Enhanced Discussion on Reassessing Lipinski's Rule of Five in the Era of AI-Driven Drug Discovery
Uppula Purushotham · Advanced Pharmaceutical Bulletin · 2025
Artificial intelligence (AI) is redefining drug discovery, prompting reevaluation of established frameworks like Lipinski's Rule of Five.Although Ro5 has been a key component in the prediction of oral bioavailability, its strict use in the AI era continues the danger of inhibiting innovation by omitting promising compounds that challenge established norms.Young recently examined Ro5's limitations and historical context in a 2023 Expert Opinion on Drug Discovery article 1 .The paper underscores that the rule's 90th-centile thresholds are often misinterpreted as absolute limits.This misinterpretation has led to the exclusion of viable compounds, especially in emerging therapeutic areas that require larger molecules such as protein-protein interaction inhibitors, PROTACs, and macrocycles.Young highlights that molecular weight (MW), the most frequently violated Ro5 parameter, is increasingly irrelevant in modern drug design, as evidenced by the success of beyond Ro5 (bRo5) drugs with MW >700 Da.These molecules leverage advanced design principles, such as conformational flexibility (chameleonicity) and transporter-mediated uptake, to achieve bioavailability. 1 Figure 1: Percentage increase in key molecular descriptors of FDA-approved oral drugs between 1994-1997 and 2013-2019.