Gvt commences 'Machine Learning AI Project' for new drug dev

MOHW and MSIT have selected 26 projects…will fund KRW 34.8 billion over 5 years until 2028

The government will actively pursue a project for faster new drug discovery.

 

It will fund KRW 34.8 billion over 5 years, from 2024 to 2028, and conduct a project utilizing federated learning of new drug discovery data owned by pharmaceutical companies, university hospitals, research institutes, and corporations.

 

The Ministry of Health and Welfare (MOHW) and Ministry of Science and ICT (MSIT) announced on July 23rd that 26 projects have been selected for the 'Machine Learning Ledger Orchestration for Drug Discovery (K-MELLODY)' project.

 

The MOHW and MSIT are jointly promoting the MELLODDY project, which aims to reduce the cost and time for new drug discovery using a federated learning-based AI model that can safely protect data while exploring.

 

Federated learning is a decentralized learning approach that enables machine learning to be trained internally on data from multiple locations, such as private and institutes, without directly sharing.

 

Analyzed data are sent to a central server.

 

The government has commenced this project and designated the Korea Pharmaceutical and Bio-Pharma Manufacturers Association (KPBMA) to run the project.

 

The project group has selected 26 subprojects and research institutes for each project through a public survey and evaluation.

 

The selected subprojects are categorized into three fields: ▲Establishing Federated Drug Discovery (FDD) platform, ▲Utilizing new drug discovery data and quality management, and ▲Developing AI solution.

 

First, for 'Establishing Federated Drug Discovery (FDD) platform,' project has been selected to establish FDD platform to enable machine learning safely without sharing data between institutes and keeping privacy.

 

As a supervision agency, Evident Inc.

 

has been chosen.

 

Also, 20 institutes have been selected as supervision agencies for AI-driven drug discovery.

 

They will share new data from the fields of drug metabolism·toxicity testing and pharmacoepidemiology.

 

Pharmaceutical companies, universities·hospitals, research centers, and corporations are included.

 

Eight pharmaceutical companies have been selected, including Daewoong Pharmaceutical, Dong-wha Pharm, Samjin Pharm, Yuhan Pharm, Jeil Pharm, Hanmi Pharm, Huons, and JW Pharmaceutical.

 

Universities·hospitals include Gachon University, Catholic University of Korea, Kyungpook National University, Seoul National University, and Seoul National University Hospital.

 

Participating research centers·foundations include Daegu Gyeongbuk Advanced Medical Industry Promotion Foundation, KRIBB, Institut Pasteur Korea, and KRICT.

 

Selected corporations include CIMPLRX and Apace.

 

For 'developing AI solution,' five institutes, including GIST, Mogam Institute for Biomedical Research, AIGEN Sciences, Chonbuk National University Industrial Cooperation Foundation, and KAIST.

 

They will develop the 'Federated ADMET Model (FAM)' to aid in drug candidate discovery, using experimental data from all stages of new drug discovery.

 

The MOHW and MSIT will allocate KRW 34.8 billion to these selected projects over a five year period, from 2024 to 2028.

 

They anticipate establishing an AI-driven new drug discovery infrastructure by collaborating with the government, pharmaceutical companies, research centers, and universities.

 

"By developing machine learning with a federated learning approach that safely shares and uses data, previously challenging large-scale data analysis and use will be enabled, accelerating drug discovery," Ko Hyeongu, MOHW's Advanced Healthcare Support Officer, said.

 

Ko said, "We will strengthen the data usage system and support AI‧Data R&D.

 

We will also strive to lead future healthcare and pharmaceutical innovation and promote citizen's welfare." "The federated learning approach will enable the safe utilization of high-quality drug discovery data accumulated from multiple institutes to develop an AI-driven drug discovery platform and machine-learning solution," Kwon Hyeonjun, MSIT's R&D Policy Director, said.

 

"We will support the R&D of digital bio-health businesses, which converges digital technology, and creating value so that South Korea emerges as a hub for cutting-edge biotechnology."