Prediction of gene expression levels in Saccharomyces cerevisiae based on chromatin accessibility using multiple...

Area of Science:

• Molecular Biology
• Systems Biology
• Bioinformatics

Background:

• Chromatin accessibility is crucial for gene transcription, influenced by regulatory proteins.
• The link between chromatin accessibility and gene expression is vital but not fully understood.
• Predicting gene expression from chromatin data in Saccharomyces cerevisiae is an underexplored area.

Purpose of the Study:

• To develop a predictive model for gene expression using chromatin accessible regions in yeast.
• To identify sequence features within accessible chromatin that correlate with gene expression levels.
• To explore the potential of these features for designing regulatory elements in synthetic biology.

Main Methods:

• Developed Yeast-Gene, a supervised machine learning model.
• Utilized k-mer features from chromatin accessible regions.
• Focused on local DNA sequences (hundreds of base pairs).

Main Results:

• Achieved an Area Under the Curve (AUC) of 0.90 in predicting gene expression.
• Identified AAGAA and CAAGA as highly influential sequence motifs.
• These motifs are potentially linked to mRNA splicing processes.

Conclusions:

• Yeast-Gene effectively predicts gene expression from chromatin accessibility data.
• Identified motifs offer insights into transcriptional regulation and mRNA splicing.
• Findings can guide the rational design of regulatory elements for enhanced gene expression in synthetic biology.