PDEGEM

PDEGEM estimates transcript and isoform abundance from RNA-Seq data using a nonlinear, positional-dependent, energy-guided regression model that accounts for non-uniform read distribution and stacking-energy–related interactions between DNA templates and reads.

Key Features:

• Nonlinear Regression Model: Uses a nonlinear regression framework to estimate transcript abundance from RNA-Seq read data.
• Positional Dependency: Incorporates positional dependency and positional weight to model non-uniform read distribution along transcripts.
• Energy-Guided Expression: Integrates stacking energy terms to represent interaction dynamics between DNA templates and newly synthesized reads.

Scientific Applications:

• RNA-Seq Quantification: Precise estimation of transcript and isoform expression levels from RNA-Seq datasets.
• Gene and Isoform Expression Comparison: Derivation of reliable expression measures with demonstrated higher correlation to alternative assays.
• Method Benchmarking: Comparative evaluation of expression-estimation methods, explicitly compared against mseq.
• Cross-platform and Cross-species Trend Analysis: Analysis of common trends across sequencing platforms and species to investigate DNA binding–related mechanisms.

Methodology:

Adapts the Positional Dependent Nearest Neighborhood (PDNN) framework to RNA-Seq by implementing a nonlinear regression model that incorporates positional dependencies, positional weights and stacking-energy terms to fit read distributions and improve correlation with alternative assays.