Proteios-Mini
Link to open source: https://github.com/prabhsuratsingh/Proteios-Mini
Link to Live Project: https://proteios.streamlit.app/
Proteios is an innovative platform designed to simplify and enhance protein research and drug discovery. It integrates cutting-edge Google AI and ML technologies to provide efficient protein analysis and comparison.
🔬 Protein Analysis – Users can enter queries like "Find the proteins that cause Alzheimer's", and Proteios leverages Gemini's flash 2.0 inferencing API along with extensive protein databases(alphafold, uniprot, chembl) to fetch relevant features, applications, and biological effects. Additionally, it provides 3D visualization of proteins using Python-based rendering libraries.
💊 Drug Discovery – It offers a streamlined, AI-driven drug discovery pipeline that begins with identifying proteins linked to a specific disease or condition. Once a target protein is selected, the platform analyzes its structural and biochemical properties to calculate a druggability score—an indicator of how likely the protein is to be a successful drug target. Proteios then performs virtual screening, evaluating thousands of known compounds for potential binding, and also supports De Novo drug design, where novel molecular structures are generated using advanced AI models. This end-to-end approach accelerates the discovery of viable drug candidates tailored to specific proteins and diseases.
🧬Protein Anomaly Detection - Proteios features a powerful anomaly detection module that distinguishes structurally or functionally abnormal proteins from their normal counterparts. This is achieved through a custom-trained Variational Autoencoder (VAE) model, which learns a compressed representation (latent space) of normal protein structures.
By combining AI-powered insights, molecular visualization, and machine learning, Proteios makes protein research more accessible, efficient, and insightful.
Deployed app (give it a few mins🙂)- https://proteios.streamlit.app/






