Related Experiment Video
Updated: Apr 15, 2026
07:41
Performing Data Mining And Integrative Analysis Of Biomarker in Breast Cancer Using Multiple Publicly Accessible Databases
Published on: May 17, 2019
9.7K
Analyzing large gene expression and methylation data profiles using StatBicRM: statistical biclustering-based rule
Ujjwal Maulik1, Saurav Mallik2, Anirban Mukhopadhyay3
1Department of Computer Science and Engineering, Jadavpur University, Kolkata, West Bengal, India.
Plos One
|April 2, 2015
Summary
We developed StatBicRM, a novel computational framework for identifying gene biomarkers by integrating statistical and biclustering methods. This approach efficiently mines biological data, improving rule generation and classification accuracy for gene expression and methylation analysis.
Area of Science:
- Bioinformatics and computational biology
- Genomics and epigenetics
- Data mining and machine learning
Background:
- Microarray and beadchip technologies are crucial for gene expression and methylation analysis.
- Biclustering enables simultaneous clustering of genes and samples.
- Identifying reliable biomarkers and understanding gene regulation requires advanced analytical methods.
Purpose of the Study:
- To propose StatBicRM, a computational rule mining framework using statistical and biclustering techniques.
- To identify significant genes, extract special rules, and discover potential biomarkers from biological datasets.
- To evaluate the performance and classification accuracy of the proposed method against existing algorithms.
Main Methods:
- A novel statistical strategy for gene filtering based on data distribution (normal or non-normal).
- Data discretization and post-discretization followed by binary inclusion-maximal biclustering.
- Extraction of special rules from maximal frequent closed homogeneous itemsets and subsequent pathway/Gene Ontology analysis.
Main Results:
- StatBicRM outperforms other rule mining algorithms by generating maximal frequent closed homogeneous itemsets, improving efficiency on large datasets.
- Pathway and Gene Ontology analyses identified key genes within the evolved rules.
- Frequency analysis of genes in evolved rules pinpointed potential biomarkers, and classification accuracy was competitive with other rule-based classifiers.
Conclusions:
- StatBicRM provides an efficient and effective framework for biomarker discovery and rule mining in gene expression and methylation data.
- The method demonstrates superior performance in generating biologically relevant rules and classifying data.
- Integrated analysis of gene expression and methylation reveals epigenetic effects on gene expression levels.
