It has been nearly 25 years since a now-seminal proteomics paper first demonstrated the potential of proteome-scale biology - a milestone that ultimately helped pave the way for platforms like HuProt™. To mark this quarter-century journey from early yeast kinase arrays to today's comprenhensive, functional human proteome microarrays, Heng Zhu shared his blog reflecting on how the field has evolved and where it's heading next.
In February 2001, I co-authored a paper describing what was, at the time, among the first proteome-scale microarrays. Looking back nearly 25 years later, what strikes me most is how far the field has come, and how ideas that once required enormous manual effort have since grown into robust, scalable technologies.
Proteome-wide interrogation was not a distant aspiration. It was something we demonstrated could be done, long before automation, standardized workflows, or commercialization caught up. The question was never if proteome-scale biology was possible, but whether we were willing to do the work required to make it real.
| Learn how HuProt™ can power your next discovery |
The first array: 108 proteins and a year of work
My first protein array project was whilst I was working with Michael Synder at Yale University in the Department of Molecular, Cellular, and Developmental Biology. It involved constructing a nano-well protein array containing 108 purified yeast kinases. It “only” took me about a year to purify all of them.
When we published the work in Nature Genetics in 2000, it was considered the world’s largest protein array. Not because it was dense, but because every spot represented a different functional protein. Around the same time, other groups were publishing high-density arrays that achieved scale by spotting the same few proteins repeatedly on glass. Those were impressive demonstrations of printing technology, but they weren’t proteome biology. That distinction mattered to me then, and it still does.
Saying yes to the yeast proteome
Soon after, Michael asked me a question that would define the next phase of my career:
Could you make a yeast proteome-wide array?
I said yes.
Almost immediately, I realized what I had agreed to. The task required subcloning nearly 6,000 yeast ORFs into expression vectors using homologous recombination in yeast, followed by the development of a high-throughput method to purify ~5,800 individual proteins. At the time, none of this existed at scale.
Fortunately, I had the privilege of working with the exceptionally talented Ronda Bangham, a technician who helped complete the subcloning in record time. But cloning was only the beginning. Protein purification—especially at scale—was the real bottleneck.
The solution that changed everything
Yeast, like plants, have a tough cell wall. Back then, one common approach was to generate yeast protoplasts using Zymolase before lysing the cells — an effective but prohibitively expensive method. Another option involved grinding cells with zirconia beads under liquid nitrogen, which was difficult to adapt to a 96-well format.
Whether by design or serendipity, our breakthrough was the difference between an idea and actually having a functional proteome array, and came about from a casual conversation with colleagues discussing how they broke down rice leaves for DNA extraction. Having done my share of rice DNA preps, I knew how resilient plant cell walls could be. That conversation sparked an idea: if this approach worked for rice, why not yeast?
Brainstorming (plus lots of sketches!) led us to an innovation on how we could overcome this with yeast cells at scale, in a 96-well format. This approach changed everything, and became CDI’s standard operating procedure for high-throughput protein purification.
From yeast to human
After joining the faculty Johns Hopkins University, Jef Boeke, Seth Blackshaw and myself became convinced we could take the next step: creating a human proteome array.
We subcloned human ORFs into both yeast and E. coli expression systems but very quickly it became clear that yeast offered a much higher success rate for producing functional human proteins. After years of effort, along with Jef meeting with every department director at Johns Hopkins to convince them to invest in the project, we eventually purified ~20,000 human proteins, assembling what was then the world’s largest proteome array.
At that point, the original vision from 2001 was no longer theoretical. Proteome-wide interrogation had become a reproducible, scalable reality.
| Contact us for a consultation |
What’s changed—and what hasn’t
Over the past 25 years, the field has matured dramatically. Automation, quality control, standardization, and commercialization have transformed what was once an academic proof-of-concept into robust platforms capable of supporting large-scale biology.
Technologies like HuProt™ represent the industrialisation of that original idea, enabling comprehensive and unbiased antibody specificity assessment at a level that was unimaginable in the early days, facilitating new discoveries into the mechanisms of disease.
And yet, the core principle remains unchanged: If you want to understand biology without bias, you have to look at everything, not just what’s convenient to measure.
Looking forward
When I think back to purifying 108 kinases by hand, or struggling to crack yeast cells one plate at a time, I’m reminded that today’s “routine” technologies were once considered unrealistic, unnecessary, or impossible.
Proteome-scale biology has always been about persistence—about doing the hard, repetitive, often invisible work required to turn a vision into a tool others can rely on.
Twenty-five years later, the journey continues. And in many ways, we are still just getting started.
Learn how HuProt™ can power your next discovery - contact us for a consultation