Evaluation of order preserving triclustering for 3-dimensional gene expression data and functional interpretation using gene ontology in breast cancer patients

Abstract

Breast cancer accounts for approximately 30% of cancer-related deaths among women globally. Recent advancements in data science have enabled in-depth analysis of various diseases, including breast cancer, through the field of bioinformatics. This study aims to analyze gene expression data from breast cancer patients using the OPTricluster method and to evaluate the biological interpretation of the results through Gene Ontology analysis. To reduce the complexity of breast cancer data, gene filtering techniques are applied. Triclustering, a bioinformatics approach, is particularly effective for processing three-dimensional gene expression data. One such method, Order-Preserving Triclustering (OPTricluster), classifies genes based on similar expression orders across patients over multiple time points. In this research, OPTricluster was employed across various scenarios, utilizing gene filtering simulations with δ parameters in comparison to the TD Score. The optimal scenario was identified with an interquartile range (IQR) < 0.75 and δ = 1.1. The OPTricluster analysis identified a total of 68 triclusters, consisting of 7 constant triclusters, 46 conserved triclusters, and 15 divergent triclusters. Functional analysis revealed that constant and divergent patterns were associated with protein transport within the Biological Process category. Conserved patterns were linked to apoptotic processes. Furthermore, the Cellular Component analysis highlighted cytosol involvement, while the Molecular Function category consistently identified protein binding across all patterns.