VeloHARMO: a deep learning framework for mRNA coding sequence generation and design via translation velocity harmonization.

VeloHARMO achieves superior translation velocity harmonization across organisms using deep learning and genetic algorithms, outperforming existing methods in prediction accuracy.

• Why it matters: Optimizing heterologous gene expression is limited by differences in translation dynamics that are not captured by codon usage frequency alone, impacting protein yield and activity.
• What they did: VeloHARMO integrates organism-specific deep learning models trained on Ribo-seq data with a genetic algorithm to design synonymous mRNA sequences that preserve native translation velocity patterns.
• The result: Computational evaluations demonstrate that VeloHARMO more effectively harmonizes translation velocity than current approaches, and analyses suggest it may explain differences among alternative coding sequences.