Biology subjects

Cong, L.

Publications and source records attributed to Cong, L..

At least 19 recordsLinked to original sources

bioRxiv2026

ctDNA Release Heterogeneity in Luminal Breast Cancer Revealed by Single-Cell Multi-Omics Spatial Analysis

Backgroud Breast cancer is a leading cause of cancer-related mortality, and HR+/HER2-; is the predominant subtype. Molecular pathology, including PIK3CA and ESR1 mutation, informs prognosis and treatment selection. Circulating tumor DNA (ctDNA) offers a non-invasive approach for cancer monitoring, but its clinical utility is limited by tissue plasma discordance, reflecting heterogeneous ctDNA release. The cellular and spatial determinants of this release remain poorly understood, particularly in HR+/HER2-; subtype, where heterogeneity across subclones complicates liquid biopsy interpretation. Results To identify the cellular and spatial determinants of ctDNA release, we developed scMASTER, integrating single-cell transcriptomes and mutations, spatial transcriptomics and matched tumor and plasma DNA sequencing. In ten patients with non-metastatic HR+/HER2- breast cancer, eight recurrent malignant epithelial states showed distinct ctDNA release potential. Antigen-presentation, stress, mitosis and mesenchymal-like states showed high release. PIK3CA and ESR1 mutations were linked to distinct functional states and altered ctDNA detectability. Three spatial architectures, EMT-enriched regions, intraductal luminal progenitors and avascular hypoxic zones, showed high release. Tumor-microenvironment interactions were bidirectional: anti-tumoral immunity correlated with reduced release, whereas immunomodulatory components correlated with increased release. A stratification model based on these features identified high-risk, immunosuppressed patients with poor prognosis, validated across external cohorts. Conclusions Our study establishes a single-cell spatial framework for decoding ctDNA release heterogeneity, linking tumor states, somatic mutations, spatial architecture, and immune interactions to ctDNA release. These findings provide mechanistic insights into tissue-plasma discordance, offer a biological rationale for interpreting ctDNA dynamics, refine patient risk stratification and liquid biopsy-based monitoring in HR+/HER2-; breast cancer.

cancer biology↗

bioRxiv2026

AutoScreen: AI Co-Scientist System for Target Discovery in Functional Genomics

Target discovery in functional genomics remains largely manual and time-consuming, lacking systematic tools for efficient and reproducible gene-level hypothesis generation. We introduce AutoScreen, an AI co-scientist system supporting target discovery through both Pre-screen Design, which constructs perturbation libraries de novo from free-text research descriptions, and Post-screen Analysis, which re-ranks experimental screen hits by integrating statistical scores with biological context. AutoScreen leverages the complementary strengths of multiple specialist agents that perform deep research into multi-modal evidence, information restructuring, parallel searches across >26 biomedical databases, evidence synthesis, and target review to provide transparent, explainable gene prioritization with pipeline provenance. Across 320 genome-scale CRISPR screens as expert-curated benchmarks, AutoScreen achieved a ~19% increase in validated-hit recovery among its top 100 predictions, relative to the strongest agent baseline, and required a 1.2-fold smaller library to recover the same number of hits at the top-500 reference point. AutoScreen reached mean average precision more than two orders-of-magnitude above random baseline. Further, we validated AutoScreen in cancer immune-evasion case studies focusing on natural killer (NK) and T-cell therapeutics. AutoScreen recovered NK-resistance genes that were initially lower-ranked in a leukemia screen, moving MUC1, PDPN, and LRRC15 from raw ranks of 118, 81, 1384 to 5, 44, 659, respectively. In follow-up tumor killing assay with primary human NK cells, individual perturbation validated all three hits successfully. Next, in prospective benchmarking across two cytotoxic T-cell-killing screens, AutoScreen recovered 77.1% of ground-truth hits called by the consensus of two gold-standard analysis pipelines (FDR<0.1), achieving 8% improvement over the strongest general-purpose LLM agent baseline. Finally, we re-analyzed >500 public datasets to construct the AutoScreen Resource Hub, a growing knowledge base of pre-computed, annotated reports for CRISPR screens, RNA-seq differential expression, and gene-level UK Biobank genome-wide association studies. This agentic AI approach enables real-time genomics benchmark construction that are continuously updated, expanded, for evaluating frontier AI co-scientists. Overall, AutoScreen enables AI-powered target discovery to be more auditable, scalable, extensible, and reproducible.

