It has been nearly 25 years since a vet, on quest to gain a greater molecular understanding of disease, and a group of scientists, with a mission to produce a functional representation of the human proteome on a slide, met.  To mark this quarter-century journey from early yeast kinase arrays to today's comprehensive, functional human proteome microarrays, Ignacio Pino shared his blog reflecting on how infrastructure, people and funding came together to deliver the most comprehensive ‘proteome on a chip’.

From Vet Clinic to “Human on a Chip”: The Early Years of CDI Labs

When people see HuProt™ today, more than 21,000 full-length human proteins, individually expressed and quality controlled, it can feel like an inevitable outcome of academic proteomics. It wasn’t!

I wasn’t part of the original team building proteome arrays at Johns Hopkins University. I was practising veterinary medicine in Puerto Rico. Veterinary science and biotechnology have always been closely tied. Many vaccines and immune sera were first developed in animals long before they reached human medicine. But after years in clinical practice, I found myself asking different questions.

Being a veterinarian is rewarding. It is also, at times, repetitive. I am naturally inquisitive. I didn’t want to just treat disease; I wanted to understand what was happening at a molecular level. I began asking myself whether I should be doing something that extended beyond my veterinary examination room. Around that time, I was introduced, almost by chance, to a group of scientists working on something unusual: a functional representation of the human proteome printed on a slide.

My first reaction was: this is the whole human on a chip.

At the time, gene therapy was not yet dominant. Drug discovery focused on proteins: real, folded, functional proteins. Producing thousands of them in a reproducible format and screening them systematically would change how researchers approached targets altogether.
Recognising the idea was one thing; building it was another. A platform like that needs infrastructure, funding, and people prepared to commit for the long term.

So, the question became simple: where do we build it?

Why Puerto Rico Became the Home of a Proteome Array Platform

Puerto Rico has long been a hub for pharmaceutical manufacturing. Initially focused on small molecules, the island later attracted major biologics and medical device companies, including Becton Dickinson and Sartorius. The workforce here is highly trained, experienced in regulated environments.

But what Puerto Rico did not yet have was a research tools company operating at proteome scale.

Research tools were not fashionable investments at the time. Funding tended to flow toward therapeutic programmes searching for the next drug candidate with blockbuster potential. Building enabling infrastructure was harder to justify. Fortunately, local government agencies and non-profit organisations recognised the long-term value of research platforms and early seed funding and grants allowed us to begin.

Through a mutual connection, I was introduced to the founding scientists, including Joe Bonaventura, a retired Duke biochemist who had relocated to the island and became what I still think of as the “molecular glue” of the founding team. Joe understood the scientific potential and the urgency of building something durable.

CDI Labs was formed:

Addressing the Reproducibility Crisis with Proteome-Wide Antibody Specificity Validation

Our original goal was to demonstrate that this proteome-scale platform solved a real problem. At the time, the scientific community was becoming increasingly aware of the reproducibility crisis. Poorly characterised antibodies were contributing to inconsistent findings across thousands of studies. Funding bodies such as the National Institutes of Health recognised that unreliable reagents were undermining research they were supporting.

If we could screen antibodies against thousands of human proteins simultaneously, we could identify off-target binding early. In principle, we could help remove a major source of experimental variability from biomedical research by fully characterizing antibody specificity.

Scientifically, the argument was clear, but commercially, it was much more complicated.

It is difficult to persuade a company to screen its antibody catalogue when the outcome might reveal cross-reactivity. Even when the long-term benefit is obvious, the short-term implications are uncomfortable. That was our first lesson: being scientifically correct does not guarantee rapid market adoption.

Nevertheless, the platform proved its value. We generated data. We secured grants. We established credibility. Work continued with companies who recognised the need for better antibody validation. However, we knew this incredible tool had much more to offer. We started to consider another direction, where the value of a proteome-scale platform was becoming more and more evident.

From Antibody Specificity Validation to Autoantibody Seromics

While CDI Labs was building operations in Puerto Rico, our co-founder Heng was using the platform in an academic environment at Johns Hopkins University. An immunologist suggested using the array to profile autoantibody responses. We ran exploratory studies and the results were compelling.

Rather than evaluating whether an antibody reagent was specific, we were uncovering biological signals directly from serum samples. We were no longer diagnosing a problem in the research ecosystem; we were generating new knowledge.

From the outset, I did not see autoantibodies as universal standalone diagnostics. In some conditions, such as rheumatoid arthritis, they are highly informative. But more broadly, I viewed them as one layer within a multimodal framework. Genomics had already transformed cohort studies. Proteomics added another dimension. Autoantibody profiling could contribute an additional “omic” layer – Seromics.

So, we began to expand our analysis to other immunoglobulin classes with larger cohorts, more refined questions and more publications.

The pivot allowed CDI Labs to move beyond validation and into discovery. The proteome array was no longer only a quality-control instrument. It was becoming a discovery engine.

In our next article, we explore how that discovery engine evolved from high-content screening and mining the natural antibody repertoire to interactomics and AI-driven specificity modelling and what that means for the next 25 years of CDI Labs.