Hello, today I will be talking about the research from MIT about the, "New AI model could cut the costs of developing protein drugs". MIT researches developed an AI model that learns the genetic patterns of yeast to potentially make the cost cheaper, manufacturing faster, and more efficient.
- This article emphasizes the use of AI to improve how industrial yeast manufactures proteins.
- The AI model focuses on optimizing the codons, which are 3-letter DNA sequences, in order to increase the protein production of cells.
- The AI model uses the genetic data from the yeast Komagataella phaffi, which is a major organism that is used in biopharmaceutical manufacturing.
- 6 proteins were tested and the AI designed codon sequences that outperformed previous commercial codon optimization tools in most cases.
- With early development rapidly improving, the technology could reduce 15-20% of the price tied to the design of drug production.
- Previous alternatives only measured frequency, unlike the AI model, which analyzed codon patterns, placement, and cellular resource balance.
- The model recognized harmful DNA repeats and harmful biochemical amino acid codons.
- The use of this AI model is scalable, although it requires organism-specific training to work well with different species.
Add comment
Comments