Insilico teases cyclic‑peptide design workflow

Drug hunters often want a molecule small enough to slip into tight spaces but sturdy enough to hold its shape. Cyclic peptides try to do both by taking a short chain of amino acids and locking it into a ring, like tying a loose string into a bracelet. (nature.com) That ring shape changes how the molecule behaves. Reviews in Nature and Wiley say cyclic peptides can bind targets tightly and often resist breakdown better than straight-chain peptides, which is why drug companies keep revisiting them for hard protein targets. (nature.com) (onlinelibrary.wiley.com) The hard part is not drawing one cyclic peptide on a whiteboard. The hard part is sorting through huge numbers of possible ring sizes, amino-acid choices, and three-dimensional shapes without spending months making compounds that fail in the lab. (sciencedirect.com) Insilico Medicine is now saying its Pharma.ai platform has a new workflow aimed at that bottleneck. On the company’s Pharma.ai site, it is promoting a “2026 Q1 Spring Kickoff” webinar scheduled for Tuesday, April 14, from 10:00 to 11:00 a.m. Eastern time, and the teaser points to cyclic-peptide design inside its generative biologics stack. (pharma.ai) Insilico’s biologics page says its Biology42 system already combines more than 10 generative and predictive models with physics-based tools. The company says that setup is meant to generate novel biologic candidates and then score them for quality before researchers commit to more expensive experiments. (pharma.ai) That “physics-based” step is the part worth watching. In plain English, it means using structure and energy calculations as a filter, so the software is not just making hundreds of pretty guesses but trying to reject candidates that look unstable or unlikely to bind the target well. (pharma.ai) (onlinelibrary.wiley.com) Insilico has been pushing this idea before. In a November 2025 release, the company said Biology42 completed a peptide-design campaign against glucagon-like peptide-1 receptor, a hormone-related drug target, in 72 hours, which shows the company has already been framing speed as a selling point for peptide work. (prnewswire.com) If the new cyclic-peptide workflow really narrows a giant idea list down to a short bench-ready list, it could compress the earliest part of discovery. Insilico’s own pipeline page describes Pharma.ai as the engine behind target discovery, hit finding, and hit-to-lead work, so better triage at the front end would feed directly into those later steps. (insilico.com) The unanswered part is everything that happens after a ranking screen. Reviews of cyclic-peptide drug development still flag familiar problems such as synthesis complexity, cell permeability, bioavailability, and broader developability, which means a strong design workflow can speed selection without solving every downstream bottleneck. (onlinelibrary.wiley.com) (sciencedirect.com) So the near-term news is not a new approved drug or even a disclosed candidate. It is a platform company using an April 14, 2026 webinar to argue that cyclic peptides can move from a slow, artisanal search toward a faster software-and-filter loop, with the real test coming when the company shows wet-lab validation instead of teaser slides. (pharma.ai) (insilico.com)