K-MELLODDY

Project 1

Federated Learning Platform Development

Build a Federated Learning-based Drug Discovery (FDD) platform and operating the FAM solution.

Project 2

Federated Learning Data Provision

Providing data from pharmaceutical companies, universities, hospitals, and research institutions and utilizing the FAM model.

Project 3

Federated Learning Model Development

Discovery of FAM solutions and application models (selecting 5 projects each year for three years).

The Future of AI-Driven Drug Discovery with Federated Learning

The K-MELLODDY (Korea Machine Learning Ledger Orchestration for Drug Discovery) project aims to develop an advanced AI model for drug discovery. Inspired by the EU MELLODDY Project (2019–2022), which involved leading pharmaceutical companies like Merck, Pfizer, Novartis, and AstraZeneca, K-MELLODDY leverages Federated Learning (FL) to allow pharmaceutical companies, hospitals, universities, and government research institutes to train AI models without the need to share sensitive data.

Platform Development

4 institutions

Data Provide

21 institutions

AI Model Development

8 institutions

About the Project

The project aims to produce a Federated ADMET Model (FAM), which predicts key ADMET (Absorption, Distribution, Metabolism, Excretion, and Toxicity) and PK parameters, which are essential for drug discovery and clinical trials. By integrating longitudinal data from in-vitro, in-vivo, preclinical, and clinical sources, FAM will significantly improve prediction accuracy, accelerating the drug discovery process and reducing development risks.

After demonstrating the effectiveness of FAM, K-MELLODDY plans to address further drug development challenges such as drug-target interactions, drug-drug interactions, and pharmacogenetics. Supported by the Korean Ministry of Science and ICT and the Ministry of Health and Welfare, this five-year project (April 2024 – December 2028) has a total budget of $25 million USD and aims to foster a robust bioindustry ecosystem in Korea through secure, data-driven collaboration.

Message from the Director

AI has become an essential tool in drug development. Global big tech companies like Google and NVIDIA are equipping themselves with AI, partnering with major global pharmaceutical companies to reshape the landscape of drug development and the digital bioindustry.

While Korea is also focusing on bio and AI as next-generation core industries, the current approach may not be enough to catch up with the already leading advanced countries. We urgently need a strategy to lead in the next-generation bio-AI convergence industry.

The K-MELLODDY project aims to create an ecosystem where expensive and difficult-to-share bio data can be safely utilized collaboratively and applied effectively to drug development.

We hope that the successful execution of this project will serve as a turning point for Korea to become a leading country in the AI-driven bioindustry.

Thank you.

Kim Hwa-Jong, Director of the K-MELLODDY Project, Korea Pharmaceutical and Bio-Pharma Manufacturers Association [August 2024]