Machine learning identifies candidates for drug repurposing in Alzheimer’s disease
Problem
Alzheimer's disease drug trials are slow, expensive and largely failing, so existing approved drugs need to be screened for reuse.
Why it matters
Alzheimer's affects millions with no effective treatment, and repurposing approved drugs is faster and cheaper than new drug development.
What the AI did
The DRIAD machine learning framework quantifies associations between AD pathology severity (Braak stage) and gene lists, then ranked 80 FDA-approved drugs as repurposing candidates.
Results
Validated on known AD-associated gene lists; produced ranked list of repurposing candidates from 80 FDA-approved drugs.
Impact
ML prioritized approved drugs for Alzheimer's repurposing, but no clinical efficacy is demonstrated in the abstract.
Record
| Venue | Nature Communications |
|---|---|
| Published | 2021 |
| Work type | article |
| Authors | Steve Rodriguez, Clemens B. Hug, Petar V. Todorov, Nienke Moret, Sarah A. Boswell, Kyle E. Evans, George Zhou, Nathan T. Johnson et al. |
| Citations | 293 (OpenAlex, at import time) |
| References | 95 |
| Model or method | DRIAD |
| Organization | not stated in source |
| Primary source | https://doi.org/10.1038/s41467-021-21330-0 |
| Open access | gold |
| Keywords | drug repurposing, Alzheimer's disease, machine learning, DRIAD, Braak stage, FDA-approved drugs |
Importance rating
significant — thousands to millions affected
AD trials have consumed large time/resources with largely negative results; repurposing 80 FDA-approved drugs offers faster, cheaper path, but abstract states no affected-population scale or clinical efficacy.