Studies on the Clustering Algorithm for Analyzing Gene Expression Data with a Bidirectional Penalty.

Area of Science:

• Bioinformatics
• Computational Biology
• Data Mining

Background:

• High-dimensional gene expression data presents significant clustering challenges.
• Existing methods struggle with determining the optimal number of clusters and handling data noise.

Purpose of the Study:

• To develop a robust clustering method for high-dimensional gene expression data.
• To simultaneously determine the unknown number of clusters and mitigate noise.
• To enhance the accuracy of cluster identification and signal feature extraction.

Main Methods:

• A novel clustering approach based on the k-means algorithm.
• Incorporation of bidirectional penalties to constrain cluster number and centroids.
• Application to benchmark gene expression datasets.

Main Results:

• The proposed algorithm demonstrates superior clustering performance.
• It effectively determines the correct number of clusters and identifies signal features.
• Performance is comparable or superior to existing methods in numerical studies.

Conclusions:

• The developed algorithm is effective for clustering high-dimensional gene expression data with an unknown number of clusters.
• Bidirectional penalties offer a robust mechanism for cluster number determination and noise handling.
• The method shows promise for analyzing complex biological datasets.