Protein expression improved by continuous codon optimization method
Predictor-Guided Latent Space Codon Optimization for Maximizing Protein Expression
Artificial IntelligenceMachine Learning
Summary
Making proteins efficiently in the lab requires choosing the right DNA letters, called codons, but this is a tricky problem because there are many ways to spell the same protein. The authors present a new method that turns this tricky problem into one that can be solved with smooth mathematical tools, using a model that understands mRNA sequences. Their method, LSCO, balances making the protein more likely to be made well, keeping its structure stable, and ensuring the protein stays the same. They tested it on antibody production and found it works better than older tricks and newer AI methods.
What this means in practice
- •For biotech manufacturing teams: Optimize DNA sequences to increase therapeutic protein yields in industrial production processes.
- •For synthetic vaccine developers: Design mRNA sequences with improved protein expression for more effective mRNA vaccine candidates.
Authors
Alberto Caron, Tianyu Cui, Dmytro S. Lituiev, Mangal Prakash, Artem Moskalev, Amina Mollaysa, Bo Zhai, Hirsh Nanda, Daniel M. Poole, Zhongyin Liu, Iman Farasat, Robert Davidson, Nikolay V. Manyakov, Tommaso Mansi, Scott Oloff, Rui Liao
Abstract
Codon optimization, the process of selecting synonymous codons to improve mRNA translation efficiency and protein expression, is central to therapeutic protein production and mRNA vaccines, yet it remains a hard problem. The design space is discrete and combinatorially large, precluding gradient-based methods, and existing tools rely on heuristic proxies (e.g., Codon Adaptation Index or GC-content) that poorly capture true expression. We introduce Latent-Space Codon Optimization (LSCO), which recasts this discrete problem as a continuous one by mapping sequences into the latent space of a pretrained mRNA language model, enabling efficient gradient-based search. LSCO combines four components: a data-driven expression objective from an uncertainty-aware predictor, a Minimum-Free-Energy regularizer for structural stability, a naturalness prior from a protein-to-codon back-translation model, and constrained decoding for protein fidelity. On a real-world, wet-lab antibody expression dataset, LSCO outperforms simple frequency-based, as well as modern deep generative baselines in predicted expression, while retaining suitable biophysical properties.