Centralization errors in comparative genomic hybridization array analysis of pituitary tumor samples

Area of Science:

• Genomics
• Cancer Research
• Bioinformatics

Background:

• Comparative genomic hybridization array (aCGH) is a powerful tool for detecting DNA copy number variations.
• Accurate interpretation of aCGH data depends heavily on effective centralization and normalization techniques.
• Existing normalization methods may not adequately address centralization bias, particularly in samples with numerous genomic alterations.

Purpose of the Study:

• To evaluate the reliability of standard aCGH centralization methods.
• To develop and validate a novel method for correcting centralization bias in aCGH data using fluorescence in situ hybridization (FISH).
• To improve the accuracy of aCGH analysis for genomic profiling of pituitary tumors.

Main Methods:

• Sixty-six pituitary tumors were analyzed using Agilent aCGH+SNP 4x180K microarray.
• aCGH raw log2(ratios) were compared with FISH-based log2(ratios) for a subset of chromosomes.
• A new normalization-centralization process was developed, incorporating FISH-based adjustments and loess regression on non-altered probes.
• Comparison of results between aCGH, CGHnormaliter, and the new FISH-based method.

Main Results:

• Significant discrepancies were observed between raw aCGH and FISH results in 11 out of 66 tumors.
• Nine tumors showed differing results between the CGHnormaliter software and the new FISH-based method.
• Discrepancies were more frequent in tumors with a high number of abnormalities (0%-40% normal probes).
• Five tumors with insufficient normal probes for standard normalization exhibited frequent, uncorrected centralization bias.

Conclusions:

• Standard aCGH centralization and normalization methods can be unreliable, especially in tumors with extensive genomic alterations.
• A FISH-based external control is essential for ensuring the accuracy and reliability of aCGH data interpretation.
• The proposed FISH-based normalization-centralization method offers improved accuracy for aCGH analysis in challenging samples.