Decoding yeast transcriptional regulation via a data-and mechanism-driven distributed large-scale network model

Area of Science:

• Systems Biology
• Computational Biology
• Genomics

Background:

• Transcriptional regulatory relationships govern gene expression and cellular phenotypes.
• Understanding these complex networks is crucial for biological insights.

Purpose of the Study:

• To introduce a novel distributed, large-scale transcriptional regulatory neural network model (DLTRNM).
• To integrate prior knowledge for pre-trained machine learning models and fine-tuning.
• To accurately predict regulatory correlations and identify key transcription factors (TFs).

Main Methods:

• Developed DLTRNM, a novel neural network model.
• Integrated prior knowledge into pre-trained machine learning models.
• Pre-trained DLTRNM on pan-transcriptomic data and fine-tuned with time-series data using Saccharomyces cerevisiae.
• Defined interactions between transcription factors (TFs) and target genes (TGs).

Main Results:

• DLTRNM accurately predicts regulatory correlations.
• The model simplifies complex transcriptional regulatory networks (TRNs).
• Identified potential key TFs and complemented existing subnetworks.
• Demonstrated reduced computational demands and enhanced interpretability.

Conclusions:

• DLTRNM is a powerful tool for studying transcriptional regulation.
• This advancement aids in understanding complex cellular transcriptional regulation within systems biology.
• The model offers a more interpretable and computationally efficient approach.