Proteostasis
Proteostasis Collapse & Neurodegeneration · Neurodegeneration
Predict and reverse age-driven protein aggregation — the proteostasis collapse behind Alzheimer's and Parkinson's.
📋 The problem
Alzheimer's and Parkinson's track with the aggregation of proteins (Aβ, tau, α-synuclein) as the proteostasis network fails with age. Predicting and reversing aggregation kinetics could prevent or slow neurodegeneration.
🧗 Why it's a grand challenge
Aggregation is governed by coupled nucleation/elongation rates that are hard to measure in vivo; biomarkers lag pathology; clearance interventions have narrow therapeutic windows.
🧮 Governing model
dM/dt = 2 k_+ m(t) P(t); dP/dt = k_n m^{nc} + k_2 m^{n2} M − k_clear·PNucleation–elongation aggregation master equation (Knowles): aggregate mass M(t) from primary/secondary nucleation and elongation rate constants; clearance by proteostasis network.
Current best: Knowles aggregation kinetics + anti-amyloid antibody trials
🧭 Possible approaches
- Master-equation (Knowles) inversion to infer rate constants from biomarker trajectories
- Neural-ODE emulators of aggregation / clearance
- Design of proteostasis-boosting interventions
🎯 Build the benchmark
Predict the fractional aggregate-load reduction over time under a proteostasis-boosting intervention, within ~10%.
Metric: load_reduction — fractional aggregate-load reduction
Datasets to start from: CSF amyloid/tau longitudinal (ADNI-like), In-vitro aggregation kinetics panel, Proteostasis-network perturbation screen
🤖 Build an AI agent to solve it
An agent that fits a patient's aggregation kinetics and proposes clearance-boosting regimens with predicted load reduction.
Once a benchmark exists, an AI4Science agent can iterate solutions against it — every verified solution earns PWM.
This is a frontier framing page — an open problem, not yet benchmarked or verified, unlike PWM's mature computational-imaging benchmarks.