bioengineering↗

bioRxiv2026

BiomniBench: Process-level Evaluation of LLM Agents for Real-world Biomedical Research

LLM agents now perform real biomedical research, but evaluating them rigorously is hard. Outcome-only benchmarks fail in two ways. First, a correct final answer can come from memorization, reward hacking, or wrong reasoning that produces the right number by chance. Second, valid alternative analyses are marked wrong simply because they differ from the reference. We introduce BiomniBench, a process-level evaluation framework that scores the full agent trajectory against expert-designed, task-specific rubrics. Our first release, BiomniBench-DA, contains 100 data-analysis tasks across 17 task types, 5 disease areas, and a general-biology category, each based on a paper from journals such as Nature, Cell, and Science and co-developed with an original author or a domain expert. Benchmarking frontier and open-weight models across four agent harnesses reveals three findings. Frontier and open-weight bases cluster within a few points of each other, with substantial headroom for all models. The agent harness shifts scores by more than the gap between successive model generations. Agents reliably ground claims in real sources yet consistently fall short on method selection, biological interpretation, and scientific reasoning. BiomniBench is the first process-level benchmark for LLM agents in biomedical research, providing the dimension-level diagnostics that outcome scoring cannot. Datasethuggingface.co/datasets/phylobio/BiomniBench-DA

bioinformatics↗

bioRxiv2025

De Novo Computational Design of VHH Nanobodies Against LGR5

VHH discovery traditionally relies on animal immunization or large-scale library screening, methods that are slow, costly, and often ineffective for challenging targets such as GPCRs. We present a fully de novo computational pipeline for epitope-directed VHH design, integrating generative backbone modeling, deep learning-based sequence optimization, and iterative experimental feedback. Using LGR5 as a model, we progressed from in silico design to functional binders without structural templates. Across three design-test-learn cycles, millions of candidates were reduced to epitope-specific binders with nanomolar affinity and high thermal stability (melting temperature [Tm] > 65 {degrees}C). Cryogenic electron microscopy (cryo-EM) confirmed atomic-level agreement (RMSD {approx} 2.2 [A]). This structure-validated approach accelerates timelines, reduces cost, and is broadly applicable to GPCRs and other membrane proteins, enabling "on-demand" therapeutic antibody generation.

bioengineering↗

bioRxiv2025

LabOS: The AI Co-Scientist That Sees and Works With Humans

Modern science advances fastest when thought meets action. LabOS represents the first AI co-scientist that unites computational reasoning with physical experimentation through multimodal perception, self-evolving agents, and XR-enabled, embodied human-AI collaboration. To empower agentic AI with embodied intelligence, we built LabSuperVision, the first vision-language-model (VLM) benchmark spanning biomedical and materials science laboratories, designed to evaluate scientific perception and reasoning. This endowed LabOS with visual reasoning capabilities that surpass those of general-purpose models on laboratory-specific tasks. By connecting multi-model AI agents, smart glasses, and human-AI-robot collaboration, LabOS allows AI to see what scientists see, understand experimental context, and participate in real-time execution through immersive XR interaction and robotic autonomation. Across applications-from cancer immunotherapy target discovery to cell engineering-LabOS shows that AI can move beyond computational design to participation, turning the laboratory into an intelligent, collaborative environment where human and machine discovery evolve together.

bioengineering↗

bioRxiv2025

AI-guided design of common light chains to enable manufacturable bispecific antibodies

Bispecific antibodies (BsAbs) offer therapeutic advantages but face manufacturing bottlenecks from light chain mispairing, which can generate a substantial fraction of incorrect products and increase manufacturing complexity and cost of goods. Common light chains (cLC) eliminate mispairing, yet existing approaches require screening thousands of variants per target. We present an AI-driven framework that computationally designs cLCs through structure-guided pairing of non-cognate VH-VL interfaces, reducing experimental screening by three orders of magnitude. The platform successfully engineers therapeutic antibodies lacking experimental structures, expanding applicability beyond crystallographic databases. Among 10 therapeutic targets, we successfully generated designs for 7 targets, comprising 55 unique BsAb pairs. Of these, 43.6% (24/55) were successfully as BsAb cLCs. Three bispecific antibodies reached production-ready specifications: >90% purity and 1.6-1.8 g/L titers. This platform democratizes bispecific antibody development, expanding access beyond well-resourced programs.

