PairGP

PairGP models paired longitudinal gene expression time series using non-stationary Gaussian processes to compare condition-specific temporal dynamics.

Key Features:

• Non-Stationary Modeling: Employs a non-stationary Gaussian process framework to capture time-varying patterns in gene expression.
• Paired Longitudinal Design Compatibility: Accounts for paired longitudinal designs by modeling correlations between measurements from the same subjects across conditions or time points.
• Pairing Effect Modeling: Explicitly models the pairing effect to enhance detection of differences in temporal dynamics between conditions.
• Temporal and Condition-Specific Effects: Incorporates temporal dependencies and condition-specific effects into trajectory modeling.

Scientific Applications:

• Genomics and Transcriptomics Longitudinal Studies: Analyzing temporal changes in gene expression across conditions in genomics and transcriptomics research.
• RNA-seq Time Series Analysis: Applied to RNA-seq time series data with multiple conditions to compare temporal expression profiles.
• Simulated Data Validation: Validated on simulated datasets to assess robustness and accuracy in recovering true temporal signals.

Methodology:

Application of non-stationary Gaussian processes to model gene expression trajectories with incorporation of temporal dependencies and condition-specific effects and explicit modeling of the pairing effect.