End of Year Message 2025

SHARE

Another year draws to a close very shortly. How was 2025 for you? For Cresset we achieved major milestones including a significant technology acquisition, our first release of AI functionality, major improvements in our CADD platform Flare™, and we have, with the help of customers, been revisiting our mission and vision for the future.

In the summer we announced the acquisition of Molab.ai. Their novel ML algorithm shows great promise in generating high accuracy predictions based on customers’ own data sets and ADME data in particular. This technology is an important extension to our existing ML based methods and is an enabling factor in digital molecular discovery. One of the most pleasing aspects of this acquisition has been the way in which the people have come together across the two companies – it has been fantastic to see the positive, can-do attitude of a small AI based startup fit seamlessly with Cresset’s existing people and culture. The Molab technology has generated interest and I am looking forward to supporting more of our customers to enable this transformative technology in 2026.

This year we have continued our investment in AI technologies alongside our core physics-based methods. We believe that using AI to enrich our science brings the acceleration and deep insights that can be gained by AI without compromising on the fundamental scientific integrity of our approach, and I am sure this will be the cornerstone of our strategy for a long time to come. As I write, Cresset’s product team is rolling out the latest version of Flare which is packed with new functionality including AI. Alongside the co-pilots that help you effortlessly query the user manual, we have two exciting AI features that I think represent a major advancement in CADD software solutions. We have strengthened our unrivalled molecule ideation algorithms by adding AI generative chemistry. This can be left to roam wild or, more interestingly in my view, trained on an existing set of molecules and asked to generate something novel. I believe this method complements our molecule generation solution Spark™ very well, enabling users to draw inspiration for their next move from the widest possible range of chemistry.

Alongside our help bot and generative chemistry capabilities is another help-based AI assistant, trained on Flare’s extensive python API, that enables users to build custom functionality, customize the user interface, and create workflows. With the help of the AI co-pilot this becomes a rapid process, turning what was a specialist task into one that can be accomplished in minutes with little or no coding experience. I see this as the first step of a major evolution in how we all use software and I am looking forward to hearing user feedback now that this feature is widely available.

Every year we add new functionality into Flare and this year has seen major new features introduced. Highlights include the introduction of absolute free energy calculations and the enhanced relative free energy calculations which provide one of the key advances necessary to truly digitize molecular discovery. Alongside FEP, we continue to invest in our ligand-based methods, aware that not every project is amenable to structure-based approaches. This means that we now have new QSAR methods and (amongst many other improvements) a new ‘prevalence’ score that reports how commonly a particular fragment of bond has been synthesized. This score is useful when combined with the AI generative chemistry so that you can focus on makeable molecules, but I particularly like the visual representation – the parts of the molecule that are truly novel are highlighted giving you a clear indication on where to focus any synthesizability discussion.

Central to all our development is conversations with customers – we are driven by your problems and the feedback you give to us. This has been true through my (approaching) 24 years at Cresset but this year we went a little further – we asked you what you thought of us and what we do. This process helped us to commit to paper a new mission and vision statement. This was an interesting process for me, as we haven’t done it like this before. However, I am delighted with the feedback we received – it was everything that I have striven to achieve across the years. Look out for the formal release announcement and the reveal of our new branding in the new year.

It just remains for me to wish all our customers, partners and friends in the industry a happy and peaceful holiday season on behalf of the entire Cresset team. I’m very much looking forward to supporting your molecular discovery journey in 2026.

Happy Holidays,
Tim

Left: Ligand from PDB 3FI3 (JNK1 protein) used to generate training data for MolGenAI in Flare V11; Center: A high scoring AI generated molecule with good prevalence score (0.81); Right: AI generated molecule with low prevalence score (0.36) colored to show the reason for the poor score.

Subscribe & Don't Miss Out

Receive our newsletter to be among the first to hear about product releases, case studies, opinion articles, events and more.