bioengineering↗

bioRxiv2025

Surfaceome CRISPR activation screening uncovers ligands regulating tumor sensitivity to NK cell killing

Natural killer (NK) cell-based immunotherapies represent a promising avenue for cancer treatment due to their ability to eliminate cancer cells independently of antigen presentation and potential for "off-the-shelf" use. However, the molecular determinants governing tumor cell susceptibility to NK cell-mediated cytotoxicity remain incompletely understood. Here we employed CRISPR activation (CRISPRa) screening to systematically identify cancer cell surface regulators of NK cell killing across multiple cancer types. Using a comprehensive surfaceome-focused library, we screened human and murine cancer cell lines co-cultured with NK cells, identifying both known and novel ligands that modulate NK cell cytotoxicity. Our screens revealed established factors including CD43 (encoded by SPN), while uncovering previously uncharacterized regulators such as CD44, PDPN, and Siglec-1/CD169. Validation through complementary cDNA overexpression and genetic knockout approaches confirmed that disruption of CD43, CD44, PDPN, and Siglec-1 significantly altered cancer cell susceptibility to NK killing both in vitro and in humanized mouse models. Analysis of clinical datasets show that expression of identified factors correlates with patient survival outcomes in an NK-context dependent manner supporting their therapeutic relevance. Most notably, our mechanistic studies demonstrate that CD43-mediated NK cell resistance operates independently of its previously proposed interaction with Siglec-7 on NK cells. Furthermore, we find that targeting CD43 on either NK cells or engineered T cells substantially enhances their cytotoxic activity against leukemia cell lines. These results establish gain-of-function screening as a powerful approach for discovering immunoregulatory surface proteins and identify multiple promising targets for enhancing NK cell-based cancer immunotherapies.

immunology↗

bioRxiv2025

BioLab: End-to-End Autonomous Life Sciences Research with Multi-Agents System Integrating Biological Foundation Models

Scientific discovery in the life sciences remains hindered by fragmented workflows, narrow-scope computational models, and inefficient links between in silico prediction and wet-lab validation. We present BioLab, a multi-agent system that integrates domain-specialized foundation models to automate end-to-end biological research. BioLab comprises eight collaborating agents, including a Planner, Reasoner, and Critic, orchestrated through a Memory Agent that enables iterative refinement via retrieval-augmented generation and a suite of 219 computational xBio-Tools spanning five biological scales (DNA, RNA, protein, cell, and chemical). These tools are built on the xTrimo Universe, a collection of 104 models derived from six foundation models (xTrimoChem, Protein, RNA, DNA, Cell, and Text), the majority of which achieve state-of-the-art (91-100% SOTA ratios) on domain benchmarks. Across standard reasoning tasks (PubMedQA, MMLU-Pro/Biology, GPQA-diamond), BioLab consistently outperformed leading large language models, including GPT-4o, Gemini-2.5, and DeepSeek-R1. Beyond benchmarks, BioLab autonomously executed a fully computational pipeline for de novo macrophage-targeting antibody design, progressing from target mining to multi-objective antibody optimization, where molecular dynamics simulations revealed structural mechanisms underlying enhanced affinity of optimized variants. Closing the computational-experimental loop, BioLab designed optimized antibodies (Pem-MOO-1, Pem-MOO-2) that achieved IC50 values of 0.01-0.016 nM, markedly surpassing the parental Pembrolizumab (0.027 nM) for PD-1. Functional assays confirmed enhanced pathway blockade and improved multi-parameter performance profiles. Together, these results establish BioLab as a generalizable framework for AI-native scientific discovery, demonstrating how multi-agent systems coupled with foundation models can autonomously generate, execute, and experimentally validate novel biological hypotheses.

