OpenProtein.AI

OpenProtein.AI is a protein design cloud platform developed by pioneers in protein language modeling. The platform enables researchers and organizations across the life sciences to design, evaluate, and optimize proteins with high precision and efficiency using state-of-the-art AI models, including their next-generation foundation model PoET-2, as well as integrations with AlphaFold2, ESM2, and Clustal Omega.

The platform is built on the principle that proteins, like natural languages, follow learnable statistical patterns shaped by billions of years of evolution. OpenProtein.AI distills these patterns from natural protein sequence databases to capture evolutionary, structural, and functional properties, training large-scale machine learning models that can map the protein universe and generate novel, functional proteins. Users can further fine-tune these models with their own proprietary data to accelerate design toward specific objectives.

Core Platform Capabilities

• Learn: Predict variant effects and map mutagenesis hotspots to identify the most impactful positions in a protein sequence.
• Generate: Create diverse types of libraries including substitution libraries, combinatorial variant libraries, and bespoke sequence libraries tailored to specific design goals.
• Review: Visualize predicted protein structures to evaluate and assess generated variants before experimental testing.

Breadth of Application

• Supports projects ranging from small-scale 96-well plate experiments to high-throughput pipelines generating hundreds of thousands of data points.
• Compatible with any protein type, including antibodies, capsid proteins, and enzymes.
• Capable of optimizing for any protein property, including activity, expressibility, and thermostability.
• Learns directly from user data to predict and optimize properties specific to each project.

Scientific Foundation and Research

• Built by protein AI experts with peer-reviewed publications in leading journals and conferences, including Nature Communications (2023), NeurIPS (2023), Cell Systems (2021), and ICLR (2019).
• Demonstrated real-world impact: technology delivered antibodies with 20x greater potency than conventional mutagenesis in a head-to-head comparison.
• PoET, a generative model of protein families as sequences-of-sequences, represents a key foundational innovation underlying the platform.

Platform Philosophy and Principles

• Data ownership: User data remains the property of the user at all times.
• Open science: Commitment to transparent science as critical for human progress.
• Hands-on assistance: The team provides direct support to help users navigate the complexity of protein design.
• Accessible AI: Dedicated to the democratization of AI tools for protein engineering across the scientific community.

OpenProtein.AI is operated by NE47 Bio and is headquartered in Singapore. Early access to the platform is available for teams looking to achieve their protein engineering objectives in fewer, faster experimental iterations.