Mining and Design of Ribosome binding site in Paracoccus denitrificans

ABSTRACT: Paracoccus denitrificans is an important heterotrophic nitrifying–aerobic denitrifying bacterium that has attracted considerable attention due to its unique applications in wastewater treatment and bioremediation. However, systematic studies on the ribosome binding sites (RBS) of this bacterium remain insufficient, limiting the precise optimization of gene expression regulation and the efficient implementation of metabolic engineering. In this study, we first utilized advanced bioinformatics tools to screen and identify a large number of candidate RBS sequences from the P. denitrificans genome, and constructed a functional library containing 1,335 RBS sequences by integrating transcriptomic and proteomic data. Based on this library, we developed an RBS strength prediction platform built on a Transformer deep learning architecture, achieving a high correlation with experimental results (R² = 0.88). Furthermore, by integrating a WGAN-GP generative adversarial network with a filter model, we successfully designed novel RBS sequences with predetermined target strengths, with the Pearson correlation coefficient between the predicted values and experimental measurements reaching as high as 0.86. This study not only establishes a comprehensive RBS functional library for a non-model bacterium but also develops an efficient RBS prediction and design platform using deep learning and generative models. The platform provides a novel technical approach for the precise regulation of gene expression and the optimization of metabolic pathways, laying the foundation for the application of P. denitrificans in environmental management and biomanufacturing, and serving as a valuable practical reference for gene regulation research in non-model microorganisms.

ORGANISM(S): Paracoccus Denitrificans

SUBMITTER: Yu Deng  

PROVIDER: PXD071494 | iProX | Tue Dec 02 00:00:00 GMT 2025

REPOSITORIES: iProX