bioinformatics↗

bioRxiv2025

Computational Design of Soluble CCR8 Analogues with Preserved Antibody Binding

G protein-coupled receptors (GPCRs) represent the largest class of drug targets, yet their membrane-embedded nature poses significant challenges for structural studies and therapeutic development. Here, we report the successful computational design and experimental validation of soluble CCR8 analogues that maintain native antibody binding properties. Using an integrated pipeline combining ProteinMPNN-sol sequence design, and structure-based filtering, we generated 13 CCR8 analogues from 272 initial designs across three N-terminal truncation strategies. Experimental validation confirmed 62% success rate (8/13 designs) with protein yields of 1.19-73.72 mg/L in aqueous buffer, representing a significant improvement over traditional membrane protein production method. Surface plasmon resonance analysis demonstrated that all analogues retained mAb1 binding with dissociation constants ranging from 77-857 nM, comparable to wild-type CCR8 (KD = 190 nM). Despite extensive sequence divergence (10-13% identity with wild-type CCR8), structural integrity was preserved as evidenced by binding affinity maintenance and computational structural validation. This work demonstrates the feasibility of computationally designing functional soluble analogues of challenging membrane proteins, with implications for accelerating drug discovery, antibody development, and structural biology studies. Our approach addresses critical limitations in membrane protein accessibility while preserving native epitope presentation, opening new avenues for therapeutic target characterization and binder discovery.

bioengineering↗

bioRxiv2025

STELLA: Self-Evolving LLM Agent for Biomedical Research

The staggering complexity of modern biomedical research has intensified the aspiration for a generalist "Biomedical World Model", yet current AI agents remain constrained by static capabilities and a lack of self-evolution mechanisms. To bridge this gap, we present STELLA, a self-evolving multimodal agent designed to progressively refine its computational reasoning and physical execution through interaction. STELLA operates via a collaborative multi-agent framework (comprising Manager, Developer, Critic, Critic, and Tool Creation agents) that continuously updates reasoning templates and autonomously expands a dynamic "Tool Ocean". We demonstrate STELLAs capabilities on the created Tool Creation Benchmark, where it attains a score of 4.01/5 with 100% task completion, significantly outperforming state-of-the-art models including GPT-5, Claude 4 Opus, and Biomni. Beyond computational metrics, STELLA drives experimentally validated scientific discovery. In oncology, the agent identified Butyrophilin Subfamily 3 Member A1 (BTN3A1) as a novel negative regulator of natural killer (NK) cell function in acute myeloid leukemia (AML), verified via CRISPR knockout studies. In protein engineering, STELLA orchestrated a complete directed evolution workflow for the enzyme strictosidine synthase, identifying variants, notably M276L, exhibiting more than a two-fold improvement in catalytic activity. Finally, the system extends to physical laboratory automation by training Vision-Language-Action (VLA) models through a Decompose-Monitor-Recover mechanism, which increased success rates from 17% to 82%. By integrating autonomous tool evolution, biological discovery, and robotic control, STELLA offers a blueprint for a self-evolving world model in the life sciences.

bioinformatics↗

bioRxiv2025

Genome-Bench: A Scientific Reasoning Benchmark from Real-World Expert Discussions

In this short report, we present an automated pipeline tailored for the genomics domain and introduce Genome-Bench, a new benchmark constructed from over a decade of scientific forum discussions on genome engineering. Our pipeline transforms raw interactions into a reinforcement learningfriendly multiple-choice questions format, supported by 3000+ high-quality questionanswer pairs spanning foundational biology, experimental troubleshooting, tool usage, and beyond. To our knowledge, this is the first end-to-end pipeline for teaching LLMs to reason from scientific discussions, with promising potential for generalization across scientific domains beyond biology. The dataset is available at https://huggingface.co/datasets/Mingyin0312/Genome-Bench.

genomics↗

bioRxiv2025

Biomni: A General-Purpose Biomedical AI Agent

Biomedical research underpins progress in our understanding of human health and disease, drug discovery, and clinical care. However, with the growth of complex lab experiments, large datasets, many analytical tools, and expansive literature, biomedical research is increasingly constrained by repetitive and fragmented workflows that slow discovery and limit innovation, underscoring the need for a fundamentally new way to scale scientific expertise. Here, we introduce Biomni, a general-purpose biomedical AI agent designed to autonomously execute a wide spectrum of research tasks across diverse biomedical subfields. To systematically map the biomedical action space, Biomni first employs an action discovery agent to create the first unified agentic environment - mining essential tools, databases, and protocols from tens of thousands of publications across 25 biomedical domains. Built on this foundation, Biomni features a generalist agentic architecture that integrates large language model (LLM) reasoning with retrieval-augmented planning and code-based execution, enabling it to dynamically compose and carry out complex biomedical workflows - entirely without relying on predefined templates or rigid task flows. Systematic benchmarking demonstrates that Biomni achieves strong generalization across heterogeneous biomedical tasks - including causal gene prioritization, drug repurposing, rare disease diagnosis, micro-biome analysis, and molecular cloning - without any task-specific prompt tuning. Real-world case studies further showcase Biomnis ability to interpret complex, multi-modal biomedical datasets and autonomously generate experimentally testable protocols. Biomni envisions a future where virtual AI biologists operate alongside and augment human scientists to dramatically enhance research productivity, clinical insight, and healthcare. Biomni is ready to use at https://biomni.stanford.edu, and we invite scientists to explore its capabilities, stress-test its limits, and co-create the next era of biomedical discoveries.

