PRAM: a novel pooling approach for discovering intergenic transcripts from large-scale RNA sequencing experiments.

Area of Science:

• Genomics
• Bioinformatics
• Transcriptomics

Background:

• Publicly available RNA sequencing (RNA-seq) data is crucial for understanding biological mechanisms.
• Joint analysis of large RNA-seq datasets enables novel transcript discovery.
• Existing transcript discovery methods use a two-step approach, potentially limiting discovery power.

Purpose of the Study:

• To develop a novel one-step approach for enhanced transcript discovery from pooled RNA-seq data.
• To improve the accuracy and efficiency of identifying novel transcripts compared to existing methods.
• To apply the new method to identify and validate previously unannotated transcripts.

Main Methods:

• Developed Pooling RNA-seq and Assembling Models (PRAM), a one-step transcript discovery method.
• Performed computational benchmarks comparing PRAM (1-Step) against traditional two-step approaches.
• Applied PRAM to 30 human ENCODE RNA-seq datasets and mouse hematopoietic RNA-seq data.

Main Results:

• PRAM demonstrated superior performance in predicting overall transcript structures and splice junctions.
• The method identified unannotated human transcripts with characteristics of known transcripts.
• New transcripts were discovered and validated in mouse hematopoietic data, showing conserved expression patterns with disease-implicated genes.

Conclusions:

• The one-step PRAM approach significantly enhances novel transcript discovery from large RNA-seq collections.
• PRAM offers improved accuracy and efficiency over two-step methods.
• The identified novel transcripts provide new insights into gene regulation and potential roles in hematopoietic phenotypes.