aptardi

aptardi predicts sample-specific polyadenylation sites and refines transcript 3'-ends by integrating genome-aligned RNA-Seq, DNA nucleotide sequences, and an initial transcriptome.

Key Features:

• Dual Data Integration: Integrates genome-aligned RNA-Seq data, DNA nucleotide sequences, and an initial transcriptome as inputs.
• Machine Learning Paradigm: Employs a machine learning model to predict expressed polyadenylation sites, achieving approximately twofold higher precision and over threefold improved recall compared to standard transcriptome assemblers.
• Transcript Refinement: Refines transcript 3'-ends based on predicted polyadenylation sites to reflect sample-specific expression.
• Cross-Species Applicability: Trained on the Human Brain Reference RNA commercial standard and maintains high performance across diverse tissues and mammalian species.

Scientific Applications:

• Gene regulation and post-transcriptional modification: Provides precise 3'-end annotations to support investigation of gene regulation and post-transcriptional modifications.
• Alternative polyadenylation and transcript diversity: Enables analysis of alternative polyadenylation and resulting transcript diversity.
• Tissue- and species-specific studies: Facilitates exploration of tissue-specific expression patterns and species-specific regulatory mechanisms.
• Quantitation and differential expression: Produces refined transcript ends usable for quantitation and differential expression analyses.

Methodology:

Inputs comprise DNA nucleotide sequences, genome-aligned RNA-Seq data, and an initial transcriptome; a machine learning model is trained on these inputs to predict expressed polyadenylation sites; predicted sites are used to adjust transcript 3'-ends.