Clustering Gene Expression Time Series with Coregionalization: Speed propagation of ALS

TL;DR

This paper introduces a novel clustering method for gene expression time series that captures correlations across conditions, revealing insights into gene regulation and disease progression in ALS.

Contribution

The paper develops a new coregionalization-based clustering approach that models correlations in gene expression across different conditions, enhancing understanding of ALS progression.

Findings

Identified significant gene expression profiles related to ALS

Clusters are less likely to be formed by chance, indicating meaningful biological insights

Unveiled shared latent information across multiple conditions

Abstract

Clustering of gene expression time series gives insight into which genes may be coregulated, allowing us to discern the activity of pathways in a given microarray experiment. Of particular interest is how a given group of genes varies with different model conditions or genetic background. Amyotrophic lateral sclerosis (ALS), an irreversible diverse neurodegenerative disorder showed consistent phenotypic differences and the disease progression is heterogeneous with significant variability. This paper demonstrated about finding some significant gene expression profiles and its associated or co-regulated cluster of gene expressions from four groups of data with different genetic background or models conditions. Gene enrichment score analysis and pathway analysis of judicially selected clusters lead toward identifying features underlying the differential speed of disease progression. Gene ontology overrepresentation analysis showed clusters from the proposed method are less likely to be clustered just by chance. In this paper, we develop a new clustering method that allows each cluster to be parameterised according to whether the behaviour of the genes across conditions is correlated or anti-correlated. Our proposed method unveil the potency of latent information shared between multiple model conditions and their replicates during modelling gene expression data.