Comparative analysis of RNA-Seq alignment algorithms and the RNA-Seq unified mapper (RUM)

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• High-throughput sequencing requires aligning millions of short reads to a reference genome.
• RNA sequencing (RNA-Seq) alignment is complicated by RNA splicing, necessitating specialized algorithms.
• Existing RNA-Seq aligners claim high accuracy and efficiency, but a comprehensive benchmark was lacking.

Purpose of the Study:

• To develop a robust RNA-Seq simulator for benchmarking alignment algorithms.
• To evaluate the accuracy and robustness of available RNA-Seq alignment tools.
• To introduce and assess the performance of the RNA-Seq Unified Mapper (RUM) pipeline.

Main Methods:

• Developed an RNA-Seq simulator modeling splicing, errors, and other alignment challenges.
• Measured algorithm accuracy and robustness at base and junction levels using the simulator.
• Validated novel transcript feature detection using RT-PCR and Sanger sequencing.
• Developed the RNA-Seq Unified Mapper (RUM) pipeline for performance comparison.

Main Results:

• The developed simulator accurately models RNA alignment impediments.
• RUM performs comparably to the best current RNA-Seq aligners.
• RUM offers an advantageous combination of accuracy, speed, and usability for RNA-Seq data analysis.
• RT-PCR and Sanger sequencing validated RUM's ability to detect novel exons and alternative splicing.

Conclusions:

• The RNA-Seq simulator provides a reliable platform for algorithm benchmarking.
• The RUM pipeline is a valuable tool for accurate and efficient RNA-Seq alignment.
• This study addresses the need for rigorous comparison of RNA-Seq alignment algorithms.