Probabilistic identification of bacterial essential genes via insertion density using TraDIS data with Tn5 libraries.

Area of Science:

• Microbiology
• Genomics
• Bioinformatics

Background:

• Transposon-directed insertion-site sequencing (TraDIS) is crucial for identifying bacterial essential genes.
• Existing methods primarily focus on mariner transposons, leaving Tn5-based data analysis underdeveloped.
• Tn5 transposon data presents challenges due to high insertion density and genomic resolution.

Purpose of the Study:

• To present a novel probabilistic Bayesian approach for classifying bacterial essential genes using Tn5 TraDIS data.
• To address the limitations of existing methods in analyzing Tn5 transposon insertion data.
• To provide a flexible tool for essential gene identification based on user-defined costs.

Main Methods:

• Developed a probabilistic Bayesian framework for essential gene classification.
• Implemented Markov chain Monte Carlo (MCMC) sampling to estimate posterior probabilities of essentiality.
• Utilized Bayesian decision theory for gene selection.
• Assessed performance using simulated data and published datasets from *Escherichia coli*, *Salmonella Typhimurium*, and *Staphylococcus aureus*.
• Compared results against Bio-Tradis, a standard gene classification tool.

Main Results:

• Achieved high classification accuracy with Area Under the Curve (AUC) values ranging from 0.967 to 0.983 across three bacterial datasets.
• Simulated data analysis revealed that insertion number and tolerance in distal regions of essential genes are key factors for accuracy.
• The method allows users to customize essential gene classification based on false discovery and non-discovery costs.

Conclusions:

• The novel Bayesian approach effectively identifies bacterial essential genes from Tn5 TraDIS data.
• The method demonstrates superior performance compared to existing tools like Bio-Tradis.
• An R package implementing the method is available, facilitating its application in bacterial genomics research.