EBSeqHMM

EBSeqHMM models gene and isoform expression dynamics in ordered RNA-seq experiments (time-course or spatial) using an auto-regressive hidden Markov model to identify differentially expressed features.

Key Features:

• Auto-Regressive Hidden Markov Model: Implements an auto-regressive hidden Markov model that accommodates dependencies in gene expression across ordered conditions such as time points or spatial positions.
• Empirical Bayes Mixture Modeling: Employs an empirical Bayes mixture modeling framework to enhance statistical inference for identifying differentially expressed genes in ordered data.
• Gene and Isoform Expression Analysis: Identifies genes and isoforms with non-constant expression profiles across ordered conditions and characterizes gene-specific expression paths.
• Simulation and Case-Study Validation: Methodology has been validated through simulations and real-world case studies demonstrating detection of differentially expressed genes and specification of expression trajectories.

Scientific Applications:

• Time-course and spatial RNA-seq analysis: Identify genes that exhibit changes over time or across spatial positions in ordered RNA-seq experiments.
• Characterization of expression dynamics: Specify and characterize the nature and trajectories of expression changes across ordered conditions.
• Isoform-specific expression inference: Infer isoform-level expression changes and patterns relevant to complex gene regulation.

Methodology:

Combines an auto-regressive hidden Markov model with an empirical Bayes mixture modeling framework to model dependencies across ordered conditions and infer differentially expressed genes and isoforms.