AI in drug discovery has shown limited clinically relevant impact despite decades of development, highlighting the need for improved translational focus.
- 7 cites
Moving in Bioinformatics, Nature Reviews Drug Discovery.
Rebuilt
AI in drug discovery has shown limited clinically relevant impact despite decades of development, highlighting the need for improved translational focus.
PCIM-DTA achieves state-of-the-art performance in drug-target affinity prediction under various cold-start scenarios, including cold-drug and cold-target conditions.
Newest first · last 60 days
A two-step machine learning approach enables transferable prediction of small-molecule retention times across different chromatographic conditions, outperforming existing methods.
The Virtual Biotech, a multi-agent AI system, improves drug development success rates by 48% and reduces adverse events by 32% through integrated analysis of biological data.
11.3% of randomly modified ligands across six targets improved potency tenfold or more, revealing a surprising baseline for ligand optimization success.
Spurious model comparisons are prevalent in biomedical AI, with nearly two-thirds of studies using invalid tests that inflate false-positive rates.
TDiMS descriptor outperforms existing methods by accurately capturing nonlocal intramolecular interactions with 20% higher predictive accuracy in property prediction tasks.
Machine learning identifies key chemical features, including nitrogen-containing aromatic scaffolds, that enhance mycobacterial outer membrane permeation, improving antibiotic potential.
Benchmarking foundation models in biomedicine reveals significant challenges in testing their limitations and evaluating their utility.
Gen-COMPAS enables efficient sampling of biomolecular transition pathways within nanoseconds to microseconds, surpassing traditional methods by reconstructing key states without predefined variables.
Journals this month
Moving areas, week to 3 Oct 2026