TECHNOLOGY LICENSING OPPORTUNITY: SignalGen

A platform that predicts the right signal peptide for a specific protein Many medicines, vaccines, and biotechnology products rely on proteins reaching the right cellular compartment or secretion outside a cell. Whether a protein gets there in enough quantities depends on its signal peptide or leader sequence. Scientists can identify many existing signal peptides, but predicting which one will work best for a specific protein remains difficult.

SignalGen uses generative AI to predict the signal peptide most likely to guide a protein to the right location. To make that prediction, it considers the protein, where it needs to go, and the organism it comes from. This approach gives researchers a new way to improve protein expression for medicines, vaccines, and biotechnology products.

Overview: Proteins carry out many of the body's essential functions, but they can only do their job if they reach the right place inside or outside a cell. Many proteins rely on signal peptides to get there. Typically, a signal peptide is a short sequence segment at the beginning of a protein that acts like a built-in address label, helping direct the protein to the location where it is needed.

Predicting the right signal peptide for a specific protein remains challenging. Existing software can identify known signal peptides or determine whether a protein already contains one, but it cannot predict which signal peptide is best suited to produce the desired protein expression and cellular location. Researchers provide the protein they want to study, where they want it to go to a specific compartment in a cell, and the organism.

SignalGen predicts the signal peptide best suited for those conditions, giving researchers a stronger starting point before laboratory testing. SignalGen is even critical for AI designed de novo proteins as it would require a signal peptide.

Advantages: Predicts signal peptides for specific proteins instead of only identifying known ones Considers the protein, where it needs to go, and the organism in a single prediction Works with both human and non-human proteins Natural-language interface reduces the need for programming or scripting Demonstrated approximately 90% prediction accuracy using the all-organism model Technology Description: At the core of SignalGen is a protein language model trained on thousands of proteins and their associated signal peptides. By learning the relationships between proteins, signal peptides, where proteins are located inside the cell, and the organisms they come from, the model predicts the signal peptide best suited for a specific protein. The model combines information about the protein, where it needs to go inside or outside the cell, and the organism it comes from to generate its prediction.

It was initially trained using human proteins and later expanded to include proteins from many different organisms, improving its versatility and achieving approximately 90% prediction accuracy. SignalGen also includes an AI agent that makes the platform easier to use. Researchers interact with the system using natural language rather than programming or scripting.

The AI agent gathers the required information, runs the prediction, and returns the recommended signal peptide through a guided workflow.

Market Applications: Therapeutic protein development Vaccine research and development Drug discovery Biomanufacturing Protein engineering Synthetic biology Industrial biotechnology Molecular biology research TRL: 3 U.S.

Patent pending LA-UR-26-27860 LANL Tech Partnerships: Unlock the Innovative Potential Los Alamos National Laboratory offers a wide range of cutting-edge technologies and capabilities that may provide your company with a competitive edge in the market and unlock the innovative potential that can enhance, refine, and revolutionize your products. LANL's licensing program focuses on moving inventions developed by our researchers to commercial innovations. Patented and patent pending inventions and copyrighted software are available to existing and start-up companies through exclusive and non-exclusive licensing agreements.

For specific discussions, please contact [email on the source notice].

Note: This is not a call for external services for the development of this technology. https://www.lanl.gov/engage/collaboration/feynman-center/partner-with-us/licensing-technology m.lanl.gov/tech-search