Highlights from Discngine Meetup Vol. 5: Pushing boundaries in peptide discovery with innovative science and technology
Since 2021, Discngine has hosted its annual online Meetup, bringing together early drug discovery scientists and IT experts to discuss recent developments and challenges in research projects.
This year, the event celebrates its fifth edition, with the theme “Peptide Discovery: Pushing Boundaries with Innovative Science and Technology.”
The agenda featured four industry experts who presented their ongoing work and approaches in peptide design and discovery. The presentations were then followed by a roundtable discussion, during which the speakers compared experiences and reflected on the advancements and hurdles shaping this rapidly evolving field.
150+ participants | 5 speakers | 3 hours of live discussions
Peptide Therapeutics: A Growing Focus in Drug Discovery
The peptide therapeutics market is experiencing remarkable growth, driven not only by blockbuster drugs (including two new approvals in 2024), such as Semaglutide and Tirzepatide, but also by its expansion beyond metabolic diseases into areas such as oncology, infectious diseases, CNS disorders, and immunotherapy.
However, working with peptides presents inherent challenges.
Peptides occupy a unique space between small molecules and biologics: they are structurally more complex than traditional small molecules, yet not as large or well-defined as biologics. This intermediate nature makes their informatics particularly demanding.
Managing peptide-related data often requires combining chemical and biological approaches, extending small molecule tools and rethinking existing cheminformatics tools and workflows.
For example, tasks such as sequence alignment or structural registration can be difficult to handle in traditional small-molecule registration systems or biologics-oriented sequence databases, as neither is optimized for peptides.
3D Model of a Bicyclic Peptide (Visualized in 3decision)
To further explore the topic, Discngine Meetup Vol. 5 featured an insightful agenda that included:
A plenary session on therapeutic peptide generative models, scoring functions and retrosynthesis
A presentation on advancements in peptide late-stage functionalization for drug discovery
In this article, we provide a summary and key takeaways of each of these discussions and the recordings available to replay.
Therapeutic peptide generative models, scoring functions, and retrosynthesis
Our first speaker, Dr. Leonardo De Maria, Principal Scientist at AstraZeneca, opened the plenary session by giving an overview of therapeutic peptides and their synthesis, describing them as synthetically accessible biologics that sit at the interface between chemistry and biology.
He explained how the incorporation of non-natural amino acids (NNAAs) can expand the chemical diversity of peptides and improve their therapeutic potential, though this area remains only partially explored.
Dr. De Maria then shifted to the computational side of peptide discovery, introducing generative models and explaining how SMILES (Simplified Molecular Input Line Entry System) strings can be used to represent peptide structures. He presented PepINVENT¹, a generative model-based framework for de novo peptide design, built as an extension of the small molecule design tool REINVENT². PepINVENT combines a chemistry-aware pretrained model with reinforcement learning to propose valid, novel and diverse peptide designs.
Watch the full session recording here
He noted that, while scoring functions are available for natural peptides, there are none that can handle peptides containing non-natural residues, which makes optimizing these molecules difficult.
To address this gap, he and his colleagues developed the CamSol-PTM method³ as a scoring function to predict the solubility of peptides with modified amino acids and integrated it into PepINVENT to help guide it towards designing peptides with better solubility profiles.
Expanding Chemical Space for Macrocyclic Peptides
Building on the theme of expanding peptide design capabilities, our second speaker, Jennifer Hanisak, Associate Principal Scientist, Peptide Discovery, Merck & Co., gave a presentation on Advancements in peptide late-stage functionalization for drug discovery.
During her presentation, she highlighted the role of non-canonical amino acids (ncAA) in improving peptide stability, potency and solubility, while also pointing out the necessity and challenges of accessing a diverse and novel range of ncAA building blocks.
Jennifer Hanisak presented new synthetic and computational approaches to expand the accessible chemical space for peptide design, including high-throughput screening and late-stage functionalization techniques.
A key focus was a pyridinium-based coupling method that enables efficient modification of macrocyclic peptides under mild conditions, compatible with both protected and unprotected peptides.
Watch the full presentation recording here
This approach helps overcome limitations in ncAA availability, broadens structural diversity, and opens new possibilities for multi-dimensional structure–activity relationship (SAR) studies and future peptide optimization.
Roundtable discussion: Future Outlooks on Peptide Discovery
The event continued with an insightful roundtable discussion on the future of peptide discovery, the challenges that persist and best practices to overcome them.
Our chair, Dr. Leonardo De Maria, was joined by 3 expert panelists: Jennifer Hanisak, Dr. Alec Flyer, Associate Director, Novartis Institutes for BioMedical Research (NIBR) and Dr. Markus Kossner, Scientific Services Manager at Chemical Computing Group (CCG).
Together, they brought unique perspectives to the discussion, touching on key issues such as improving the oral bioavailability of peptides, enabling intracellular delivery and advancing computational modeling to better understand and optimize backbone modifications, topics that resonated with the audience and raised questions.
Replay the full roundtable discussion here
The panel also noted that many existing SAR tools were developed for small molecules or natural peptides and are not well-suited for more complex peptide formats, such as cyclic peptides or those containing non-natural amino acids. This creates an opportunity for developing SAR tools that can accurately capture the complexity of modern peptide therapeutics.
We ended this year’s Discngine Meetup with questions from the audience, which made for a lively and engaging discussion.
Thank you to all who took part in this year’s event. We look forward to welcoming you at Meetup Vol. 6!
Stay updated on our upcoming events, scientific blog articles and other activites by subscribing to our newsletter!