AI Model May Slash Protein Drug Development Costs

MIT engineers use large language model to optimize protein production in industrial yeast, potentially reducing costs of new biologic drugs.

Published on Feb. 18, 2026

Got story updates? Submit your updates here. ›

MIT chemical engineers have developed an AI model that can optimize the genetic code used by the industrial yeast Komagataella phaffii to produce proteins, including human growth hormone and a cancer-treating monoclonal antibody. The model outperformed commercially available codon optimization tools, potentially reducing the time and cost of developing new protein-based drugs and biologics.

Why it matters

Protein-based drugs and biologics are a rapidly growing segment of the pharmaceutical industry, but the development process is complex and expensive. This new AI-powered approach to optimizing protein production in yeast could help streamline the process, making it faster and more cost-effective to bring new biologic medicines to market.

The details

The MIT team used a large language model to analyze the patterns of codon usage in K. phaffii, the industrial yeast commonly used to manufacture protein drugs and vaccines. The model was able to predict which codon sequences would work best for producing six different target proteins, including human growth hormone and a cancer drug monoclonal antibody. When tested experimentally, the AI-optimized sequences outperformed those generated by commercially available codon optimization tools for five out of the six proteins.

  • The new study was published in the Proceedings of the National Academy of Sciences in February 2026.

The players

J. Christopher Love

The Raymond A. and Helen E. St. Laurent Professor of Chemical Engineering at MIT, a member of the Koch Institute for Integrative Cancer Research, and faculty co-director of the MIT Initiative for New Manufacturing (MIT INM).

Harini Narayanan

The lead author of the study and a former postdoc at MIT.

Komagataella phaffii

An industrial yeast used to manufacture vaccines, biopharmaceuticals, and other useful compounds.

Got photos? Submit your photos here. ›

What they’re saying

“Having predictive tools that consistently work well is really important to help shorten the time from having an idea to getting it into production. Taking away uncertainty ultimately saves time and money.”

— J. Christopher Love, Professor (Proceedings of the National Academy of Sciences)

“We made sure to cover a variety of different philosophies of doing codon optimization and benchmarked them against our approach. We've experimentally compared these approaches and showed that our approach outperforms the others.”

— Harini Narayanan, Lead Author (Proceedings of the National Academy of Sciences)

What’s next

The researchers plan to continue using the new AI model to optimize protein production in K. phaffii and have made the code available for other researchers to use with this and other organisms.

The takeaway

This AI-powered approach to optimizing protein production in industrial yeast could significantly reduce the time and cost of developing new biologic drugs, helping to accelerate the growth of this important pharmaceutical sector.