range, demonstrating the platform’s utility for the discovery of antibodies with therapeutically relevant affinities. This work establishes a robust and deep sequencing-compatible yeast display system that overcomes key limitations of previous mating-based platforms.
- Open Access
- 17 September 2025
- Published by Elsevier in Cell Systems
- Vol. 16 (9) , 101391
Synthetic coevolution reveals adaptive mutational trajectories of neutralizing antibodies and SARS-CoV-2
- 17 September 2025
- Published by Elsevier in Cell Systems
- Vol. 16 (9) , 101391
Open AccessSynthetic coevolution reveals adaptive mutational trajectories of neutralizing antibodies and SARS-CoV-2