bioinformatics↗

bioRxiv2025

Gene Editing for ATXN3 Inactivation in Machado-Joseph disease: CRISPR-Cas9 as a Therapeutic Alternative to TALEN-Induced Toxicity

Machado-Joseph disease (MJD) is an autosomal dominantly-inherited neurodegenerative disorder, caused by an over-repetition of the polyglutamine-codifying region in the ATXN3 gene. Strategies based on the suppression of the deleterious gene products have demonstrated promising results in pre-clinical studies. Nonetheless, these strategies do not target the root cause of the disease. In order to prevent the downstream toxic pathways, our goal was to develop gene editing-based strategies to permanently inactivate the human ATXN3 gene. TALENs and CRISPR-Cas9 systems were designed to target exon 2 of this gene and functional characterization was performed in a human cell line. After the demonstration of TALENs and CRISPR-Cas9 efficiency on gene disruption, a sequence of each system was selected for further in vivo experiments. Although both TALENs and CRISPR-Cas9 systems led to a drastic reduction of ATXN3 aggregates in the striatum of a lentiviral-based mouse model of MJD/SCA3, only CRISPR-Cas9 system allowed the improvement of key neuropathological markers of the disease. Importantly, the administration of the engineered system in YAC-MJD84.2/84.2 mice mediated a delay in disease progression, when compared with non-treated littermates. These data provide the first in vivo evidence of the efficacy of a CRISPR-Cas9-based approach to permanently inactivate the ATXN3 gene in the brain of two mouse models of the disease, supporting its potential as a new therapeutic avenue in the context of MJD/SCA3.

neuroscience↗

bioRxiv2024

RNAGenesis: Foundation Model for Enhanced RNA Sequence Generation and Structural Insights

RNA molecules are central to gene regulation, catalysis, and molecular recognition, and offer broad opportunities for therapeutic applications. However, uncovering their complex sequence, structure, and function relationships, particularly for non-coding RNAs, remains a formidable challenge. Here, we introduce RNAGenesis, a Generalist RNA foundation model that integrates sequence representation, structural prediction, and de novo functional design within a single generative framework. Trained on diverse clustered non-coding RNAs, RNAGenesis leverages a BERT-style encoder, query-based latent compression, and a diffusion-guided decoder enhanced by inference-time alignment with gradient guidance and beam search strategies. Through comprehensive evaluations, RNAGenesis achieves state-of-the-art performance on 11 of 13 tasks in the BEACON benchmark and surpasses structure-aware models in inverse folding, 3D structure prediction, and de novo structure design. We further introduce RNATx-Bench, a dedicated benchmark for RNA therapeutics comprising over 100,000 experimentally validated sequences. RNAGenesis demonstrates strong predictive performance across ASOs, siRNAs, shRNAs, circRNAs, and untranslated region (UTR) variants. Furthermore, RNAGenesis enables functional RNA design, including aptamers targeting IGFBP3 and structurally constrained sgRNA scaffolds. Wet-lab validation confirms aptamer binding with KD values as low as 4.02 nM and up to 2.5-fold improvement in editing efficiency across CRISPR-Cas9, base editing, and prime editing systems. These results position RNAGenesis as a next-generation general-purpose RNA foundation model with broad utility for computational modeling and experimental therapeutic design.

bioinformatics↗

bioRxiv2024

FoldMark: Protecting Protein Generative Models with Watermarking

Proteins are the principal architects of life, fueling advances in bioengineering, drug discovery, and synthetic biology. The integration of generative AI with computational protein science has revolutionized protein design while also posing dual-use risks, such as enabling the creation of pandemic-capable proteins that require strong biosecurity safeguards. Here, we introduce FoldMark, a first-of-its-kind watermarking strategy leveraging distributional and evolutionary principles tailored for protein generative models, balancing watermark capacity and structural quality. FoldMark achieves over 95% watermark bit accuracy at 32 bits with minimal impact on structural quality (>0.9 scTM scores) for leading models including AlphaFold3, ESMFold, RFDiffusion, and RFDiffusionAA. For user tracing, FoldMark can successfully trace up to 1 million users. To validate FoldMark in wet lab, we applied it to structure-based design of EGFP and CRISPR-Cas13, showing wildtype-level function (98% fluorescence, 95% editing efficiency) and >90% watermark detection, demonstrating its practical utility for safeguarding AI-driven protein research.

