Let's discuss a key aspect of vertical integration for diagnostics companies involved in #oncology testing:

Synthetic Biology ( #SynBio)

We think maximum operating leverage will accrue to those companies who have bought, or built, the ability to print their own DNA.

First, let's define the scope of this thread. I'm focusing in on a type of oncology testing called minimal residual disease (MRD).

MRD tests are given to patients who've just started treatment for an early-stage cancer. The goal of MRD is to monitor how well a patient responds.

A good MRD test identifies, sometimes years in advance, if a patient's tumor is recurring. As shown below, MRD testing happens between the green line (surgery) and yellow line (relapse), which hopefully shouldn't happen.

https://www.nature.com/collections/haffgaicaf

Here, we have our first fork in the road. There are two approaches to building an MRD test:

1. Tumor-Naive = 'off-the-shelf' panel with a rapid turnaround time

2. Tumor-Informed = a bespoke, personalized test that tracks patient/tumor-specific mutations, w/ slower turnaround

Both of these MRD tests are liquid biopsies. They scan tiny fragments of DNA in the blood to give clinicians the ability to see how their patients are responding to treatment.

Now, why did I say that personalized MRD tests are associated w/ slower turnaround time?

'Personalizing' an MRD test means that I first must analyze the patient's tumor tissue (which I can acquire from the initial surgery). Once I've decided on the mutations to track, I need to do use synthetic biology to create the patient-specific panel. I'll elaborate.

Let's say I want to track 40 patient-specific mutations scattered across the genome. I want to focus my sequencing solely on these regions, otherwise my sequencing costs will be too high and I will lose money. This is accomplished through targeted sequencing.

To target a mutation, I need a probe that targets that site in the genome. A probe is a tiny piece of DNA that targets and captures just that spot in the genome so I can focus my sequencing there and nowhere else.

It's not trivial to manufacture that DNA probe!

Of course, if I know the mutations I want to look at, I can order DNA from a list of vendors like IDT, QIAGEN, or Twist for example. Even though this may be cheap, I'm beholden to their manufacturing timeframe and I have to pay the sales markup.

This isn't a deal-killer, but it does increase my operating expenses and could make testing so slow that I'll be outcompeted by faster vendors.

However, if I print my own #DNA probes in-house, those last two issues go away almost entirely.

Now, it sure isn't easy to print DNA accurately and to scale. Companies that do this for enterprise also aren't small, so the opportunities to buy aren't exactly growing on trees. Building could work, but may require massive R&D investment.

Think of high-quality DNA like a manufacturing tolerance for machine parts. Little inconsistencies here and there matter A LOT--a few incorrect bases in a probe could restrict it from binding to its site on a patient sample, resulting in a false positive to the patient!

However, if I can print highly-uniform, cheap DNA at scale, I could gain a serious market advantage. Ostensibly, my test would be cheaper (if I choose to pass on savings instead of keeping the margin) and faster (by skipping 3rd-party manufacturing).

Closing the loop, you could ask why doesn't everyone just do a tumor-naive MRD test then, one that's just 'off-the-shelf'?

You could, but that's a whole separate discussion on the puts and takes between the two. I'll leave it by saying there's a market/use-case for both.