Track trials, findings, and clinical evidence as it lands.
Search papers, labs, and topics across Lattice.
This paper addresses the combinatorial challenge of codon optimization for protein expression by recasting the discrete sequence selection problem into a continuous latent space learned by a pretrained mRNA language model. The proposed Latent-Space Codon Optimization (LSCO) method integrates an uncertainty-aware expression predictor, a Minimum-Free-Energy regularizer, a naturalness prior from a back-translation model, and constrained decoding to optimize for both expression and protein fidelity. LSCO significantly outperforms existing heuristic and deep generative baselines on a wet-lab antibody expression dataset, demonstrating its efficacy in maximizing protein production.
Gradient-based optimization of protein expression is now possible by mapping sequences into an mRNA language model's latent space, bypassing combinatorial search limitations.
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.