C.A.S.S.A.N.D.R.A. – Continuous AI-Driven Smart Screening & Analysis for Nanoparticle & Dynamic Recognition Automation

A new generation of medicines---such as RNA-based therapies and vaccines---relies on delivering genetic instructions directly into cells. This is typically achieved using lipid nanoparticles (LNPs): microscopic, fat-based spheres that encapsulate RNA and carry it safely to target cells. While LNPs have proven vital in recent advances, particularly during the Covid-19 pandemic, manufacturing them reliably and efficiently remains a major challenge.

At present, there is no way to analyse LNPs during production in real time. Pharmaceutical companies must wait until a batch is completed before using complex tools like electron microscopy to check whether the nanoparticles are the right size, shape, and properly loaded with RNA. If issues are found at this late stage, entire batches---often worth hundreds of thousands of pounds---must be discarded. With LNP-based medicines growing rapidly in importance, the sector urgently needs smarter, faster ways to monitor these particles during production.

In this project, InvenireX will develop a novel digital analysis platform to solve this problem. Our system will use artificial intelligence and computer vision to detect, count, and assess LNPs as they are made---measuring particle size, shape, and RNA content in real time. This builds on our existing technology, which already performs precise, fluorescent-based quantification of DNA and RNA. By adapting our AI imaging engine---originally built using the 'You Only Look Once' (YOLO) model---to work with LNPs, we aim to provide medicine manufacturers with a powerful new quality control tool.

The result will be a first-of-its-kind solution: a smart, in-line visualisation system for LNP manufacturing that could significantly reduce waste, lower production costs, and accelerate the rollout of advanced RNA-based treatments.

InvenireX is currently raising private investment to advance our core RNA and DNA quantification platform. However, adapting this technology for LNP analysis...