bioinformatics↗

bioRxiv2024

CRISPR-GPT: An LLM Agent for Automated Design of Gene-Editing Experiments

Performing effective gene-editing experiments requires a deep understanding of both the CRISPR technology and the biological system involved. Meanwhile, despite their versatility and promise, Large Language Models (LLMs) often lack domain-specific knowledge and struggle to accurately solve biological design problems. We present CRISPR-GPT, an LLM agent system to automate and enhance CRISPR-based gene-editing design and data analysis. CRISPR-GPT leverages the reasoning capabilities of LLMs for complex task decomposition, decision-making, and interactive human-artificial intelligence (AI) collaboration. This system incorporates domain expertise, retrieval techniques, external tools, and a specialized LLM fine-tuned with open-forum discussions among scientists. CRISPR-GPT assists users in selecting CRISPR systems, experiment planning, designing gRNAs, choosing delivery methods, drafting protocols, designing assays, and analyzing data. We showcase the potential of CRISPR-GPT by knocking-out four genes with CRISPR-Cas12a in a human lung adenocarcinoma cell line and epigenetically activating two genes using CRISPR-dCas9 in human melanoma cell line. CRISPR-GPT enables fully AI-guided gene-editing experiment design and analysis across different modalities, validating its effectiveness as an AI co-pilot in genome engineering.

bioinformatics↗

bioRxiv2024

Modeling the genomic complexity of human cancer using Cas12a mice

Somatic genome editing in mouse models has increased our understanding of the in vivo effects of genetic alterations in areas ranging from neuroscience to cancer biology and beyond. However, existing models are limited in their ability to create multiple targeted edits. Thus, our understanding of the complex genetic interactions that underlie development, homeostasis, and disease remains incomplete. Cas12a is an RNA-guided endonuclease with unique attributes that enable simple targeting of multiple genes with crRNA arrays containing tandem guides. To accelerate and expand the generation of complex genotypes in somatic cells, we generated transgenic mice with Cre-regulated and constitutive expression of enhanced Acidaminococcus sp. Cas12a (enAsCas12a). In these mice, enAsCas12a-mediated somatic genome editing robustly generated compound genotypes, as exemplified by the initiation of diverse cancer types driven by homozygous inactivation of trios of tumor suppressor genes. We further integrated these modular crRNA arrays with clonal barcoding to quantify the size and number of tumors with each array, as well as the efficiency of each crRNA. These Cas12a alleles will enable the rapid generation of disease models and broadly facilitate the high-throughput investigation of coincident genomic alterations in somatic cells in vivo.

cancer biology↗

bioRxiv2023

Protein language model-guided engineering of an anti-CRISPR protein for precise genome editing in human cells

Promiscuous editing by CRISPR/Cas systems within the human genome is a major challenge that must be addressed prior to applying these systems therapeutically. In bacteria, CRISPR/Cas systems have evolved in a co-evolutionary arms race with infectious phage viruses that contain inhibitory anti-CRISPR proteins within their genomes. Here, we harness the outcome of this co-evolutionary arms race to engineer an AcrIIA4 anti-CRISPR protein to increase the precision of CRISPR/Cas-based genome targeting. We developed an approach that specifically leveraged (1) protein language models, (2) deep mutational scanning, and (3) highly parallel DNA repair measurements within human cells. In a single experiment, [~]10,000 AcrIIA4 variants were tested to identify lead AcrIIA4 variants that eliminated detectable off-target editing events while retaining on-target activity. The candidates were further tested in a focused round of screening that included a high-fidelity version of Cas9 as a benchmark. Finally, arrayed experiments using Cas9 delivered as ribonucleoprotein were conducted that demonstrated an increase in gene editing precision across two independent genomic loci and a reduction in the frequency of translocation events between an on-target and off-target site. Thus, language-model-guided high-throughput screening is an effective way to efficiently engineer AcrIIA4 to increase gene editing precision, which could be used to improve the fidelity of gene editing-based therapeutics and to reduce genotoxicity.

synthetic biology↗