Cross-platform array comparative genomic hybridization meta-analysis separates hematopoietic and mesenchymal from...

Area of Science:

• Genomics
• Cancer Research
• Bioinformatics

Background:

• Array comparative genomic hybridization (array CGH) is used to study chromosomal copy number changes in various tumor types.
• Characterizing chromosomal aberrations that distinguish different tumor classes remains a challenge.
• Cross-platform analysis of array CGH data is complicated by differing methodologies.

Purpose of the Study:

• To perform a meta-analysis of array CGH data sets to classify tumor samples across different platforms.
• To develop preprocessing methods for integrating and analyzing diverse array CGH data.
• To investigate the potential of array CGH meta-analysis for cancer subtype classification.

Main Methods:

• Integrated cell line and primary cancer data from three dual-channel array CGH platforms across four institutes.
• Developed and applied noise reduction and data transformation preprocessing methods for cross-platform compatibility.
• Analyzed array CGH profiles from 373 primary tumors using the developed preprocessing pipeline.

Main Results:

• Preprocessing enabled perfect clustering of cell line data by cell line, independent of platform or institute.
• Meta-analysis of primary tumor data showed no platform or institute-specific features, confirming cross-platform validity.
• Tumors clustered by common tissue origin, with distinct groupings for hematopoietic, mesenchymal, and epithelial tumors.

Conclusions:

• Chromosomal aberrations differ significantly between tumors of epithelial origin and those of hematopoietic/mesenchymal origin.
• Meta-analysis of array CGH data is a viable method for identifying these distinct chromosomal aberration patterns.
• This approach holds promise for prospective cancer subtype classification by integrating diverse copy number data.