Cluster analysis of comparative genomic hybridization (CGH) data using self-organizing maps: application to prostate...

Area of Science:

• Genomics
• Bioinformatics
• Cancer Research

Background:

• Comparative genomic hybridization (CGH) is a key technique for genome-wide analysis of chromosomal imbalances.
• Prostate cancer research often involves analyzing large datasets from multiple individuals.
• Organizing and interpreting complex genomic data requires advanced analytical tools.

Purpose of the Study:

• To apply a self-organizing map (Genecluster) for cluster analysis of CGH data in pT2N0 prostate cancer cases.
• To identify patterns of chromosomal gains and losses associated with prostate cancer.
• To evaluate the prognostic significance of specific chromosomal alterations.

Main Methods:

• Utilized comparative genomic hybridization (CGH) to detect chromosomal imbalances.
• Applied self-organizing maps (Genecluster), a type of artificial neural network, for unsupervised cluster analysis of CGH data.
• Analyzed data from three groups: 40 recent cases, 20 older cases with follow-up, and a combined dataset.

Main Results:

• Genecluster successfully clustered clinically similar pT2N0 prostate cancer cases based solely on genetic information.
• Frequent chromosomal losses were observed on chromosome arms 6q, 8p, and 13q.
• Losses on chromosome arm 8p demonstrated the most significant prognostic importance.

Conclusions:

• Self-organizing maps (Genecluster) are effective tools for analyzing complex CGH data in cancer research.
• Specific chromosomal losses, particularly on 8p, are associated with pT2N0 prostate cancer and may serve as prognostic markers.
• This approach facilitates the identification of genetic factors influencing cancer progression.