Aeon in Motion
What Rama Ranganathan teaches us: statistics before physics.
Most of biology science performs what Thomas Kuhn called consolidation: it catalogs new components and studies new reactions, building on top of an already-staggering picture of complexity. And yet, we still cannot comprehensively explain the behavior of a single protein. What is needed, in Rama Ranganathan’s view, is a functional decoding of complex biological systems; the problem is not a shortage of data but a shortage of laws. His wager is that living systems hide a small set of general design principles, and that finding them requires doing biology in the style of physics. His Center for Physics of Evolving Systems is designed to do exactly that.
This issue of the Project Aeon newsletter is the first (of hopefully many) that cover people and organizations that embody the spirit of scientific freedom and boundary-crossing exploration.
Who he is. Ranganathan is an MD-PhD (Berkeley, then UC San Diego) who joined The University of Chicago in 2017 as the Joseph Regenstein Professor across Biochemistry & Molecular Biology, the Pritzker School of Molecular Engineering, and the College. He leads the Center for Physics of Evolving Systems and directs BioCARS, a national user facility for structural biology located at the Advanced Photon Source at Argonne National Laboratory. For two decades the Center (and its predecessor at the University of Texas) has pursued three deceptively simple questions about biological systems: what they are, how they work, and why they are built the way they are. The through-line is a method he calls "statistics before physics."
Why he's kindred. Project Aeon exists for people who believe radically consequential science lives or dies depending on the models we use to fund and organize it—i.e., the way the machinery of discovery is engineered and how it operates. Ranganathan offers a structural diagnosis along the same vein: biology has been organized by spatial scale (biochemistry, cell biology, organismal biology, ecology) where each stratum owns its own model systems, tools, and rewards. This blocks any meaningful investigation of the fundamental principals that cut across all biological systems. That critique, like others we’ve covered in the newsletter, is an incentive-and-institution argument, not merely a scientific one. It is matched by a research-process credo we find energizing: commit to a fundamental question on a timescale of ten years to a lifetime; treat innovation as "a necessity, not a virtue"; and recognize that the truly risky move is to keep doing what everyone else is already doing.
What he's doing that's noteworthy. Rather than studying one protein at a time, Ranganathan's lab studies ensembles of related proteins and uses statistics to separate what is invariant (the "symmetries") from what is idiosyncratic. That approach surfaces "sectors": sparse, physically connected networks of amino acids that appear to be the fundamental units of a protein's design, allostery (long-range signaling within proteins), and evolvability. The decisive test of any such model is whether you can build with it. His lab has designed synthetic proteins that fold and function like their natural counterparts. More recently, working with partners at NVIDIA, the group built a layered neural-network stack that learns the rules of proteins directly from sequence data; the platform has already spun out two companies (Evozyne and Natural Machines) aimed at therapeutics and sustainability.
His roadmap is informed by Bell Labs: a cross-divisional entity with its own faculty appointment powers and dedicated resources, roughly 15–20 permanent teams, a rotating fellows program, a technology-development and commercialization arm, a visitors' program, and a biennial symposium. The framing is ambitious: just as Carnot, Shannon, and Turing each found the simple laws that unlocked the industrial, communications, and computing ages, biology, he argues, is now positioned for the same kind of transition.
Why it matters. Ranganathan's belief is that the container shapes the discovery: that biology's failure to find general laws is partly an organizational failure and the fix is a different kind of institution. The Center is a living experiment on whether the biology of the 21st century benefits from a new institutional model, and a reminder that creating a new field can be as consequential (or more) as any single result within an old one.
How to engage. This Big Brains podcast and the lab's website, which keeps an annotated bibliography of primary sources, are good places to start. In the realm of how to fund and build new scientific institutions, the Center is one to watch, and to help along.
As always, if this resonates, or raises a question, please reply.
The Idea Garden
Factors associated with scientific creativity
What makes a scientist creative, and can institutions shape it? Drawing on 324 major discoveries across a century of British, French, German, and American biomedicine, Rogers Hollingsworth argues that breakthroughs came from scientists with high "cognitive complexity": a trait he traces to those who internalized multiple cultures or pursued artistic avocations, and who could work across disparate fields as a result. His second finding is structural: weak, decentralized institutional environments out-produced strong, centralized ones, because they let unconventional careers and new disciplines take root where rigid credentialing blocked them. A case that the conditions for discovery are things we can engineer.
Science, AI, and the Fight for the Creative Frontier
If AI can now search literature, analyze data, and generate hypotheses, what is left for the human scientist to do? Davis Garner argues that the enduring edge is analogy: the capacity to borrow a structure from one domain and map it onto another. He makes the case concrete by walking through a worked example, treating a lymph node as a painting gallery to generate a testable immunology hypothesis, then using an AI agent to check whether anyone had studied it before. His conclusion is that AI can fill in an analogy's details and vet its novelty, but it cannot supply the initial creative leap or find the questions that don't yet have data.
Studying Inquiry
Scientists demand evidence in their own work, yet the systems that fund and organize science largely run on habit and assumption. In this piece, Aishwarya Khanduja and Stuart Buck lay out 32 concrete metascience experiments to test those assumptions directly, spanning funding mechanisms, scientific culture, and AI's growing role in fraud detection, peer review, and experimental design. The premise throughout is that there is no single right way to fund science, and that we should treat the ecosystem like a garden to be cultivated and pruned rather than a field to be standardized.
From systems operators to systems architects
As AI takes over pattern-finding and hypothesis generation, Seemay Chou argues the scientist's most valuable role shifts upstream: designing the data systems that set the ceiling on what machines can discover. Her case study is structural biology's "next PDB": moving from static protein snapshots to capturing how proteins move, and building datasets standardized enough to scale as a byproduct of routine work rather than brute-force spending. She calls for funders to start funding coordinated, open, systems-level design.