Discovery of TYR inhibitors from de novo molecular generation to dual-track lead optimization: 'Competition' between AI and chemists.

Clinical Snapshot

Key Findings

• P Value: Not reported in the abstract

• Effect Size: Not quantitatively specified in the abstract; described qualitatively as 'most potent' and 'notable antipigmentation activity' — specific IC50 values require full paper review

• Primary Outcome: Identification of AI10-m15 as the most potent TYR inhibitor from the expert-guided optimisation track, and AI10-a2 as a potent inhibitor with unconventional chemotype from the AI-driven track; both demonstrated antipigmentation activity in cellular models with acceptable cytotoxicity profiles

• Nnt Or Sensitivity: Not applicable at this preclinical stage; relevant metric would be IC50 (enzymatic) and EC50 (cellular antipigmentation) — values not provided in abstract. Comparison to kojic acid (standard reference) IC50 ratio would be the appropriate effect measure

• Confidence Interval: Not reported in the abstract

Clinical Application

The computational-to-in vitro pipeline is feasible and resource-efficient as a drug discovery tool. However, progression to clinical feasibility requires: in vivo melanocyte/zebrafish model validation, rodent pharmacokinetic studies, dermal formulation development, regulatory toxicology package, and Phase I/II clinical trials — none of which are addressed in this study. No current PBS listing or TGA registration is relevant to these investigational compounds. Hyperpigmentation is a common dermatological concern in Australia, particularly given high UV exposure rates and diverse skin phototypes in the population. Existing TGA-approved topical agents include hydroquinone (restricted), kojic acid, azelaic acid, and tranexamic acid. Should AI10-m15 or AI10-a2 progress through the development pipeline, TGA evaluation under the Therapeutic Goods Act 1989 would be required. RACGP and ADSA (Australasian College of Dermatologists) guidelines on hyperpigmentation management would be the relevant clinical framework for eventual integration. Australian researchers and dermatologists should monitor this line of investigation but no practice change is warranted at this stage. Preclinical only — intended eventual target population is patients with hyperpigmentation disorders (melasma, post-inflammatory hyperpigmentation, solar lentigines). No clinical application is currently supported by this evidence.

Abstract

This study introduces a unified framework combining artificial intelligence (AI)-directed de novo molecular generation with dual-track lead optimization-comprising expert-guided strategies and AI-driven pathways-to discover tyrosinase (TYR) inhibitors for hyperpigmentation disorders. Using a reinforcement learning (RL)-based generative model, the lead compound AI10 was identified. Subsequent optimization followed two parallel routes. The expert-guided approach yielded AI10-m15 as the most potent TYR inhibitor, with notable antipigmentation activity and excellent cellular safety profiles. In contrast, the AI-driven pathway explored broader chemical spaces, generating unconventional chemotypes, exemplified by the potent TYR inhibitor AI10-a2, highlighting AI's capacity to uncover nonintuitive activity cliffs despite greater output variability. Systematic comparison revealed that the AI model offers exploratory diversity, whereas expert-guided optimization provides predictable improvements in activity and developability. In summary, starting from an AI-generated lead and subsequently integrating both expert-guided and AI-driven structural optimization strategies, these findings further underscore that combining AI technologies with experts' medicinal chemistry insights can substantially accelerate the discovery of viable candidate compounds.

References

1. 1.Sun, Y., Wang, J., Chen, W., Jiang, X., Wang, S., Zhi, J., Li, F., Feng, M., Niu, X., Ju, B., Guo, J., & Bai, R. (2026). Discovery of TYR inhibitors from de novo molecular generation to dual-track lead optimization: 'Competition' between AI and chemists. Science Advances. https://doi.org/10.1126/sciadv.aeg0376