In brief
- Johns Hopkins Medicine (19 October 2026) says a team led by neurologist Xiaobo Mao has received an ARPA-H award of up to $52.1 million over four years to study how shape-shifting proteins misfold.
- For clinicians and labs, the DETR project aims to deliver three tools: a CSF test that reads alpha-synuclein shape (Parkinson’s), molecules that bind the harmful form while sparing the healthy one, and an open foundation model of disordered-protein behaviour.
- What’s still unknown: any clinical or model results — this is an award and research plan, starting with Parkinson’s disease and Ewing sarcoma.
ARPA-H’s BIOGAMI programme backs the work. DETR joins 21 investigators across 12 institutions in the US and Europe.
What will the money fund?
The release frames three endpoints: diagnostics, selective binders, and a shared computational model.
| Item | Detail (JHM) |
|---|---|
| Award | Up to $52.1m over four years |
| Funder | ARPA-H (BIOGAMI) |
| PI | Xiaobo Mao, Johns Hopkins |
| Team | 21 investigators, 12 institutions |
| First diseases | Parkinson’s; Ewing sarcoma |
| Phases | Two 24-month technical areas |
What is the scientific bet?
“We can read a protein’s sequence in seconds, but for this class of proteins, the sequence still does not tell us which shape it will take and whether that shape will harm the patient,” Mao said. “If we can predict the shape, we can design a drug that recognizes the harmful form and leaves the healthy one alone.”
Phase one builds the “molecular grammar” and foundational model from high-throughput lab measurements. Phase two targets detection and control of folding — biomarkers and interventions that spare working protein forms.
Who else is involved?
Johns Hopkins is the prime. Partners named include Columbia, Gladstone Institutes, MIT, Michigan, Emory, University of Copenhagen, IRB Barcelona, UT Austin, Arizona State, UT Dallas and the University of Washington with the Seattle Institute for Biomedical and Clinical Research.
Who is affected?
- Parkinson’s and sarcoma research groups watching open models and CSF assays
- Drug hunters needing shape-selective binders for disordered proteins
- Open science / foundation-model labs that may reuse the planned public resources
Our read
This is ARPA-H buying a shape-not-sequence AI stack for intrinsically disordered protein biology — with Parkinson’s as the beachhead and an explicit open-model promise. Until phase outputs appear, treat every diagnostic and binder claim as a milestone target, not a product. The upside for health AI is open resources other labs can audit; the risk is over-promising early detection before the CSF test exists.
What’s still unknown?
- Which model architecture and training data will be released
- Timeline to a usable CSF assay
- Whether binders reach preclinical selectivity
How we sourced this
- Sole primary: Johns Hopkins Medicine news release, 19 October 2026.
- Framed as funding award and plans; no clinical results claimed.
Sources
Related stories
How is the programme structured?
DETR sits inside BIOGAMI under ARPA-H programme manager Shannon Greene. The first technical area establishes sequence, structure and environmental rules that govern disordered regions and trains a foundational model on measurements from high-throughput platforms. The second aims to detect misfolding as a biomarker and to modulate proteins so harmful forms can be targeted while healthy forms continue to work. Knowledge from Parkinson’s and Ewing sarcoma is meant to transfer to other BIOGAMI focus areas such as Alzheimer’s, cancer, heart disease and type 2 diabetes — again as a design goal, not a delivered result.
Follow @tsnmediahub on X for breaking AI news as it happens.
Follow on X
