Biomedical Engineering Reference
In-Depth Information
Method
1 Allow scanner lasers to warm up for 5 min before starting.
2 Scan the microarray at 10
μ
m scanning resolution according to the manufacturer's
protocol.
3 Store scans from both channels as separate TIFF images
4 Perform automated spot finding, using the information from the GAL-file to position the
array grid over each image.
5 Perform automated spot exclusion.
v
6 Perform automated linking of the spot ratios to the genomic positions of the
corresponding oligonucleotides (using the information from the position file).
7 Perform global mode normalization.
w
8 Draw the genomic profile (automated in BF): order normalized ratios by chromosomal
mapping and display in a graph.
x
Notes
v
We suggest excluding spots that have a 'confidence value' lower than 0.1, or a 'quality flag'
lower than 1, which will further diminish outliers. These confidence values are calculated in a
proprietary manner by the BF feature extraction software.
w
Avoid block normalization (normalization per printed block of spots on the array slide, which
can be performed by either median, or intensity-dependent lowess), because this may compress
the profile. Mode normalization is used to set the 'normal' level and is preferred over mean or
median normalization, as it is more accurate since it ignores the ratios generated by gains,
amplifications and deletions. Block normalization is sometimes used to suppress noise, although
its suitability may depend on the type of sample analyzed; that is, for samples showing few
aberrations, block normalization may help to suppress noise, but is not recommended for
samples with multiple chromosomal aberrations (e.g. tumor samples).
x
For more sophisticated analysis procedures and to 'call' the actual gains, losses and ampli-
fications, we recommend the use of more dedicated software, such as the freeware CGH
call [14].
1.2.2 SNP aCGH
The recently developed high-density SNP microarrays were originally developed for
high-throughput genotyping for linkage analysis and association studies. These arrays
have additionally proven useful to measure both genomic copy-number variations and
loss of heterozygosity (LOH); that is, SNP aCGH. The ability of SNP aCGH, unlike
conventional CGH, to detect copy-neutral genetic anomalies offers the benefit of detecting
copy-neutral LOH [15]. Moreover, the combination of copy number abnormalities and
LOH status with the parental origin of the aberrant allele can possibly be associated with
the predisposition to hereditary cancer. Successful use of SNP aCGH has been reported
for several cancers, such as breast, colorectal and lung cancer [16-19]. While the current
