pGQL

pGQL analyzes gene expression time-course data using probabilistic graphical models to enable robust temporal queries on noisy microarray measurements.

Key Features:

• Probabilistic Time Boxes: Implements time-box query primitives based on a class of linear Hidden Markov Models (HMMs) to represent temporal expression patterns probabilistically.
• Noise-Robust Statistical Modeling: Models amplitude and frequency fluctuations and accommodates unevenly sampled points to increase robustness to noisy measurements.
• Microarray Time-Course Support: Operates on gene expression time courses derived from microarray experiments.
• gPROF Integration: Allows submission of analysis results to gPROF for further processing and interpretation.

Scientific Applications:

• Temporal gene expression analysis: Enables exploration and querying of gene expression dynamics over time.
• Yeast sporulation datasets: Has been applied to yeast sporulation time-course data as a demonstration of real-world use.
• Pattern and regulatory inference: Supports identification of temporal expression patterns and investigation of regulatory mechanisms.

Methodology:

Uses probabilistic graphical models, specifically linear Hidden Markov Models, to define and execute probabilistic time-box queries on time-course expression data.