Machine-learning from Pseudomonas putida KT2440 transcriptomes reveals its transcriptional regulatory network.

Area of Science:

• Microbiology
• Systems Biology
• Computational Biology

Background:

• Bacterial gene expression is controlled by transcription factors (TFs).
• Understanding gene regulation is key for bacterial physiology and biotechnology.
• Pseudomonas putida KT2440 is a crucial microorganism for bioproduction.

Purpose of the Study:

• To uncover the genome-scale transcriptional regulatory network (TRN) of P. putida KT2440.
• To apply a machine-learning approach for analyzing gene expression data.
• To provide a quantitative framework for understanding P. putida KT2440 transcriptome dynamics.

Main Methods:

• Independent Component Analysis (ICA) applied to 321 gene expression profiles.
• Identification of 84 independently modulated gene groups (iModulons).
• Systematic comparison of iModulons with known TF-gene interactions.

Main Results:

• 84 iModulons were identified, explaining 75.7% of gene expression variance.
• Expanded understanding of 39 TF regulatory functions.
• Detailed transcriptional shifts during growth phase transitions and carbon source utilization.
• Revealed evolutionary strategies for transcriptome reallocation and defined an osmotic stimulon.

Conclusions:

• This study presents the first quantitative, genome-scale TRN for P. putida KT2440.
• The identified iModulons provide a basis for understanding complex transcriptome changes.
• This work facilitates future engineering of P. putida for bioproduction applications.