Pay Attention to Sequence

The Hidden Pattern Behind Scientific Breakthroughs and Careers

“Sequence, sequence, sequence. When something doesn’t make sense, pay close attention to the sequence.”

My dad gave me this advice when I was about 16 years old. Perhaps I should clarify up front that he was not referring to genetic sequences, but rather the temporal order of events. Exactly why he gave me this advice…I honestly don’t remember. It was something going on in his life, not mine, and I generally stayed in my lane.

It’s been a sporadically useful piece of advice, and it popped up in my mind a few days ago. I always have a few trains of thought going at any given moment; in this moment I was simultaneously thinking about the lead-up to the discovery that DNA is the primary molecule of inheritance, and a talk I’m giving this week at Genspace about applying your skills to biotech automation. Sequence plays a quiet but immutable role in innovation.

Before it was shown that transferring DNA would transfer traits between bacteria, and that bacteriophages inject DNA into cells, there was a strong contingent amongst scientists who believed that protein, not DNA, would prove to be the molecule of inheritance.

It was a reasonable interpretation of the information available at the time. It was known that chromosomes were mediators of inheritance, and that chromosomes are made up of proteins and DNA. Proteins, which are built from varying combinations of 20 different amino acids, were known to have high diversity and complexity. This seemed suitable to the task of encoding the complexities of life.

By contrast, there are just four types of nucleic acid in DNA. How…simple. Far more likely, they thought, that this seemingly simpler molecule was a structural scaffold for the more complex proteins.

As compelling as the functional evidence was, DNA was not widely embraced by scientists as the molecule of inheritance until the structure of DNA was elucidated and the role of the sequences of nucleic acids was demonstrated.

Knowledge Shapes the Perception of Possible

A big part of the hang-up was around the idea that, given the complexity of different life forms, genetic material must also be complex. Proteins, built from 20 different amino acids in countless combinations and configurations, seemed like the obvious candidate.

I don’t think it’s a coincidence that this description of proteins sounds a lot like a very commonly used encoding system: the alphabet. In English, we use 26 letters to form a wide variety of words with different meanings and purposes. We use the written word to document what is in our minds, enabling our thoughts to last long enough into the future that others might read it as well. The so-called molecule of inheritance must do similarly by storing critical information to pass on to future generations.

The pivotal work demonstrating DNA as the genetic material was performed in the 1940s and 1950s. Computer programming existed, but it was still in its infancy. It was not something that biologists had at the top of their minds.

What if the development of computational technologies had been a few decades faster? If encoding information in 0s and 1s had already been shown to be so versatile and powerful, would it have seemed more plausible that the 4 nucleotides of DNA could encode for the diversity of life?

We know that diversity of knowledge is a crucial component of innovation. What is equally essential, though discussed far less, is the role of sequence: the timing of when different pieces of knowledge become available and intersect.

X-rays, Crystals, and the Shape of DNA

This timing principle becomes clear when we look at how the structure of DNA was actually discovered. While biochemical studies had demonstrated that DNA was the molecule of inheritance, the underlying structure and mechanism of DNA remained a mystery.

Meanwhile, through the 1930s and 1940s, the technique of X-ray crystallography (a tool for uncovering the shapes of molecules) began to be applied to biological molecules. By the 1950s, it was a developed technique that had successfully been used to reveal the structure of cholesterol and penicillin.

The timing of the biochemical inquiry around DNA and the technological progress of X-ray crystallography came together perfectly. Rosalind Franklin, a skilled X-ray crystallographer, captured the diffraction data for DNA that resulted in the discovery of its helical shape. The need and the tool converged at exactly the right moment.

Thank Jellyfish for Modern Cell Biology

A few decades later we got to see another example of this phenomenon. In the 1960s and 1970s, marine biologists were curious to understand how some types of jellyfish glow. This study of bioluminescence resulted in the discovery of green fluorescent protein (GFP), as well as the DNA that encodes it. One of the strengths of GFP is that it can be produced in many types of cells with little effect on the cell itself. In fact, you can add it to another protein and in many cases the protein continues to function as usual.

Meanwhile, astronomers and physicists were developing new tools for imaging. Lenses, mirrors, and filters all improved, but most impactful of all was the development of the charge-coupled device (CCD). CCDs paved the way for more sensitive detection of low levels of light, meaning that something didn’t have to be bright to still get a clear image of it.

In the 1990s these two paths came together. While GFP can be expressed in cells and tethered to proteins, it does not produce a lot of light per molecule. The combination of GFP and microscopes that were more sensitive to light transformed our understanding of cell dynamics. Suddenly, we could see where proteins were located and how their locations changed under different conditions. As these technologies matured further, we were able to observe additional colors of fluorescent proteins at lower levels and in shorter periods of time.

