Matwings Launches MatwingsVenus™: A Conversational AI Agent for Personalized Protein R&D
Matwings Technology
AI Protein Design Service Provider
Editor | Wang Duoyu
Typesetting | Shuicheng Wen
April 24, 2026Matwings TechnologyPublishConversational Protein Research and Development Agent——MatwingsVenus™(Xiaowu™). Users on the platformThrough conversational agents, complete industry research, tag database retrieval, protein design, automated experimental validation, and online expert collaboration.And a series of other tasks, achieving“Design is verification, verification is iteration”The intelligent R&D model.
In this model, individual users can independently invoke Matwings Technology's services on the MatwingsVenus™ platform.AI Design Capability, Automated Wet Lab Platform, completing a closed loop from model inference to experimental validation and then to iterative optimization through natural language dialogue. In addition, users can alsoConsult experts at any time`, providing authoritative insights for design solutions.`
ExperienceAddress:https://matvenus.com/
Agent-Centric
Integrate computation and wet lab experimentsAnd expertsThe Complete Scientific Research Process of Interaction
Over the past decade, the capabilities of AI have gradually expanded from perceptual recognition and content generation to reasoning, collaboration, task decomposition, and tool invocation, among others.The significance of MatwingsVenus™ lies in extending the intelligence of the digital world further into the real world.。
MatwingsVenus™ Agent-CentricBuilt a one-stop protein R&D platform, SupportBillion-Level Real Tag Protein Data Retrieval, Integration200+Protein design tools,50+Experts certified by the platform, and30+ Optimized by experts in various fields Skills，Users only need to input task objectives through natural language, and the system can automatically break down tasks, schedule corresponding design, prediction, analysis, and screening capabilities to complete in-depth research, enzyme mining, directed evolution, de novo design, and automated wet-lab collaboration, etc.。
Unlike traditional tool platforms, MatwingsVenus™ doesn't just pile functions together, butAutomatically organize work paths around task objectives and directly link the results to an automated shared laboratory for execution, forming an executable, feedback-enabled, and iterative protein design and development closed loop.。
The key to this closed loop lies in the deep collaboration between cloud-based design and physical experiments.. Agent（Agent）After completing the design, the platform can import the results into the plasmid ordering and experimental arrangement process through its self-constructed communication mechanism, automatically link subsequent experimental tasks, and drive robots to complete sample preparation, protein purification, and functional testing. Eventually, the experimental results are fed back into the next round of AI design, forming a complete cycle."Conversational Dry-Wet Closed Loop" Iteration Driven by Computational Wet Experiments and Wet-Experiment-Feedback to Computation。
From Closed-loop to Implementation
Verify Platform Capabilities in Real Projects
Currently, the MatwingsVenus™ platform has already beenInnovative DrugAndSynthetic BiologyCompleted multiple cases in the field, demonstrating the practical application capabilities of the AI design and automated verification integrated platform.
AI-Based De Novo Design and Validation of Immune Modulatory Receptor Targets
Immune regulatory receptors are important targets in the research and development of innovative biologics, widely used in areas such as cancer, autoimmune diseases, and inflammatory conditions, demonstrating high clinical value and market potential.
InDe Novo Design Project of a Certain Immunomodulatory Receptor TargetIn China, Matwings Technology, based on its independently developed MatwingsVenus™ platform, has successfully obtained dozens of novel binders with in vitro cell blocking activity.（binder）Molecules, completed the full-process validation of de novo design binders, demonstrating the strength of AI-driven protein innovation in drug development.
This achievement was not easy to attain.Due to the novelty of target selection, the lack of reference drug molecules of the same kind, and the predominantly polar regions on the target surface with a lack of typical high druggability binding hotspots, the difficulty of designing proteins with stable and tight binding has increased. Meanwhile, the natural ligand already possesses nanomolar affinity.（nM）Higher affinity also further increases the difficulty for candidate molecules to achieve effective blocking.
Relying on the MatwingsVenus™ platform, Matwings Technology takes target structure and functional requirements as inputs, and the Agent automatically completes all computational tasks including scaffold screening, interface design, sequence optimization, and drug-likeness prediction, rapidly outputting high-quality binder design sequences. This significantly enhances design efficiency and success rates. Subsequently, the binder samples prepared via the automated experimental platform demonstrated outstanding performance in in vitro cell activity assays.，Dozens of molecules possess clear cell-blocking activity, combining functional inhibition with high affinity potential.Force.
De Novo Design of Antibodies Based on Target-Specific Epitopes
Modification of Sweet Protein Monellin Increases Sweetness by More Than Tenfold
MonellinIt is a classic protein-based sweet substance. The natural form typically consists of two chains maintained by non-covalent interactions, but its stability is relatively poor, easily affected by pH, temperature, ionic strength, and long-term storage conditions. For this type of protein, the research and development challenges lie not only in enhancing sweetness but also in maintaining structural stability.
MatwingsVenus™ adopts the strategy of "Agent design—automated experiments—AI feedback—Agent redesign," progressively narrowing the search space and optimizing the candidate set round by round. Ultimately, through continuous iterative optimization, a set of 24 representative candidates was formed, yielding multiple superior candidate variants. The results of automated experiments showed,The sweetness of multiple samples increased more than tenfold compared to the wild type, while maintaining high heat resistance at approximately 75°C.
From de novo design of real targets to multi-round optimization of complex proteins, MatwingsVenus™ has completed the entire process of AI design, experimental validation, and result iteration in multiple real projects, which means that,The platform has moved beyond the conceptual level.Is to truly close the loop on conversational dry and wet systemsTransform into actionable R&D capabilities.
Conclusion
When the predictive power of AI is deeply integrated with automated execution capabilities, MatwingsVenus™ represents not just the iteration of a tool, but the transformation of R&D capabilities that were once only accessible to large enterprises or institutions into infrastructure that individuals can also utilize.Promote Protein Research and Development from "Large Platform-Driven" to "Individually Accessible"。In the future, MatwingsVenus™ will continue to promote the democratization of protein R&D capabilities, enabling more individuals and small teams to access shared scientific research resources through conversational AI agents, quickly transforming ideas into real products and effects, allowing everyone to develop the proteins they desire.
Matwings Technology（Matwings）Is aAI-Driven Full-Stack Protein R&D Platform Company, by Shanghai Jiao Tong UniversityLiang HongFounded by a professor, through the construction of an ultra-large-scale protein sequence dataset/tag library that is "unique and unparalleled," and using this as a cornerstone, a general large model for protein design was developed, achieving precise functional protein design directly tailored to industrial needs.
With its self-developed large protein model as the core engine, Matwings Technology has developed scientific research agents and automated experimental devices, creating an AI protein R&D full-stack platform.: From market/literature/patent research, to molecular design, small-scale verification, pilot process optimization, and eventually production scale-up. By fully integrating AI into industrial application scenarios, we can surpass the current cognitive boundaries of human experts, and develop various bioproducts based on proteins as the underlying chips faster and better, such as innovative drugs, health products, agriculture, functional foods, and bio-based materials.
PointIn View, Convey Your Taste