DENR

DENR quantifies the abundance of precursor (pre-)RNA isoforms from nascent RNA sequencing data to deconvolute isoform-level transcription from 3' end nascent reads such as PRO-seq.

Key Features:

• Annotation-Based Transcript-Level Quantification: Performs annotation-based transcript-level quantification for nascent RNA sequencing data where each read can represent the 3' end of a transcript being synthesized (e.g., PRO-seq).
• Mixture Modeling Approach: Models nascent RNA read counts at each locus as a mixture of user-provided isoforms to quantify both whole genes and individual isoforms.
• Machine-Learning TSS Predictions and Shape Profile Adjustment: Incorporates machine-learning predictions of transcription start sites (TSSs) and adjusts for the characteristic "shape profile" of read counts along transcription units to improve accuracy.
• Validation on Simulated Data: Demonstrates improved performance compared to simple read-count-based methods in simulations.
• Application to PRO-seq Datasets: Applied to PRO-seq data from K562 and CD4+ T cells, revealing widespread transcription of multiple isoforms per gene, frequent dominant isoforms using internal TSSs, and more than 200 genes with different dominant TSSs across these cell types.
• Joint Analysis with StringTie and RNA-seq: Used alongside StringTie on newly generated PRO-seq and RNA-seq for human CD4+ T cells and CD14+ monocytes to show that entropy at the pre-RNA level contributes disproportionately to overall isoform diversity across cell types.

Scientific Applications:

• Pre-RNA Isoform Quantification: Enables quantification of pre-RNA isoforms from nascent RNA sequencing data such as PRO-seq.
• Transcription Start Site Usage: Reveals internal and cell-type-specific TSS usage across datasets including K562 and CD4+ T cells.
• Isoform Diversity and Entropy Analysis: Assesses the contribution of pre-RNA entropy to isoform diversity across cell types, exemplified in CD4+ T cells and CD14+ monocytes.
• Transcription Dynamics and Alternative Splicing Studies: Supports investigation of transcription dynamics, alternative splicing, and cell-type-specific gene expression patterns.

Methodology:

Performs annotation-based transcript-level quantification of nascent RNA reads; models locus read counts as a mixture of user-provided isoforms; incorporates machine-learning TSS predictions; and adjusts for the read-count "shape profile" along transcription units.