Biomedical Engineering Reference
In-Depth Information
Table 1.
Quantitative RT-PCR for the 6 Selected Genes.
Gene Name Probe Set ID Gene Description
23k PRL vs. Luc 16k PRL vs. Luc
PRL
878 s at
Prolactin
|*
|*
IGFBP-5
38650 at
insulin-like growth
factor binding protein 5 2.3**
6.5**
CHOP
39420 at
DNA-damage-inducible
transcript 3
2.2**
4.5**
2.9
Asparagine
36671 at
Asparagines synthetase
5.4**
TB1
37178 at
Hypothetical protein
BC017169
1.1
1.14
DKK1
35977 at
Dickkopf homolog 1
(Xenopus laevis)
0.4**
0.4**
Note.
*: The prolactin gene was not included in quantitative RT-PCR test because it was
conrmed in the Kim et al.'s study.
48
**: Dierential expression by microarray.
: Non-dierential expression by microarray.
integrated bioinformatics tool to extract the relevant biological information
and eectively present the results so investigators can easily convert them
into useful knowledge.
4.1. Quality Control
The use of 2D image plot is to ensure the high quality of oligonucleotide
array data for analysis. The 2D image plot uses percentile methods to group
data, and then applies the 2D image plot to display the grouped data. Fi-
nally, a coverage rate based on an invariant band is computed to quantify
degrees of array comparability. The 2D image plot is limited to pair-wise
comparisons. When the number of arrays increases, this pair-wise compar-
ison strategy may become impractical. However, in practice, we found it
is not a major issue because, most times, the use of one or two reference
arrays is enough for us to screen out incomparable arrays. Alternatively, we
can average all arrays as the reference. However, this may introduce a con-
founding eect between the average and array incomparability, especially
when the bias among arrays is nonlinear.
4.2. Gene Selection
We use the percentile dierence of probe weighted rank to determine the
status of probe expression change. When sample size is small, such as the
Search WWH ::
Custom Search