scMAPA: Identification of cell-type-specific alternative polyadenylation in complex tissues

Area of Science:

• Genomics
• Computational Biology
• Molecular Biology

Background:

• Alternative polyadenylation (APA) impacts gene expression by altering 3'-untranslated region (3'-UTR) lengths, influencing cellular processes like proliferation and differentiation.
• Current bioinformatic methods for identifying cell-type-specific APA genes in scRNA-Seq data face limitations, including rigid assumptions on read coverage, restricted applicability to only two cell types, and inadequate control for confounding variables.
• These limitations hinder accurate and robust identification of APA genes across diverse cell populations and complex biological samples.

Purpose of the Study:

• To develop a novel computational method, single-cell Multi-group identification of APA (scMAPA), for accurate identification of cell-type-specific APA genes from scRNA-Seq data.
• To address the limitations of existing methods by developing a tool that does not rely on read coverage shape assumptions and can handle multiple cell types and confounding factors.
• To enhance the understanding of APA gene functions in complex tissues and various cell types.

Main Methods:

• scMAPA integrates a computational change-point algorithm with a statistical model to analyze scRNA-Seq data.
• It transforms 3'-biased scRNA-Seq data to represent full-length 3'-UTR signals, formulating a change-point problem to avoid read coverage shape assumptions.
• The method models APA isoforms considering cell types and undesired sources of variation, enabling robust identification.

Main Results:

• scMAPA demonstrated superior sensitivity, robustness, and stability compared to existing methods on simulated data and human peripheral blood mononuclear cells.
• Application to mouse brain data with multiple cell types identified cell-type-specific APA genes, revealing novel roles in immune cells, neurons, and brain disorders.
• The tool effectively elucidates cell-type-specific APA functions and provides new insights into APA roles in complex tissues.

Conclusions:

• scMAPA is a powerful and versatile tool for identifying cell-type-specific APA genes in scRNA-Seq data.
• It overcomes key limitations of previous methods, offering improved accuracy and broader applicability.
• The findings highlight the significant role of APA in cellular functions and disease, particularly in complex tissues.