The surgically resected tumor was handled according to guidelines by the International Neuroblastoma Pathology Committee (Shimada H, Umehara S, Monobe Y, Hachitanda Y, Nakagawa A, Goto S, Gerbing RB, Stram DO, Lukens JN, Matthay KK. International neuroblastoma pathology classification for prognostic evaluation of patients with peripheral neuroblastic tumors: a report from the Children's Cancer Group. Cancer 2001;92:2451-61) and tumor cell content was evaluated. The sample was classified as "Neuroblastoma Schwannian stroma poor tumor" containing more than 80% malignant neuroblasts. According to the criteria of the International Neuroblastoma Staging System, the tumor was classified as stage 4 neuroblastoma.
Extracted molecule
genomic DNA
Extraction protocol
Total genomic DNA was extracted from the tumor sample according to Sambrook et al. (Sambrook J, Fritsch EF and Maniatis T: Molecular Cloning. A Laboratory Manual. Second edition. Cold Spring Harbor Laboratory Press, 1989)
Label
Cy5
Label protocol
1.5 ug of test genomic DNAs were digested with AluI/RsaI restriction enzyme mix (Promega, Madison, WI). Fragmented DNA was labelled by direct enzymatic incorporation of fluorescent tags. Genomic DNA was labelled with Cy5-dUTP (Perkin Elmer, Wellesley, MA) using the Random-Primed Bioprime DNA Labeling kit (Invitrogen Life Technologies, Carlsbad, CA). Briefly, 50 ul reaction mix containing dATP, dGTP and dTTP (120 uM each), dUTP (60 uM) and Cy5-dUTP (60 uM) was incubated with Klenow Fragment (40 units) at 37°C for 2 hrs. Unincorporated nucleotides were removed on Microcon YM-30 filters (Millipore, Billerica, MA) according to the manufacturer protocol.
1.5 ug of reference genomic DNAs were digested with AluI/RsaI restriction enzyme mix (Promega, Madison, WI). Fragmented DNA was labelled by direct enzymatic incorporation of fluorescent tags. Genomic DNA was labelled with Cy3-dUTP (Perkin Elmer, Wellesley, MA) using the Random-Primed Bioprime DNA Labeling kit (Invitrogen Life Technologies, Carlsbad, CA). Briefly, 50 ul reaction mix containing dATP, dGTP and dTTP (120 uM each), dUTP (60 uM) and Cy3-dUTP (60 uM) was incubated with Klenow Fragment (40 units) at 37°C for 2 hrs. Unincorporated nucleotides were removed on Microcon YM-30 filters (Millipore, Billerica, MA) according to the manufacturer protocol.
Hybridization protocol
Dye incorporations of post-labelling products were measured by ND-1000 spectrophotometer (NanoDrop Technologies) and the parameters that predicted successful hybridization were a minimum Cy3 incorporation of 0.5 pmol/ul and Cy5 incorporation of 0.3 pmol/ul. Fluorescent-labelled reference and tumour DNAs (1 ug each) were mixed with 50 ug of human Cot-1 DNA (Invitrogen Life Technologies) and control targets (Agilent Technologies) and were hybridized to Human Genome CGH Microarray 4x44K (Agilent Technologies). Slides were hybridized in SureHyb gasket (Agilent Technologies) placed in rotisserie (20 RPM rotation speed) in hybridization oven at 65° C for 24 hrs. After hybridization the slides were washed with Agilent washing solutions (including Stabilization and Drying solutions), according to Agilent Oligonucleotide array-based CGH for genomic DNA analysis protocol (version 5.0).
Scan protocol
Slides were scanned at 10 µm resolution by using the Agilent G2565BA Scanner. Scan Settings: • Scan region was set to Scan Area (61 × 21.6 mm). • Green PMT was set to 100%. • Red PMT was set to 100%.
Description
Microarray images were processed by Feature Extraction software (version 9.5.3.1, Agilent Technologies).
Data processing
Array CGH analysis was performed by using the aCGH Analytics Software (version 3.5.14, Agilent Technologies). A statistical analysis based on hypergeometric Z-scores was used to identify all aberrant regions in each sample. The scoring method has two steps. First, it identifies the total number of probes with log ratios significantly different from zero in a normal sample. These probes are referred to as aberrant probes and the normal sample is referred to as calibration sample. These statistics can be calculated using a sample with no genetic abnormalities or they can be calculated from the test array itself, which is the recommended procedure and the method used in this study. The log ratios from each sample are converted to normalized log ratios. In the normalization step, the expected mean log ratio is subtracted from all log ratios, and these modified log ratios are divided by the estimated variance of the log ratios. In this case, the expected mean is set to 0 and variance is computed as the Derivative Log Ratio Spread (DLRSpread) of the array. Chromosomes X and Y are not included in the estimation of mean and variance as differences between arrays in these probes may offset the statistics. The number of probes for which normalized log ratios are above and below the user specified thresholds (2.0, in this case) are counted and reported as the calibration statistics. In the second step, aberrant regions are identified by comparing the proportion of aberrant probes identified in a given region of the genome in the test sample as compared to that calculated in the first step. If significantly more probes are detected as aberrant in a given genomic region in the test sample then this region gets a higher score and is called as an aberrant region. In this step the over or under abundance of aberrant log ratios is analyzed for each sample using the user specified window size. In this case, a moving window of size 2 Mb is used. For each position of this window, the number of probes with log ratios above and below the user specified cut off of 2.0 is calculated. Then, a z-score is computed that measures the over or under abundance of aberrant positive log ratios in that particular position of 2 Mb window using an equation that reflects a hypergeometric distribution. The score is assigned to the probe in the middle of that window. This score identifies statistically significant groups of probes that appear to deviate from the typical distribution of log ratios around the mean for the given microarray and provides predictive power to make calls of gain/amplification or deletion events. Probes with a log ratio value > 2 were considered as amplified.
