Computing Prediction and Functional Analysis of Prokaryotic Propionylation

Area of Science:

• Biochemistry
• Proteomics
• Bioinformatics

Background:

• Protein propionylation, a post-translational modification, is crucial for understanding molecular mechanisms.
• However, many lysine-propionylated substrates and their roles in disease remain unknown.
• Existing proteome-scale methods have limitations in identifying these substrates.

Purpose of the Study:

• To identify and systematically analyze candidates for protein propionylation in prokaryotes.
• To develop a robust computational tool for predicting lysine-propionylated sites.
• To explore the biological functions and pathways associated with prokaryotic propionylation.

Main Methods:

• Collated experimental prokaryotic propionylation data from databases and literature.
• Trained a support vector machine model using a three-step feature selection method.
• Developed and validated the PropSeek online tool (http://bioinfo.ncu.edu.cn/PropSeek.aspx).

Main Results:

• PropSeek demonstrated stable and robust performance in predicting potential lysine-propionylated sites.
• Leave-one-out and n-fold cross-validation yielded similar, reliable results.
• Analyses suggested prokaryotic propionylation's involvement in protein synthesis and metabolism.

Conclusions:

• PropSeek is a valuable tool for identifying potential protein propionylation sites.
• This research enhances the understanding of the functional roles of protein propionylation in prokaryotes.
• Further investigation into propionylation's role in protein synthesis and metabolism is warranted.