Modeling and analysis of RNA-seq data: a review from a statistical perspective

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Next-generation RNA sequencing (RNA-seq) has revolutionized transcriptomic studies by enabling the analysis of RNA molecules.
• RNA-seq data analysis presents complex statistical and computational challenges at sample, gene, transcript, and exon levels.

Purpose of the Study:

• To review RNA-seq analysis tools from a statistical perspective.
• To highlight practical biological and statistical questions in RNA-seq data analysis.
• To compare statistical models used in RNA-seq analysis to aid method selection.

Main Methods:

• Statistical review of RNA-seq analysis tools.
• Comparison of statistical models based on assumptions and performance.
• Identification of key biological and statistical considerations.

Main Results:

• RNA-seq analysis tools are evaluated across sample, gene, transcript, and exon levels.
• Commonly used statistical models are discussed concerning their assumptions.
• Performance variations of different statistical models under various scenarios are noted.

Conclusions:

• Significant advancements in statistical and computational methods for RNA-seq analysis have occurred.
• Diverse statistical models exist for similar biological questions, with varying performance.
• This review assists users in selecting appropriate methods and informs future method development.