TL;DR
Turn quiet afternoons into listening time
- Thousands of audiobooks, podcasts and originals
- Listen on your phone, tablet or Echo — also offline
- Cancel anytime
The NATIVE-ID research project will receive up to $28.6 million to investigate early protein dysfunction linked to neurodegenerative diseases. A Stowers Institute team will receive about $4.1 million over two years to measure aggregation across 50,000 proteins in yeast, generating data intended to help train AI models. The effort will initially focus on frontotemporal lobar degeneration, but whether its predictions will apply in human neurons or lead to treatments remains to be tested.
A new multi-institutional research project will receive up to $28.6 million to study how proteins begin malfunctioning in neurodegenerative diseases and to develop AI predictions of protein dysfunction. The Stowers Institute for Medical Research says its team, led by investigator Randal Halfmann, will receive about $4.1 million over two years to measure aggregation across 50,000 proteins, producing data for a project that will initially focus on frontotemporal lobar degeneration (FTLD).
The effort, called NATIVE-ID, is led by the Innovative Genomics Institute at the University of California, Berkeley and is part of the Advanced Research Projects Agency for Health’s BIOGAMI program. Researchers from UC Berkeley, Brown University, Emory University, Johns Hopkins University, the Parallel Squared Technology Institute and Texas A&M University are also involved. ARPA-H program manager Shannon Greene leads the broader BIOGAMI program, according to the project announcement.
Halfmann’s laboratory will use Distributed Amphifluoric FRET (DAmFRET), a method developed by the team in 2018, to measure protein self-assembly inside individual living yeast cells. The researchers plan to test 50,000 proteins under varied conditions intended to mimic changes that occur in human cells as they age. The team expects to examine more than 1 million samples and generate over 10 billion measurements of aggregation, the announcement says.
The measurements are intended to help address a gap in current AI capabilities. Many protein-prediction tools have advanced in forecasting stable protein structures, but about one-third of proteins are intrinsically disordered and do not maintain a single fixed shape. Their behavior can still be influenced by their amino-acid sequences, but their shifting forms and interactions make them harder to characterize. NATIVE-ID aims to give AI models large-scale experimental evidence about those interactions, rather than relying on structure predictions alone.
From Protein Data to Early Signals
Protein aggregation is associated with several neurodegenerative diseases, including Alzheimer’s, Parkinson’s, ALS and Huntington’s. The project’s premise is that studying changes before clumping is readily detectable could help researchers identify processes linked to disease earlier. That is a research aim, not evidence that the project can currently forecast an individual’s disease or prevent it.
If the data supports reliable predictions, it could help researchers decide which protein interactions merit further study and guide the design of experiments in human cells. Halfmann said better estimates of disease probabilities and onset ages could help people seek preventive or early-stage treatments or join clinical trials. Those are potential future applications: the project announcement does not report a validated prediction tool, a clinical test or a new treatment.
The work also tackles an important translation challenge. Measurements made in yeast can reveal protein behavior, and Halfmann’s earlier studies found that disease-relevant patterns in yeast could inform behavior in human cells. But the new model’s predictions must still be tested in human neurons. The project combines the yeast experiments with partner expertise in generating human neurons from different genetic backgrounds, according to the announcement.
As an affiliate, we earn on qualifying purchases.
Building on Earlier Protein Studies
NATIVE-ID draws on Halfmann’s work on proteins associated with neurodegenerative disease. In 2023, his lab reported experimentally determining the structure of an initiating step in amyloid formation associated with Huntington’s disease. The team has also studied TDP-43, a protein strongly associated with ALS and FTLD. Halfmann described those studies as pilot work: they examined hundreds of protein sequences, while the new effort is planned to examine 50,000.
The initial focus on FTLD reflects its significant genetic and biological features in common with ALS, according to the project description. The stated longer-term goal is to develop approaches that may apply to other diseases involving protein misfolding. That wider ambition does not mean the research has established that one mechanism or prediction approach will work across all such diseases.
Intrinsically disordered proteins (IDPs) are the central subject because they can shift among multiple shapes rather than settling into one stable structure. Some can form harmful aggregates. The project is designed to measure how differences in amino-acid sequence relate to aggregation tendencies, then connect those experimental results to AI modeling and validation in human neurons.
“If we can better predict the probabilities and onset ages of disease, it could allow many more people to seek preventive or early-stage treatments or enroll in clinical trials.”
— Randal Halfmann, Stowers Institute investigator
Prediction Still Needs Human Testing
The project announcement describes planned experiments and expected data output; it does not report that the AI model has been trained, that predictions have been validated, or that the work has identified a usable early-warning signal. It is also unclear when the full dataset or a working model will be available.
Researchers will need to determine whether patterns measured in yeast hold in human neurons and whether they apply across different genetic backgrounds. The announcement does not specify the model’s eventual performance, how it would estimate disease probability or onset age, or whether those predictions could be used in clinical care. Nor does it establish that the planned work will produce preventive treatments.
Experiments and Neuron Validation
Over the funded two-year period, the Stowers team plans to conduct the large-scale yeast experiments and generate measurements of protein aggregation. The wider NATIVE-ID collaboration is intended to combine that dataset with work on human neurons, where researchers can test whether model predictions hold in a more disease-relevant setting.
The next milestones will be the completion and analysis of the measurements, development of the AI model and validation of its predictions in human neurons. The project announcement gives no dates for those milestones beyond the Stowers award’s two-year period. Any claims about clinical use will depend on results from this research and further testing.
Key Questions
What is the NATIVE-ID project?
NATIVE-ID is a multi-institutional research effort led by the Innovative Genomics Institute at UC Berkeley. It aims to collect experimental data about protein aggregation and use it to develop AI predictions about intrinsically disordered proteins.
How much funding has been announced?
The broader effort is set to receive up to $28.6 million. The Stowers Institute team led by Randal Halfmann is slated to receive about $4.1 million over two years.
What will the Stowers team measure?
The team plans to use DAmFRET to study aggregation across 50,000 proteins expressed in yeast cells under varied conditions. It expects to examine more than 1 million samples and produce over 10 billion measurements.
Will the project predict whether a person will develop a neurodegenerative disease?
That has not been established. The project aims to develop and test AI predictions about protein dysfunction; the announcement does not describe a validated clinical prediction tool or a test for individuals.
Which disease will researchers study first?
The collaboration will initially focus on frontotemporal lobar degeneration (FTLD), which shares genetic and biological features with ALS. The broader goal is to explore approaches that could apply to other diseases involving protein misfolding.
Source: rss
Halloween Picks
halloween
As an affiliate, we earn on qualifying purchases.
