Method Profile
Separate models predict activation/inhibition direction and affinity, followed by formulation selection and human-cell injury measurements.
Prioritize complementary molecular interventions and examine whether a selected coformulation changes measured injury-related outcomes.
23,204 annotated molecule–protein direction pairs and approximately 2.18 million affinity pairs; rat liver transcriptome reanalysis defines targets and human HepG2 cells test candidate formulations.
Independent classification/regression models with warm, protein-cold and scaffold-cold tenfold validation; HepG2 comparisons of single agents, free combination and coassembled formulation.
Computational evidence
Comparative Toxicogenomics Database direction labels; BindingDB/PDBbind affinity data; Open TG-GATEs rat liver profiles at 3, 6, 9 and 24 hours with time-matched controls.
Human-cell model
HepG2 cells exposed to 5 mM acetaminophen (APAP) for two hours, then treated for 24 hours with an equivalent hesperidin concentration of 42.4 μg/mL.
Candidate forms
Free hesperidin, free ursodeoxycholic acid, their physical mixture and coassembled HES–UDCA nanomedicine.
Keep the direction and affinity datasets/models separate. Compare generalization splits before interpreting the screen and preserve the distinction between human-cell, mouse macrophage and mouse whole-animal evidence.
Compare model generalization and test whether coassembly changes measured cellular protection relative to a free combination.
23,204 direction pairs; approximately 2.18 million affinity pairs; separate HepG2 cell-line experiments.
Pair-random warm start, unseen protein and unseen scaffold splits; free molecules/physical mixture versus HUNMs in HepG2 injury assays.
Prediction performance fell in protein/scaffold cold-start tests. HUNMs increased uptake and improved viability/GPX responses relative to free combination in the APAP injury assay.
Independent direction and affinity models prioritize complementary molecules for liver-injury research. A coassembled hesperidin–ursodeoxycholic acid formulation improves selected HepG2 responses, while cold-start prediction and mechanistic coverage remain limited.
