Imagine some diseases are caused by 'sticky' proteins that current medicines can't grab onto. This project built a smart computer system that acts like a matchmaker, designing special molecules that glue these bad proteins to the cell's own trash disposal system. Once glued, the cell simply dissolves the harmful protein, potentially curing diseases that were previously thought untreatable.
Many disease-causing proteins are 'undruggable' because traditional medicines cannot bind to them. Current computational methods like docking are too slow, expensive, and inaccurate to efficiently find new ways to degrade these proteins.
The Celeris One cloud platform, featuring three modules: Hades (target ID), Xanthos (interaction prediction/ligand design), and Hephaistos (automated synthesis and validation).
If you are a large pharmaceutical company dealing with 'undruggable' targets for cancer or Alzheimer's — this project developed the Celeris One platform that uses deep learning to predict biomolecular interactions and design degraders more accurately than traditional docking.
If you are a biotech SME dealing with high R&D costs and slow lead optimization — this project developed a cloud-based orchestration platform that reduces the cost of developing medicines and increases the probability of success.
If you are a digital health company dealing with inaccurate molecular docking simulations — this project developed graph-based algorithms to predict ternary complexes, providing a faster and more precise alternative for medicinal chemists.
The platform uses deep learning and generative AI to predict interactions and design fragments in silico, which reduces the need for expensive and slow trial-and-error laboratory docking.
The company has established an 800sqm wet lab in Graz, Austria, with a team of 22 FTEs to translate computational designs into real biology and chemistry.
Based on available project data, the company develops in-house drug discovery programs and collaborates with big pharma companies, though specific licensing terms are not listed.
It provides a web application called Celeris One with three modules (Hades, Xanthos, Hephaistos) that replace or augment traditional docking methods used by medical and computational chemists.
The project period ran from 2022-07-01 to 2024-02-29, following initial research and development conducted in 2021 and 2022.
The project is led by a single Austrian SME, Celeris Therapeutics GmbH. With a 100% industry ratio and a lean structure, the company has successfully leveraged EU funding and venture capital to build a significant physical infrastructure (800sqm lab) and a specialized team of 22 people, focusing on rapid translation from AI design to wet-lab validation.
Contact Celeris Therapeutics GmbH in Graz, Austria
Contact us to find similar AI-driven drug discovery platforms for your pipeline.
