The scientific problem at the heart of the project is one of the most stubborn in modern medicine. Protein misfolding and aggregation are associated with Alzheimer’s disease, amyotrophic lateral sclerosis (ALS), some cancers and a range of other chronic and degenerative conditions. Yet many of the proteins involved have long been considered essentially undruggable, because they lack the stable three-dimensional shapes and well-defined binding pockets that most drug development strategies are built to exploit. When a protein does not hold still, designing a molecule to grip it becomes a formidable challenge.

Team SEEDS, whose full name is Sequence-to-disease: Environment-aware Ensembles of Disordered proteins at Scale, is co-led by Priyanka Joshi of Georgetown University and Chuankai (Kai) Zhou of the Buck Institute. The consortium unites expertise in protein science, artificial intelligence, computational modeling, cancer biology and neuroscience, drawing on researchers from nine academic institutions and seven industry partners. Its shared goal is to build a platform capable of mitigating diseases driven by misfolded proteins, moving the field from describing these molecules one at a time to predicting their behavior from their amino acid sequence and their surroundings alike.

The proteins under study are known as intrinsically disordered proteins, or IDPs. Unlike many proteins that fold into relatively fixed structures, IDPs constantly shift among many different conformations. Conditions inside the cell, including acidity, molecular crowding, metals, metabolites and chemical modifications, can influence which forms they adopt and how they behave. That flexibility allows IDPs to carry out important functions in healthy cells, but it also makes them extraordinarily difficult to understand and target when something goes wrong. Current techniques can capture snapshots of these proteins, often only after they have already clumped together, but researchers cannot reliably predict how they will behave as cellular conditions change, or when those changes begin to become harmful.

Artificial intelligence tools such as AlphaFold have transformed protein science by learning from decades of accumulated structural data, but the same wealth of information does not exist for intrinsically disordered proteins, which makes building predictive models far harder. SEEDS aims to help close that gap by developing new experimental and computational ways to study these molecules and by mapping their behavior at unprecedented scale. Crucially, the team plans to pair computational predictions with experimental measurements, testing protein behavior under varying conditions rather than relying on computer models alone.

Zhou framed the effort in explicitly preventive terms. The team aims to uncover the principles that determine when these proteins remain in healthy functional ensembles and when they begin to shift toward disease-associated states, he said, adding that the work could ultimately enable researchers to detect and potentially redirect harmful conformational changes before they lead to irreversible cellular damage. Central to that ambition is understanding how a protein’s environment shapes its behavior, since the same sequence can follow very different trajectories depending on the molecular company it keeps.

Joshi described the challenge with a question that brought the team together: can researchers recognize when a protein is headed in a harmful direction before it arrives, and if so, can they change its course? She compares the task to learning the grammar of a language. Scientists need to examine proteins at scale while also studying some closely enough to understand how their behavior shifts with context. By combining physics-guided artificial intelligence with experiments that test and refine the model’s predictions, the consortium hopes to uncover the molecular grammar connecting protein sequence and environment to behavior. The objective, in her words, is to catch the seed before it grows into disease.

SEEDS operates under ARPA-H’s BIOmolecular Grammar for protein Aggregation Modulation and Intervention (BIOGAMI) program, led by ARPA-H Program Manager Shannon Greene. BIOGAMI concentrates on the earliest stages of protein dysfunction, before misfolding or clumping inflicts lasting cellular damage. Within the consortium, two complementary research teams attack the problem from different directions. At the Buck Institute, SHIFT, short for Structural Characterization of Intrinsically Disordered Regions at Scale, will map the context-dependent shapes that disordered regions can take across a broad range of human proteins, revealing patterns that single-protein studies cannot capture. At Georgetown, INTREPID, the INTegRated Environment-aware Prediction Engine for Intrinsically Disordered Protein Pathology, will examine in depth how sequence and cellular environment affect the range of shapes a protein can adopt. Researchers there will combine high-throughput and high-resolution measurements, computation and artificial intelligence, focusing on tau, implicated in Alzheimer’s disease, and EWS::FLI1, implicated in Ewing sarcoma, studying both from purified systems through human tissue. Data from the two teams will be integrated to train foundation models designed to predict the context-dependent conformational ensembles of intrinsically disordered regions.

The consortium assembles a deep bench of specialists. At Georgetown, Joshi will work alongside Jeffrey Toretsky, whose research centers on Ewing sarcoma and the disordered EWS::FLI1 protein; R. Scott Turner, director of Georgetown’s Memory Disorders Program, who brings longitudinal clinical, imaging and biomarker expertise; Carlo Tornatore, who leads the neurology department at Georgetown and MedStar Health; Rodrigo Maillard, an expert in single-molecule biophysics; Steven Metallo, who studies small-molecule interactions with disordered proteins; Junfeng Ma, a specialist in mass spectrometry-based proteomics; and Mark Burns and Harshul Pandit, whose work includes preclinical models of tau pathology. At the Buck Institute, Zhou will be joined by Tara Tracy, who specializes in translational tauopathy research and therapeutic discovery, and Birgit Schilling, director of the Institute’s Proteomics and Metabolomics Center, with expertise in protein-protein interaction networks and mass spectrometry-based assays. Academic collaborators include principal investigators at the University of Maryland, Baltimore; the University of Maryland, Baltimore County; CIRMMP; Stanford University; the University of California, San Francisco; the University of Denver; and Memorial Sloan Kettering Cancer Center. Industry partners, among them Nautilus Biotechnology, Nanost, Emergente, Neuroene Therapeutics, Post Translational Medicines, Praio and Omphalos Lifesciences, contribute technologies spanning protein mapping, diagnostics, computational modeling and therapeutic development.

Leaders at both institutions emphasized the broader stakes of the initiative. Eric Verdin, president and CEO of the Buck Institute, said the project reflects how the institute approaches diseases of aging by moving beyond treating disease after it presents and working instead to prevent the earliest molecular changes that drive dysfunction, adding that SEEDS has the potential to create an entirely new framework for understanding, and someday intervening in, the chronic diseases of aging. Norman J. Beauchamp Jr., executive vice president for health sciences at Georgetown University, noted that the most difficult health challenges are rarely solved by one discipline or institution working alone, and that many diseases linked to these proteins have resisted the field’s best efforts while taking an enormous toll on patients and families. If the consortium can better predict when these proteins are headed toward disease and find ways to intervene, he suggested, researchers may change what is possible for patients and families. The research is funded in part by ARPA-H, and the views expressed by the participating institutions do not necessarily represent the official policies of the U.S. Government.

Subject of Research: Prediction and intervention in intrinsically disordered protein misfolding linked to neurodegenerative disease and cancer

Article Title: Up to $51.7 million ARPA-H award backs Georgetown University and Buck Institute co-led “super team” to predict and intervene in harmful protein changes before disease begins

Article References: Up to $51.7 million ARPA-H award backs Georgetown University and Buck Institute co-led “super team” to predict and intervene in harmful protein changes before disease begins. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: ARPA-H, protein misfolding, intrinsically disordered proteins, Alzheimer's disease, Ewing sarcoma, artificial intelligence, BIOGAMI, Georgetown University, Buck Institute, tau, EWS::FLI1, drug development