Oligonucleotide Manufacturing

Unlocking the Future of Medicine: Oligonucleotides, Innovation, and Machine Learning at CPI’s 2025 Symposium

Last week, I had the privilege of presenting at the Innovation in Oligonucleotide Manufacturing Symposium hosted by CPI at the Medicines Manufacturing Innovation Centre in Glasgow. It was an energising day of shared learning, visionary talks, and genuine collaboration across industry and academia. I came away excited—not just by the work we showcased, but by the broader momentum building toward smarter, more scalable, and data-driven oligonucleotide production.

But before diving into that, let’s take a step back…

---

🧬 What Are Oligonucleotides—and Why Do They Matter?

Oligonucleotides (or “oligos”) are short strands of synthetic DNA or RNA—typically between 10 and 100 nucleotides long. These molecules can bind precisely to a target sequence of genetic material. That makes them incredibly powerful tools for:

• Gene silencing (e.g., antisense therapies)
• Splice modulation
• Gene editing (CRISPR)
• RNA-targeted therapeutics

Unlike traditional small molecules, oligos work at the genetic level. That means they can be used to treat diseases once considered “undruggable,” like rare genetic disorders or certain cancers. Several oligo-based therapies have already reached the market, and dozens more are in development pipelines around the world.

Here’s a quick visual of how DNA strands—like oligos—pair up via complementary base pairing:

---

💡 My Talk: Machine Learning for Smarter Synthesis

At the symposium, I shared recent progress from our team on using machine learning to enhance solid-phase oligonucleotide synthesis (SPOS)—the dominant method for manufacturing these molecules. Our tool, Alchemite, uses adaptive design of experiments (DOE) to learn from process data and predict outcomes.

Here’s what we’re achieving:

• Higher yields
• Lower impurities
• Improved overall purity

This isn’t just incremental improvement—it’s a step toward commercial-scale, high-quality oligonucleotide production with lower environmental impact and faster development timelines. The conversation around Process Mass Intensity (PMI)—a key sustainability metric—was especially compelling. With machine learning, we can start optimising PMI systematically and early in development.

---

🤝 What the Industry Is Saying

CPI’s symposium brought together engineers, chemists, regulators, and biomanufacturers for a packed day of talks and panel discussions. A few standout takeaways:

• UK infrastructure is expanding with new GMP oligo facilities and a dedicated training academy opening soon.
• Modified oligos (like triazole-linked or PMO-based structures) are pushing the boundaries of therapeutic design.
• Enzymatic and hybrid synthesis methods may help the industry scale production without relying on hundreds of batches per year.
• Cross-functional teams are essential. You can’t just drop a small molecule chemist into an oligo line and expect success.

There was a clear sense that the field is maturing—from niche R&D to mainstream manufacturing challenge. As one panellist put it:

“Everyone wants to be the first to be second.”

De-risking innovation through collaboration, smart design, and public investment (like from Innovate UK) will be crucial.

---

🚀 Looking Ahead

We’re entering a golden era for oligonucleotide therapeutics. The technology is ready. The tools—from smart synthesis to ML-driven optimisation—are here. And the talent across the industry is committed to solving the hard problems of scale, quality, and sustainability.

I’m proud of the work we’re doing with CPI and excited to help lead this transformation—one base pair at a time.

If you’re working in oligos or want to collaborate on smarter synthesis approaches, let’s connect!

---

#MachineLearning #Oligonucleotides #Manufacturing #AdaptiveDOE