A zero-inflated Poisson model for insertion tolerance analysis of genes based on Tn-seq data

Area of Science:

• Genomics
• Microbiology
• Statistical Genetics

Background:

• Transposon insertion sequencing (Tn-seq) is a powerful technique for identifying genes involved in bacterial survival.
• Tn-seq data is characterized by high dimensionality and a large proportion of zero counts, posing analytical challenges.
• Existing methods are insufficient for robustly analyzing the excess of zeros in Tn-seq data.

Purpose of the Study:

• To develop a novel statistical method for analyzing high-dimensional Tn-seq data with an excess of zeros.
• To accurately infer gene functions related to bacterial growth and survival using Tn-seq data.
• To provide a robust framework for categorizing gene tolerance states while controlling the false discovery rate.

Main Methods:

• Proposed a zero-inflated Poisson model to handle the excess zeros in Tn-seq data.
• Employed an expectation-maximization (EM) algorithm for maximum likelihood estimation of model parameters.
• Developed a multiple testing procedure using pseudogenes to classify genes into hypo-tolerant, tolerant, and hyper-tolerant states.

Main Results:

• The proposed zero-inflated Poisson model effectively analyzes high-dimensional Tn-seq data with excess zeros.
• The method successfully categorizes genes into different tolerance states, controlling the false discovery rate.
• Applied the method to Campylobacter jejuni Tn-seq data, demonstrating its practical utility.

Conclusions:

• The developed statistical method provides a robust approach for Tn-seq data analysis.
• This method enhances the ability to infer gene functions related to bacterial fitness.
• The R code and user guide are available for broader application in microbial genomics research.