A new structural class of antibiotic candidates against MRSA
A new structural class of antibiotic candidates against MRSA is graded peer reviewed on whataifound.org, with the AI's role graded search scaffold.
Graph neural networks trained on 39,312 assayed compounds and applied to 12 million molecules surfaced a chemical class active against MRSA in mice, with the substructures driving each prediction made explicit rather than left opaque.
- Verification
- Peer reviewed
- Autonomy
- Search scaffold
- Lab
- MIT / Broad Institute / Harvard
- Model
- Ensembles of graph neural networks with substructure attribution
- Field
- Chemistry
- Date
- 2023-12-20
- Human collaborators
- Felix Wong, Erica Zheng, James Collins
What was found
The team measured antibiotic activity and human-cell cytotoxicity for 39,312 compounds, trained network ensembles on that data, and predicted both properties for over 12 million molecules. Rather than reading off top scores, they extracted the chemical substructures the models were keying on, which let them pick a class rather than isolated hits. Two lead compounds cleared MRSA infection in mouse models, topically and systemically, and appear to kill by collapsing the electrochemical gradient across the bacterial membrane. Published in Nature.
Novelty check
This is the same group's follow-up to the 2020 halicin work and to abaucin (2023); the new element is selecting a structural class via model interpretation instead of screening for individual hits. "New structural class" is a claim about scaffold novelty relative to clinical antibiotics, checked against known antibiotic chemotypes in the paper.
Caveats
These are candidates, not drugs: mouse models only, with no clinical development at the time of this entry. The mechanism — disrupting membrane potential — is the same one halicin uses, so the novelty is structural rather than mechanistic, and membrane-active compounds carry a known mammalian-toxicity risk. The screening library, assays and interpretation were human-run; the model scored and explained.
Sources
How this is graded
whataifound.org grades every entry on two axes: verification (how solid the result is, from a machine-checked proof down to refuted) and autonomy (how much the AI did versus its human collaborators). This finding is peer reviewed and search scaffold. Full definitions are in the methodology.
Cite this entry
whataifound.org (2023). A new structural class of antibiotic candidates against MRSA. whataifound.org: A Registry of AI Scientific and Mathematical Discoveries. https://whataifound.org/finding/2023-12-20-antibiotic-structural-class