Understanding and evaluating ambiguity in single-cell and single-nucleus RNA-sequencing

Area of Science:

• Single-cell and single-nucleus RNA sequencing (scRNA-seq/snRNA-seq) analysis
• Gene expression and splicing dynamics
• Computational biology and bioinformatics

Background:

• Accurate classification of spliced (mature) and unspliced (nascent) RNA reads is crucial for understanding gene regulation in scRNA-seq and snRNA-seq data.
• A recent modification to splicing classification models by Hjörleifsson and Sullivan et al. aims for high conservatism.
• Evaluating the theoretical underpinnings and practical utility of this new classification method is essential.

Conclusions:

• While conservative, the proposed method's handling of ambiguous reads presents practical limitations compared to simultaneous mapping approaches.
• Conservative splicing classification rules can be effectively implemented in existing tools by adjusting indexed reference targets.
• The development of tools like the piscem index facilitates comprehensive analysis of both nascent and mature transcripts.
• Future work could explore probabilistic approaches for inferring splicing status, potentially offering advantages over discrete UMI classification.