Machine learning-driven discovery of insect-repellent proteins

Overview

Shiru Inc. leverages advanced machine learning (ML) and protein language models (PLMs) through its proprietary Protein Discovery Engine (PDE) to discover natural proteins and peptides with insect-repellent properties. This innovative approach targets specific insects in domestic and outdoor environments, offering a more efficient and resource-conscious alternative to traditional empirical methods. The technology focuses on identifying peptides that bind to the Aminopeptidase-N (APN) receptor, crucial for ensuring insect-repellent activity.

Technical specifications

Key features:

• Utilizes a curated database of over 450 million natural proteins.
• Incorporates ML models to predict peptide activity and stability under various conditions such as pH and temperature.
• Employs a peptide classifier for APN receptor binding prediction.
• Predicts soluble expression in heterologous hosts, facilitating scalable production.
• Focuses on peptides with high stability and expression for formulation and delivery testing.

Technology readiness level

The technology is at TRL 3, with ongoing collaborations to define product-specific constraints and validation plans. Future development includes peptide expression in microbial systems and screening for activity and stability, bridging the gap towards commercial scalability.