Time-synchronized clustering of gene expression trajectories

Area of Science:

• Genomics
• Computational Biology
• Systems Biology

Background:

• Clustering gene expression time course data is crucial for understanding biological process dynamics.
• Existing methods struggle with varying gene activation rates and temporal patterns.
• Accurate clustering requires methods that handle both time and shape variations.

Purpose of the Study:

• To develop a novel clustering method for gene expression time course data.
• To address the challenge of simultaneous time and shape variations in gene activation patterns.
• To improve the discovery of biologically relevant gene expression clusters.

Main Methods:

• A novel clustering approach combining pairwise curve alignment with a clustering algorithm.
• Implementation of a cluster-specific time synchronization technique.
• Evaluation using simulation data and real-world datasets from yeast and human fibroblasts.

Main Results:

• The developed method demonstrates superior performance in cluster quality compared to standard methods.
• Pairwise curve alignment effectively adjusts for time variations within clusters.
• Discovered clusters in yeast data show high concordance with known biological processes.

Conclusions:

• The novel clustering method accurately identifies gene expression patterns with varying temporal dynamics.
• This approach enhances the understanding of biological process unfolding.
• The method provides a valuable tool for analyzing gene expression time course data.