The identification of genetic alterations using array Comparative Genomic Hybridization (CGH) would provide important insights into the mechanisms of tumorogenesis. High resolution array CGH are very expensive and the experiment require a separate sample, therefore, we developed a computational method for predicting gains and losses of genomic DNA segments based on mRNA expression profiles of tumor cell lines, and called this method a Virtual CGH (vCGH) predictor. VCGH is performed through a novel algorithm in which each chromosomal segment is evaluated by the gene transcriptional profiles. The calculation yields a log of odds (LOD) score for each chromosomal segment and this likelihood-based score is used to predict the correlation between mRNA expression patterns and DNA copy number alterations. By aligning all regions of gains and losses from multiple cell lines we can identify minimal common regions of gains and losses which may contain potential oncogenes or tumor suppressors. This method can be used to screen transcriptional profiles of other malignancies for the identification of DNA segmental loss or gain.