Immunobiome, AI Drug Development Related Paper Published in International Academic Journal

Immunobiome, a microbiome-based new drug development company, has succeeded in predicting the likelihood of drug success and side effects before clinical trials using machine learning. This is expected to significantly reduce time and costs by allowing early estimation of the causes of failure in clinical trials and the approval status of drugs.

On the 18th, Immunobiome announced that research results related to predicting clinical success probability using artificial intelligence (AI) were published in 'eBioMedicine,' a sister journal of the world-renowned medical journal 'The Lancet.'

The title of the paper is "Drug approval prediction based on the discrepancy in gene perturbation effects between cells and humans."

The research team of Immunobiome and Professor Sangwook Kim from the Department of Life Sciences at POSTECH developed an AI technology that predicts clinical success rates by focusing on the risks associated with drug target or marker genes and learning the differences in gene risk between cell lines and humans. This is expected to play a key role in designing successful clinical trials for selected candidate drugs. It can not only predict clinical success rates but also identify genes related to side effects and their characteristics.

Unlike existing AI models for new drug development, this technology is applicable to microbiome therapeutics. While conventional AI models for drug development are generally applicable only to compounds or proteins with well-known structures, Immunobiome’s technology is based on gene information responsive to drugs, making it effectively applicable even to strains with complex material properties.

An Immunobiome representative explained, "Many drugs deemed safe in the preclinical stage fail in actual clinical trials due to side effects. This technology combines a gene-centered approach reflecting human genetic differences with the chemical approach of drugs to more accurately predict drug safety and success probability."

This AI technology is expected to generate high synergy when integrated into Immunobiome’s proprietary new drug development platform, 'Abatiome.' Abatiome is a technology that comprehensively evaluates the therapeutic efficacy of strains and their MoA (mechanism of action). Using the precise gene response information derived from Abatiome, it is possible to predict the success rate of clinical trials.