Model Proteins Could Bring Order to Protein Science

Protein scientists could improve reproducibility and coordination across the field by rallying around a small, shared set of "model proteins," according to a new Perspective by Connecticut College chemist Marc Zimmer.

The article appears in the 40th-anniversary issue of Protein Engineering, Design and Selection (Oxford University Press). Zimmer argues that protein science is ready to adopt a framework similar to the one that transformed research using model organisms such as fruit flies, mice, yeast and C. elegans.

Those organisms became powerful research tools not only because their biology is conserved, Zimmer notes, but because scientific communities coordinated around them. Shared protocols, databases and benchmarks made results easier to compare, reproduce and build upon.

Zimmer proposes extending that logic to the molecular scale. His framework calls for formally identifying a small group of "model proteins" already serving as informal standards across subfields, then pairing them with shared benchmarks, curated reference datasets and minimal reporting requirements.

Protein researchers often struggle to reconcile studies that rely on different experimental assumptions or measurements, Zimmer said. A model protein system would reduce that friction by giving the field common reference points.

The Perspective suggests beginning with five widely used proteins: green fluorescent protein (GFP), lysozyme, hemoglobin and myoglobin, RNase A and bacteriorhodopsin. Zimmer describes the list as a practical starting point rather than a fixed canon, one that could evolve as the community debates criteria and identifies where standardization is most useful.

GFP offers a clear example of how a model protein can function in practice. The protein fluoresces when it folds correctly, providing a direct and quantitative readout across organisms. Its structure remains consistent across systems, and decades of shared tools - including engineered variants, brightness benchmarks, plasmid libraries and public datasets - allow researchers to compare results across laboratories.

Zimmer also points to the growing role of artificial intelligence in protein research. Fluorescent proteins have become common benchmark cases for machine-learning models because fluorescence provides a clear indicator of whether a designed protein works. As those tools advance, Zimmer argues, shared standards will be essential to ensure results are comparable and reusable.

Nobel laureate Martin Chalfie, whose work helped establish GFP as a foundational research tool, emphasized that coordination matters more than formal labels.

The question Marc Zimmer is addressing is how communities studying similar proteins can benefit by working together."

Nobel laureate Martin Chalfie

Rita Strack, Ph.D., chief editor of Nature Biomedical Engineering, called the proposal overdue and said it could strengthen benchmarking, reproducibility and community-wide resource sharing.

Zimmer's proposal includes convening a cross-disciplinary steering group, defining criteria for model proteins, establishing minimal reporting checklists and curating gold-standard reference datasets. The goal, he said, is to spend less time translating between incompatible practices and more time building on shared work.