<aside>

This is the third part of my deep dive into why qPCR & phenotypic testing need to be replaced—and what comes next. In Part I, I covered the limitations of current diagnostics and the urgency of moving beyond slow, incremental improvements. Part II focused on why targeted Next-Gen Sequencing (tNGS) could be the answer, offering both rapid identification and resistance profiling in a single test.

Now, in Part III, I’m tackling one of the biggest challenges to making tNGS clinically viable—automation. If sequencing is ever going to match qPCR in accessibility, we need to radically simplify sample and library prep while maintaining the flexibility that makes sequencing powerful in the first place.

</aside>

<aside>

The Challenge of Automating Sequencing for Clinical Use

Right now, if you want to do molecular diagnostics in a hospital, you use qPCR. It’s fast, simple, and has been the standard for decades. Cepheid, one of the biggest players in the space, has been around for 26 years, with 40,000 installed GeneXpert instruments and $2 billion in revenue. Their system works because it’s designed for clinicians, not lab scientists—drop in a sample, press a button, and get results. No complex prep, no custom protocols.

But Cepheid isn’t the only player anymore. Indian company Molbio Diagnostics has taken a similar approach. Their Truenat system is a real-time RT-PCR platform, designed for low-resource settings. It’s battery-operated, IoT-enabled, and already deployed in ~8,000 hospitals and clinics across the world. More importantly, the cost per test is competitive—$7.90 per sample, similar to Cepheid’s $8 cartridges.

It’s clear that automation works when the workflow is simple. Truenat, like GeneXpert, is built around one test, one result, one action. No batching, no custom protocols, no flexibility. That’s exactly why it succeeds in clinical settings.

</aside>

<aside>

Why Isn’t Sequencing as Simple as qPCR?

Sequencing is messy. It’s not just about running a machine—it’s about sample prep, library prep, sequencing, and analysis. Each of these steps requires specialised reagents, workflows, and trained personnel. That’s why sequencing is still mostly confined to research labs and high-end clinical centres. It’s powerful, but not simple.

People have tried to automate sequencing workflows. Pipette robots are one approach. These are flexible and can automate parts of the process, but they require maintenance, trained staff, and custom protocols. They also don’t eliminate complexity—they just shift it. Instead of pipetting by hand, you program the robot to pipette for you. If you need full automation, it’s still not an out-of-the-box solution.

Another approach is closed-box automation—machines like MagicPrep NGS that handle everything for you, as long as you use their kits and follow their protocol. It works well, but locks users into fixed workflows. If you need to change anything—multiplexing, target selection—you’re out of luck.

Then there are digital fluidics platforms like Illumina’s NeoPrep or ONT’s Voltrax. These are elegant in theory—miniaturising reactions into a microfluidic chip—but expensive and limited in adoption. They work, but they haven’t become standard.

One of the newest attempts to tackle this problem is Volta Labs’ Callisto System. Instead of using pipettes or cartridges, Callisto uses electrowetting-based digital fluidics—tiny droplets move around a chip, performing all the steps of DNA extraction and library preparation without physical pipetting. It’s a brilliant idea. But it’s still a $125,000 benchtop instrument, which makes it out of reach for most hospitals and small labs. It’s automation, but not yet clinical automation.

Meanwhile, Cepheid and Molbio still own the clinical molecular diagnostics market with their simple cartridge-based qPCR systems.

</aside>

<aside>

Why Not Just Build a Sequencing Cartridge Like Cepheid?

Technically, there’s nothing stopping someone from building a cheap, plastic, moulded cartridge with preloaded reagents that automates sequencing prep. But the problem isn’t technical—it’s demand.

The whole reason sequencing is powerful is its flexibility. Labs want to run different assays, multiplex samples, and tweak protocols. A single-use cartridge that locks them into a fixed workflow isn’t what they need. That’s why no one has built a sequencing equivalent of Cepheid’s GeneXpert—it would be a niche product at best.

The real problem isn’t the lack of automation. It’s that sequencing users don’t want the kind of automation that qPCR users need.

</aside>

<aside>

Where Does That Leave Us?

There’s still room for improvement. The transition from manual prep to automated sequencing isn’t fully solved. The right solution probably isn’t a single cartridge system, but something modular—automating the most painful parts while keeping flexibility where it’s needed.

Maybe that means hybrid systems that allow some customisation. Maybe it means smarter automation that adapts protocols based on sample type. But if we’re looking for a one-size-fits-all, push-button solution for sequencing, we may be searching for something that clinicians don’t actually want, and sequencing labs don’t actually need.

That’s why automation in sequencing isn’t a solved problem. Not because it’s too hard, but because it’s not clear what problem needs to be solved.

</aside>

<aside>

If you’re working on automation for sequencing, diagnostics, or AMR detection, I’d love to hear your thoughts. Where do you see the biggest gaps? What’s missing from current solutions?

Drop a comment or DM me—let’s figure this out together. 🚀

</aside>