Machine learning method for optimizing coding sequences in mammalian cells.
- 1ATUM, Newark, CA 94560, United States.
- 2Department of Genetics, Stanford University, Stanford, CA 94305, United States.
Abstract
Mammalian cell lines are the preferred hosts for producing commercially relevant therapeutic proteins such as antibodies, antibody derivatives, and cytokine fusion proteins. Even though significant investment is made to optimize upstream and downstream processes, the optimal gene design parameters for heterologous recombinant protein expression remain poorly understood. We describe a generic approach to gene optimization in which design features are systematically sampled and iteratively modulated using machine learning (ML). Synthetic gene variants encoding the Dasher fluorescent protein, differing only in synonymous codons, were used to interrogate the gene-sequence preferences of transient antibody-expressing HEK293 cells. Synonymous codon variations influenced expression by more than 2 orders of magnitude. This variation in protein yield was used to build ML models relating gene design features, which were then employed to design further-improved genes. The ML models were shown to be expression system-specific. Messenger RNA levels and ribosome occupancy were highly correlated with protein levels, suggesting that mRNA lifetime is influenced by coding bias. Our results illustrate a novel, generally applicable method to improve gene expression via synonymous re-coding for any protein target or host cell.
Related
- Optimizing DNA origami assembly through selection of scaffold sequences that minimise off-target interactions
- Artificial Avidity Optimization of T-cell Receptors by Integrating Deep Learning Model and Mammalian Cell Surface Display
- Advancing Targeted Nanoparticle mRNA Delivery Using Machine Learning
- From conservation scores to language models: a feature engineering perspective on machine learning for genetic variant interpretation
- A Drug-Referenced Machine-Learning Framework for Prioritizing Korean Herbal Formulations in Hypertension Using Linear SVM and Graph Convolutional Networks
BibTeX
@article{theodorou2026machine,
title = {{Machine learning method for optimizing coding sequences in mammalian cells.}},
author = {Elias Theodorou and Michael Stadler and Claes Gustafsson and Mark Welch},
journal = {Biology methods \& protocols},
year = {2026},
month = oct,
doi = {10.1093/biomethods/bpag057},
url = {https://europepmc.org/article/MED/42889935},
note = {CC BY-NC},
}
RIS
TY - JOUR
TI - Machine learning method for optimizing coding sequences in mammalian cells.
AU - Elias Theodorou
AU - Michael Stadler
AU - Claes Gustafsson
AU - Mark Welch
PY - 2026
DA - 2026/10/05
JO - Biology methods & protocols
DO - 10.1093/biomethods/bpag057
UR - https://europepmc.org/article/MED/42889935
AB - Mammalian cell lines are the preferred hosts for producing commercially relevant therapeutic proteins such as antibodies, antibody derivatives, and cytokine fusion proteins. Even though significant investment is made to optimize upstream and downstream processes, the optimal gene design parameters for heterologous recombinant protein expression remain poorly understood. We describe a generic approach to gene optimization in which design features are systematically sampled and iteratively modulated using machine learning (ML). Synthetic gene variants encoding the Dasher fluorescent protein, differing only in synonymous codons, were used to interrogate the gene-sequence preferences of transient antibody-expressing HEK293 cells. Synonymous codon variations influenced expression by more than 2 orders of magnitude. This variation in protein yield was used to build ML models relating gene design features, which were then employed to design further-improved genes. The ML models were shown to be expression system-specific. Messenger RNA levels and ribosome occupancy were highly correlated with protein levels, suggesting that mRNA lifetime is influenced by coding bias. Our results illustrate a novel, generally applicable method to improve gene expression via synonymous re-coding for any protein target or host cell.
SN - 2396-8923
ER -