This led to a critical maturation in our understanding of how cells function. How our cell membranes are in a constant state of turnover. How molecules like kinesin and myosin move along structural elements in the cell in a way that reminds us of trains on a track. How calcium is a critical signaling molecule for countless pathways. How our development is able to go from a single cell to a complex organism, thanks to the complex dynamics of so-called cell fate ensuring that we have skin on the outside, bones on the inside, and a digestive tract to fuel the whole body.

Five → One (The Automation Convergence)

While there have, of course, been more examples over time of this dynamic of converging technologies and scientific inquiry, the incoming explosion of full automation in laboratories represents something unprecedented in scale. It’s the result of five different developments occurring in parallel: robotics, machine learning, artificial intelligence, statistics, and so-called big data research. If just one of these were out of sync, we could not see the emergence of self-driving labs that we see today.

This industry needs brilliant minds trained across disciplines: automation experts passionate about lab problems, biostatisticians seeking new ways to generate large datasets, and scientists who want tools that leverage their minds rather than tie up their hands.

Of these, expertise in the field is currently dominated by those with engineering, computer science, and math backgrounds. Strong representation by these disciplines is essential to prototyping these technologies, especially since, as a rule, early versions are less user-friendly than later iterations.

But whether I’m speaking to people building foundries, biotech hiring managers, or academic labs, I hear the same thing: the hardest roles to recruit for are those for scientists with automation experience. They don’t even need experience with the exact automation equipment the lab uses, just something that demonstrates a real willingness to work with automated processes.

Diverse training and skill combinations provide multifold value. In multidisciplinary spaces, problems arise within individual domains or at their intersections, requiring experts who can quickly identify issues and efficiently apply solutions. You also need people who have experienced the pain points in ways that vary widely from one another; this is how robust solutions are developed. Teams need people whose multidisciplinary exposures occurred in different sequences, so that you get the benefits of deep expertise and recent learnings.

The Convergence Window is Open Now

The shortage of scientists with automation experience we’re seeing today signals something much larger: the birth of a new industry ecosystem that will demand skills we’re only beginning to understand.

It’s like the early days of the internet. The first wave needed programmers who could build websites with basic HTML. But as the technology matured, entirely new roles emerged: user experience designers, content strategists, growth hackers, social media managers.

Laboratory automation will have its own inflection point. The companies building self-driving labs today will need more than engineers, scientists, and data scientists. They will also need to fill traditional business roles, but with a twist. When selling an entirely new way to work, sales reps must understand more than features and pricing; they need to grasp how automation changes experimental workflows and why protocols fail during transition. The marketing managers need to be able to create educational content that bridges decades of scientific practice with cutting-edge robotics and AI.

This creates a unique opportunity for another type of sequence: careers. Those with specialist knowledge in a core business domain combined with cross-functional knowledge about automation will thrive as the industry begins to grow. A graduate student who spent summers working in an automated core facility. A lab manager who implemented their first liquid handling robot. A consultant who helped biotech companies evaluate automation vendors. A science communicator who’s passionate about emerging technologies.

The lack of formalized training in lab automation today is a double-edged sword. The downside is that it’s less clear what training to pursue. The upside is that experience counts more than expensive certifications and degrees. The curious and tenacious have the advantage.

The Sequence of What’s Next

The convergence we’re witnessing today mirrors the historical patterns, but with one crucial difference: scale. When X-ray crystallography met DNA research, it transformed our understanding of genetics. When GFP met sensitive imaging, it revolutionized cell biology. The current convergence of robotics, machine learning, artificial intelligence, statistics, and big data research promises to transform entire industries.

Yet convergence alone doesn’t guarantee transformation. The DNA structure breakthrough required Rosalind Franklin’s expertise in crystallography applied to biological problems. The cell biology revolution needed researchers who understood both fluorescent proteins and imaging technology. Today’s automation revolution will require people who can bridge the gap between cutting-edge technology and practical laboratory needs.

The shortage of scientists with automation experience we’re seeing isn’t just a hiring challenge. It’s a signal that we’re approaching the same inflection point that defined previous technological revolutions. Just as the early internet created entirely new roles that nobody had predicted, laboratory automation is creating opportunities that blend technical depth with domain expertise in ways we’re only beginning to understand.

This much is clear: five technologies have converged, early technical barriers are falling, and the field is approaching the moment when diverse skills become more valuable than pure technical expertise. The question isn’t whether these opportunities will emerge. It’s whether we’ll recognize them while they’re still taking shape.

Pay attention to the sequence, and choose your next steps wisely.