Picking the "Right" Technology, Part I

Do you remember multiple choice tests?

The hardest questions offered options "d) some of the above" and "e) all of the above." But you never got an "it depends" option, at least until now.

So, just like Jeopardy, here's the question answered by "it depends."

Which technology is best for immune monitoring and biomarker discovery?

Let's explore assays that study serum or plasma, in a simple and clear way, with a special focus on impact. I've color-coded the discussion points: green is positive, yellow requires some consideration, and red is a disadvantage. Summary points are highlighted, and we present a framework for figuring out which to use. A TL;DR (too long, did not read) is provided at the end for my lazy friends.

ELISA (Enzyme-Linked Immunosorbent Assays).

Easy and quick to perform.

Measures proteins that are functional mediators of an immune response.

ELISA samples - the serum or plasma fraction of blood - are easy to obtain and stable.

Assays and equipment are relatively inexpensive.

Data analysis is as easy as it gets.

But disease/immunity aren't often driven by a single circulating molecule.

(And if they are, the single driving molecule was often discovered ages ago.)

Performing multiple single analyte ELISAs gets old fast - labor-intensive and inefficient.

There's no way to know precisely which cell types produce the cytokine of interest.

Good for biomarker discovery & immune monitoring.

Poor for understanding biological mechanisms.

Multiplexed ELISA-like Assays.

These assays vary in complexity, time to readout (with Olink's sequencing assay

requiring the longest wait times), sensitivity, specificity, and cost.

Their ability to test for multiple analytes is a huge advantage over ELISAs.

Great for biomarker discovery & immune monitoring.

Better than basic ELISAs for understanding biology but still not great.

(Unless you use highly multiplexed assays which provide tons of data to infer mechanisms)

_________________________________________________________________________

How do you know which to select for your study? Here's a way to think things through:

DISCOVERY is your priority: Select from the multiplexed ELISA-like platforms. If you're looking to do the lab work yourself, go with Luminex or MSD... the lab work for these assays doesn't require much of a learning curve. If you are looking to sift through a crazy number of analytes, go for the highest end Olink system, and perhaps hire one of their certified providers to do the lab work. If you are planning to outsource, consider Nomic's offering.

As you are doing this work, remember... it's for discovery, so you don't need a fit-for-purpose validation yet. Your work in the discovery phase should be on a "training set" of samples, and you should reserve some samples - preferably collected as part of a distinct study as a test set to confirm results... you can do that work using a simpler technology. I would reserve a fit-for-purpose validation for settings that are less about discovery, and more about reproducibility, precision, stability, etc (i.e., clinical trials).

VALIDATION is my priority: Choose Luminex, MSD, or a simple ELISA. These are long-standing, well-established systems, so you'll have lots of published data to compare your validation study results... the reduced cost and relative simplicity of these systems set you up to translate your results more easily into a clinical test or companion diagnostic or an assay supporting FDA submissions, if and when that's needed.

TRANSLATIONAL IMPACT is my priority: This is, of course, not mutually exclusive with the above priorities, and one might argue that this should always be our #1 priority. In any case, when you have a unique set of samples, from a unique clinical setting, it's critical to maximize the information you obtain. We owe it to science and the patients.

So, start by using a highly-multiplexed assay as a screen. Once you find analytes that differ between patient groups and/or correlate to a clinical outcome, use the simplest possible assay (ELISA) to establish the sample matrices (serum, plasma) and conditions (time since draw, etc) appropriate for the test (that still identify the patient group differences). This is the kind of information you'll need for a clinically-relevant, real-world test.

TL;DR: There are a ton of technologies available to measure proteins in serum and plasma. They vary in how established and complex they are, and most notably how many proteins they can measure at once. These assays are amazing screening tools, well-suited for biomarker discovery, because they measure easily accessible plasma or serum, and because they are so readily translatable into clinical tests. But scientific and clinical impact often depend on novel findings, and the odds of inconsequential, me-too results can be high with these approaches. When/if you find novel correlates of disease or treatment, validate them with simple assays that allow you to examine all the possible ways a lab might receive a sample from the clinic.