This article has been reviewed according to Science X's editorial process and policies. Editors have highlighted the following attributes while ensuring the content's credibility: Viral infections can be difficult to treat because viruses evolve rapidly, allowing them to develop resistance to drugs designed to attack viral proteins directly. One alternative is host-directed therapy (HDT), an emerging medical approach that treats infections by targeting the host's own biological pathways and associated proteins rather than attacking the pathogen directly.
Researchers at Korea University, led by professor Kwang Yeon Hwang of the College of Life Sciences and Biotechnology and associate professor Hong-Rae Kim of the College of Medicine, have developed an artificial intelligence (AI)-guided strategy to discover molecules that inhibit deoxyhypusine synthase (DHPS), a human enzyme involved in eIF5A hypusination, a process that supports the replication of a broad range of viruses. Unlike many AI drug-discovery approaches evaluated only computationally, the researchers experimentally tested AI-generated molecules. The study was published online in the Journal of Cheminformatics on Aug. 5, 2026.
"DHPS is involved in activating eukaryotic translation initiation factor 5A, or eIF5A, through a unique modification called hypusination. This process is important for the replication of a broad range of viruses, including HIV-1 and SARS-CoV-2, making DHPS a potential target for host-directed antiviral drug discovery," explained Hwang. The researchers systematically benchmarked 13 machine-learning models for predicting molecular binding scores and 10 methods for generating new molecules, using DHPS as a common target.
Out-of-fold stacking (OOF-Stacking) achieved the strongest overall performance among the prediction models tested, while RL-Design achieved the highest predicted hit rate of 92.02% among the generation methods tested, and 98.0% of its top-ranked molecules passed a standard drug-likeness threshold. To experimentally assess the AI-generated molecules, the researchers selected six fragments, small compounds of 140 to 190 daltons: three predicted-active and three predicted-inactive. All three predicted-active fragments inhibited DHPS, and none of the predicted-inactive ones did, showing complete agreement between computational predictions and experimental results.
Two of the active fragments gave measurable IC₅₀ values of 31.2 and 49.1 μM, within the typical range for initial fragment hits. An X-ray co-crystal structure also provided structural evidence for the binding of one AI-generated compound at the entrance of the DHPS substrate and NAD-binding pocket. "The best-known DHPS inhibitor closely mimics a natural cellular molecule, which brings a risk of off-target activity.
Our hits are fragments built on scaffolds unrelated to the known inhibitor. In fragment-based discovery, this is exactly where you want to start: they are small and bind efficiently for their size, which leaves considerable room to grow them into more potent compounds. However, these molecules are early-stage hits rather than drug candidates," noted Kim.
Extract — continue reading at the source.
