A contextual activity score (CAS) for inferring ADAR-associated transcriptional activity across RNA-seq, single-cell,...

Area of Science:

• Molecular Biology
• Genomics
• Bioinformatics

Background:

• Adenosine-to-inosine RNA editing by Adenosine Deaminases Acting on RNA (ADARs) is crucial for neural function, immunity, and cancer.
• The Alu Editing Index (AEI) is the standard but limited metric for ADAR activity, unsuitable for single-cell and spatial transcriptomics.
• Existing methods struggle with diverse transcriptomic data types.

Purpose of the Study:

• To develop an alternative framework for inferring ADAR activity from gene expression data.
• To create a method applicable across various transcriptomic technologies, including single-cell and spatial data.
• To enable robust ADAR activity assessment independent of raw sequencing reads.

Main Methods:

• Developed the Contextual Activity Score (CAS) framework using ADAR perturbation transcriptional signatures.
• Generated context-specific signatures for human neurons, mouse neurons, and cancer models.
• Computed CAS from normalized gene expression matrices via regulon-based enrichment analysis.

Main Results:

• CAS demonstrated strong concordance with the Alu Editing Index across datasets.
• CAS proved robust to reduced sequencing depth and varied library protocols.
• CAS successfully applied to single-cell and spatial transcriptomics, enabling independent ADAR2 assessment and capturing biological variations.

Conclusions:

• CAS offers a scalable method for estimating ADAR activity from gene expression data across RNA-seq protocols.
• An open-source R package facilitates broad adoption of CAS.
• CAS expands the study of ADAR activity in transcriptomic modalities challenging for direct editing quantification.