Accurate isoform quantification by joint short- and long-read RNA-sequencing.

Area of Science:

• Genomics
• Molecular Biology
• Bioinformatics

Background:

• Accurate transcript isoform quantification is vital for understanding gene regulation and cellular functions.
• Current RNA sequencing methods, short-read (SR) and long-read (LR), have limitations in depth and resolution for comprehensive transcript analysis.
• A need exists for integrated approaches to overcome individual sequencing platform drawbacks.

Purpose of the Study:

• To develop and validate a novel generative model, MPAQT, for accurate isoform-resolved transcript quantification.
• To leverage MPAQT for analyzing transcriptomic complexity during human embryonic stem cell differentiation.
• To investigate the role of untranslated regions (UTRs) in determining transcript isoform and exon usage.

Main Methods:

• Developed MPAQT, a generative model integrating SR and LR sequencing data.
• Validated MPAQT using extensive simulations and experimental benchmarks.
• Applied MPAQT to an in vitro model of human embryonic stem cell differentiation into cortical neurons.
• Utilized machine learning to model transcript abundances and identify determinants of isoform proportion.

Main Results:

• MPAQT achieved state-of-the-art performance in isoform-resolved transcript quantification.
• Untranslated regions (UTRs) were identified as major determinants of transcript isoform proportion and exon usage.
• Isoform-specific sequence features within UTRs likely mediate these effects via RNA-binding proteins and mRNA stability.

Conclusions:

• MPAQT effectively combines complementary sequencing platforms for enhanced transcriptomic analysis.
• UTRs play a critical role in post-transcriptional regulation, influencing isoform and exon usage.
• Findings advance the understanding of transcriptomic complexity and gene regulation mechanisms.