deGPS is a powerful tool for detecting differential expression in RNA-sequencing studies

Area of Science:

• Bioinformatics
• Genomics
• Computational Biology

Background:

• Next-generation sequencing (NGS) technologies, including RNA-Seq, offer high-speed, low-cost transcriptome profiling.
• RNA-Seq provides precise transcript level and isoform measurements, surpassing methods like microarrays.
• Accurate detection of differential gene expression is crucial for understanding biological and disease conditions.

Purpose of the Study:

• To develop a robust and powerful tool for detecting differential gene expression from RNA-Seq count data.
• To address limitations of existing methods, such as type I error inflation and false discovery rate control.

Main Methods:

• Proposed deGPS, a novel framework incorporating generalized Poisson distribution for normalization of sequence count data.
• Implemented permutation-based differential expression tests within the deGPS framework.
• Evaluated deGPS using simulated datasets, compcodeR benchmark data, and real RNA-Seq data from Drosophila.

Main Results:

• deGPS demonstrates precise control of type I error and false discovery rate in differential expression detection.
• The tool exhibits robustness even with abnormal high sequence read counts common in RNA-Seq experiments.
• Systematic evaluations confirmed deGPS's superior performance across diverse datasets.

Conclusions:

• deGPS is a powerful and robust tool for RNA-Seq data normalization and differential expression analysis.
• An R package implementing deGPS with parallel computation capabilities is available.
• deGPS shows potential for enhancing data analysis in other high-throughput platforms like ChIP-Seq and MBD-Seq.