Tumor with this morphology are found within the undifferentiated and poorly differentiated Vemurafenib classes of NB and are seen most often in cases with MYCN amplified tumors. To extend these findings, we isolated RNA from 18 samples of hemizygous abdominal TH-MYCN tumors and performed Affymetrix gene expression array analysis. These tumors which were representative of full grown progressive tumors, weighing 1.5 gr+/20.75 gr, were harvested from 12 females and 6 males 8�C13 week old mice. All eighteen samples clustered tightly on principle component analysis analysis. Human NB data were obtained from three previously published neuroblastoma studies using the Affymetrix U133 v2 chip and combined to give a total of 125 gene expression arrays for comparison. The data was then examined by PCA and collated with meta data describing each tumor. The 125 arrays were adjusted by batch correction method included in Partek Genomics Suite 6.4. Data from Affymetrix human U133 v2 arrays and mouse Affymetrix 430 v2 arrays were processed using the RMA algorithm. This data was then used to examine data quality by PCA for each species separately. The corrections described above removed much of the variability across the human array data from the different studies. Next were removed genes that did not have documented orthologs, between humans and mice, using the Affymetrix��s ortholog mapping document derived from National Center for Biotechnology Information��s Homologene database. The probesets corresponding to 15646 unique human and mouse unigene pairs were retained. Each subset of the original data from mouse and human data sets was then converted to quartiles to further normalize the data and to concentrate on large differences between the species. Once quartiled, the data within each unigene id was averaged and the resulting means were compared across species. The batch BI-D1870 corrected set of human neuroblastomas and the mouse N-MYC tumors unigene pairs were then assigned grades based on the performance of probesets that were designed to measure the same transcript in the same array. The purpose of this grading was to distinguish consistent measurements from variable ones. The quartile range across all observations within a given unigene and species represents technical and biological variability but does not influence grade. However, the range of probeset signals within a unigene does influence the grade. If the range for all probesets for a particular unigene within a species was less than or equal to 1 quartile it was given an A+ grade. A total of 424 A+ unigene pairs from all stages and MYCN status exhibited expression differences of one or more quartiles between human and mouse and were termed differential.