Accurate long-read transcript discovery and quantification at single-cell, pseudo-bulk and bulk resolution with...

Area of Science:

• Computational Biology
• Genomics
• Transcriptomics

Background:

• Accurate mRNA isoform detection and quantification from nanopore long-read sequencing are challenging due to technical noise, particularly in single-cell applications.
• Existing methods struggle with sensitivity and accuracy in complex transcriptomic data.

Purpose of the Study:

• To introduce Isosceles, a novel computational toolkit designed to improve mRNA isoform detection and quantification from nanopore sequencing data.
• To enhance the fidelity of single-cell transcriptome analysis at the isoform level.

Main Methods:

• Development and application of the Isosceles computational toolkit.
• Validation using synthetic and biologically-derived datasets across single-cell, pseudo-bulk, and bulk resolution levels.
• Case study involving the analysis of neuronal differentiation lineages.

Main Results:

• Isosceles demonstrates superior performance in isoform detection sensitivity and quantification accuracy compared to existing methods.
• The toolkit significantly improves the fidelity of single-cell transcriptome quantification at the isoform level.
• Application of Isosceles revealed coordinated splicing events within and between neuronal differentiation lineages.

Conclusions:

• Isosceles is a robust computational tool for accurate mRNA isoform analysis from nanopore long-read sequencing.
• The toolkit facilitates flexible downstream analyses and is applicable to diverse biological systems.
• Isosceles enables advanced studies of cellular heterogeneity and splicing dynamics in biomedical research.