SingaporePublic sector2026 – ongoingAgency for Science, Technology and Research
PLM and WISE for identifying microbial chemical producers
A*STAR narrowed a microbial screening exercise to 4 candidates and validated 2 species
Problem
A microbe's genetic potential did not establish whether it could produce a desired chemical. Incomplete genomic data also complicated screening.
Approach
A*STAR researchers combined a protein language model with WISE, a generative-AI chemical-structure workflow. They matched biological and chemical signals to prioritize laboratory tests. Applied to over 2,000 samples, the framework identified 4 leading neomycin B producer candidates.
Results
Takeaway
AI narrowed the laboratory screening workload; experiments established which candidates produced the chemical.
Categories
- Industry
- Public sector
- Business function
- IT and R&D
- Innovation type
- Process innovation
- AI technology
- Generative AIPredictive analytics and ML
- Adoption
- Built in-house
- Evidence
- Company's own claim
Sources
Source summaries
A*STAR described its combined protein-language and chemical-structure models. It reported screening over 2,000 samples and validating 2 previously uncharacterised species after selecting 4 candidates.
Full details are in the original sources.
