Benchmarking in silico degradation predictions against accelerated and long-term stability data
Meire Y Kawamura, Jack W Hodgson, Luccas Sanches, Valeria Vassiliades, Gabriel Greco, Ariane Rivellis , Natanael Segretti and Rachel Hemingway
A Lhasa whitepaper, published 22nd September 2026.
Forced degradation studies tell you what is chemically possible. They don’t tell you what will actually appear in your drug product two years into shelf life, and that is the space where regulatory decisions are made.
Our new whitepaper investigates this query. Building on the peer-reviewed study in Organic Process Research & Development, which established 70% sensitivity for Zeneth in the potential degradation space, this study benchmarks Zeneth predictions against 213 structurally elucidated degradants from 96 drug substances, all observed under accelerated and long-term stability conditions.
The results show an average sensitivity of 60%, covering both direct drug substance degradation and drug–excipient interactions. In an environment shaped by formulation, packaging, physical state and trace contaminants, identifying three in five real-world degradants before the lab is a genuinely useful place to start a stability investigation.
“A 60% sensitivity in the stability space is a meaningful outcome. The drug product environment introduces variability that cannot be replicated in silico without assumptions. The fact that Zeneth covers three in five observed degradants, including excipient interactions, demonstrates a genuine use case for stability experiment design.” Jack W Hodgson, Scientist, Lhasa Limited
Access the whitepaper to learn more:
- See how in silico predictions performed against an external dataset of 213 degradants from accelerated and long-term stability studies, and why 60% average sensitivity is a meaningful result in this space.
- Understand where predictions hold up for drug substance degradation (61%) and drug–excipient interactions (59%), and what that means for formulation compatibility risk.
- Get a clear map of the potential, likely and actual degradation spaces, and the pathways that dominate real-world degradation: hydrolysis (36%) and oxidation (26%).
- See where predictions can cut structural elucidation effort while still meeting ICH Q3A/B expectations.