Ambiguous genes due to aligners and their impact on RNA-seq data analysis.

Area of Science:

• Genomics
• Bioinformatics
• Molecular Biology

Background:

• Ambiguous genes pose challenges in gene expression estimation using next-generation sequencing (NGS) technologies.
• Accurate identification of these genes is critical as they may be implicated in diseases, potentially leading to misdiagnosis and incorrect treatments.

Purpose of the Study:

• To identify ambiguous genes in RNA sequencing (RNA-Seq) data from the Illumina platform.
• To investigate the causes of ambiguity, focusing on mapping difficulties and the role of pseudogenes.
• To evaluate the impact of ambiguous genes on the predictive power in sample classification.

Main Methods:

• RNA-Seq data analysis using multiple aligners (mappers) to detect genes with variable expression measurements.
• Application of a generalized linear model with factors for mappers and experimental groups to identify ambiguous genes.
• Comparative analysis of classification performance using ambiguous versus non-ambiguous genes as covariates.

Main Results:

• A significant proportion of ambiguous genes were identified as pseudogenes, suggesting issues with alignment due to high sequence similarity with functional genes.
• The study successfully identified ambiguous genes using comparative analysis across different mappers.
• Ambiguous genes generally exhibited lower predictive power in sample classification tasks compared to non-ambiguous genes.

Conclusions:

• Ambiguous genes, particularly pseudogenes, present alignment challenges in RNA-Seq data.
• These genes have limited utility in sample classification, indicating their expression data may be less reliable for diagnostic purposes.
• Understanding and identifying ambiguous genes is essential for accurate disease research and clinical applications.