• CANDID-CNS™ is the new state-of-the-art AI model for BBB penetration: it achieves 87% AUPRC on bRo5 small molecules compared to 56% by Pfizer’s CNS MPO.
• CANDID-CNS™ distinguishes CNS penetrant stereoisomers at 68% compared to 50% by Pfizer’s CNS MPO.
• Published in the Journal of Chemical Information and Modeling, CANDID-CNS™ presents a breakthrough in AI Drug Discovery by unlocking the potential of bRo5 and stereochemistry to significantly expand the chemical space of small molecules that can be developed into CNS therapeutics.

1910, the only AI-native biotech pioneering small and large molecule therapeutics discovery, today announced that its CANDID-CNS™ AI model has been published in the Journal of Chemical Information and Modeling (JCIM), a journal of the American Chemical Society.

This press release features multimedia. View the full release here: https://www.businesswire.com/news/home/20251124252059/en/

1910's CANDID-CNS™ AI model has been published in the Journal of Chemical Information and Modeling (JCIM), a journal of the American Chemical Society.

The publication, titled “CANDID-CNS™: AI Unlocks Stereochemistry and Beyond Rule of 5 to Predict CNS Penetration of Small Molecules,” presents a first-of-its-kind AI model that accurately predicts blood–brain barrier (BBB) penetration for Beyond Rule of 5 (bRo5) molecules and incorporates stereochemistry – two roadblocks in neuroscience drug discovery. The publication is available here.

CANDID-CNS™ addresses one of the hardest problems in drug discovery: predicting which molecules can cross the blood–brain barrier. The BBB blocks ~100% of large molecules and more than 98% of small molecules from entering the central nervous system (CNS), making neuroscience the most difficult therapeutic area in pharma R&D. Most approved CNS drugs are small molecules that conform to Lipinski’s Rule of 5, while bRo5 compounds – larger, more complex molecules – represent an untapped chemical class with significant potential to unlock undruggable targets. Yet, these bRo5 molecules face three key challenges: exclusion from most medicinal chemistry designs, difficulty penetrating the BBB, and the inability of existing computational methods to predict their CNS penetration. Stereochemistry further influences BBB permeability, but current computational models fail to capture it. CANDID-CNS™ overcomes these limitations by accurately predicting BBB permeability for bRo5 molecules and learning stereochemical distinctions that govern CNS penetration.

“Neuroscience has long been defined by what we can’t reach,” said Jen Asher, Ph.D., Founder and CEO of 1910. “CANDID-CNS™ expands the boundaries of what’s considered druggable in the brain. By overcoming the limitations of bRo5 design and learning stereochemical effects, it opens an entirely new bRo5 chemical space for CNS drug discovery – bringing us closer to effective treatments for diseases like Alzheimer’s, Parkinson’s, and ALS.”

CANDID-CNS™ employs an attentive graph neural network (GNN) architecture that outperforms Pfizer’s CNS MPO score on predicting CNS penetrant bRo5 molecules (87% AUPRC vs 56%), distinguishing CNS penetrant stereoisomers (68% AUROC vs. 50%) and selecting CNS penetrant molecules from 1910's proprietary repository (90% AUROC vs. 81%). CANDID-CNS™ directly contributed to the discovery of 1910-102, a non-opioid, covalent small molecule inhibitor for chronic pain, a program partially funded by the NIH's Helping to End Addiction Long-term (HEAL) Initiative within the National Institute of Neurological Disorders and Stroke (NINDS). Please see 1910's Pipeline for more information.

“CANDID-CNS™ does not just classify molecules – it recovers the physicochemical principles that drive BBB transport,” said Jesse Collins, Ph.D., Senior AI Research Scientist at 1910 and lead author of the JCIM publication. “Its predictions correlate with quantum mechanical hydration free energy, indicating that the model implicitly learns the thermodynamic determinants of passive permeability. That mechanistic signal enables CANDID-CNS™ to generalize and identify brain penetrant bRo5 molecules and stereoisomers.”

CANDID-CNS™ is just one of ~100 AI models in 1910's ITO™ platform, which integrates massive multimodal data, frontier AI models, and high-throughput lab automation to identify novel disease targets and design small and large molecule therapeutics better and faster than traditional approaches. With three core capabilities spanning AI-driven Precision Target ID, AI-driven Molecular Design & Optimization and Federated Learning, 1910’s ITO™ platform is a first-of-its-kind Multimodal AI Platform for Modality Agnostic Drug Discovery™.

About 1910

1910 is the only AI-native biotech pioneering small and large molecule therapeutics discovery by integrating massive multimodal data, frontier AI models, and high-throughput lab automation into an infrastructure for AI-enabled drug discovery.

View source version on businesswire.com: https://www.businesswire.com/news/home/20251124252059/en/