iReckon: simultaneous isoform discovery and abundance estimation from RNA-seq data.

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• High-throughput RNA sequencing (RNA-seq) offers insights into gene function and disease mechanisms.
• Accurate transcript identification and quantification from noisy RNA-seq data are crucial but challenging.
• Existing computational methods struggle with novel isoforms and technical biases.

Purpose of the Study:

• To introduce iReckon, a novel computational method for simultaneous RNA transcript isoform determination and abundance estimation.
• To develop a probabilistic approach that accounts for biological and technical variations in RNA-seq data.
• To enhance the discovery of novel isoforms and improve the accuracy of transcript abundance quantification.

Main Methods:

• Developed iReckon, a probabilistic method using regularized expectation-maximization.
• Incorporated biological factors (novel isoforms, intron retention, unspliced pre-mRNA) and technical biases (PCR amplification, multimapped reads).
• Applied iReckon to simulated data, real RNA-seq datasets (cancer), and validated findings with QT-PCR.

Main Results:

• iReckon demonstrated superior performance in discovering novel isoforms with fewer false positives compared to existing tools.
• Abundance estimation accuracy surpassed other state-of-the-art methods.
• Successfully reconstructed complex splicing changes in triple-negative breast cancer and MCF7 cell line data.
• QT-PCR validation confirmed all detected isoforms and showed high agreement (r(2) = 0.94) with predicted abundances.

Conclusions:

• iReckon provides a robust and accurate solution for transcript isoform identification and quantification from RNA-seq data.
• The method significantly advances the analysis of transcriptomes, particularly in complex disease contexts.
• iReckon holds promise for revolutionizing our understanding of gene expression and its role in human diseases.