Supplementary Material and Methods
microarrays analysis of gene expression profiles.
Human 20K Expression Bioarrrays (GE Healthcare, formerly Amersham
Biosciences, Piscataway, NJ)
containing approximately 20.000 gene probes derived
annotated mRNA sequences, were used to analyze gene expression profiles.
Except where indicated, all reagents for labeling and hybridization were provided in the
expression assay reagent kit. RNA was isolated with TRIreagent (Sigma,
. Louis, MO) and its integrity was assessed using an Agilent 2100 Bioanalyser
(Agilent, Palo Alto, CA). Double stranded cDNA and biotin
labeled cRNA were
generated following manufacturer’s instructions, except that biotin
Applied Science, Pen
zberg, Germany) was used instead of biotin
labeled cRNA was purified on an RNeasy column (Qiagen, Valencia, California),
quantified by UV spectrophotometry, and analyzed
for integrity using an Agilent 2100
Bioanalyser. Then, 10 μg of cRNA were fragmented by heating at 94ºC for 20 minutes
in fragmentation buffer, and subsequently diluted in hybridization buffer and hybridized
Bioarrrays for 20 hours at 37ºC i
n an Innova 40 shaking incubator (New
Brunswick, Edison, NJ) at 300 rpm. After hybridization, microarrays were washed in
0.75x TNT buffer (1x TNT: 0.15 M NaCl, 0.05% Tween
200, 1 M Tris
for 1 hour at 46ºC, incubated with Cy5
streptavidin for 3
0 minutes at room temperature,
washed in 1x TNT four times for 5 minutes each followed by two rinses in 0.1x SSC,
20, and then dried by centrifugation. Slides were scanned in an Axon
GenePix Scanner (Arlington, TX) and analyzed using CodeLink
Software (GE Healthcare).
Background extraction and data normalization.
After scanning, background correction for
Human 20K Expression
expression was carried out using
method available in
package developed by Díez
In this step, those genes
showing L (low, Signal Noise Ratio
1) and G (good, SNR
1) flags are
selected for later normalization and analysis
. Concerning expression data normalization,
Cyclic Loess met
hod was implemented in R using the Bioconductor
function and parameters as described by Wu
Identification of g
s differentially regulated and functional analysis
tial gene expression analysis
Test and non
Whitney U test were performed. Obtained p
values were adjusted by Benjamini
Hochberg method for False Discovery Rate (FDR) correction required in multiple
testing approaches. Thos
e genes exhibiting q
selected as differentially
expressed genes among classes.
The set of differentially expressed genes
was analyzed with the
Expression Analysis Systematic Explorer
implemented in Me
V software development
in order to
identify overrepresented GO terms
Component, Molecular Function, Biological Process)
. Fisher‘s exact test with
Bonferroni correction for multiple testing was
used for comparisons.
ne Set Enrichment Analyses (GSEA, see Subramanian
.) were carried out using
publicly available Biocarta pathways (
. In all cases, wei
enrichment statistic and Student’s T
Test metric for r
anked gene lists were
statistical significance was determined using permutation testing (1000
Gene sets including less than 15 members were excluded from the
enrichment analysis. Following non
with FDR<0.25 were considered significantly enriched among classes (stable vs
Díez D, Álvarez R, Dopazo A.
odelink: An R package for analysis of GE Healthcare Gene Expression Bioarrays.
tics. 2007 Mar 7.
Subramanian A, Tamayo P, Mootha VK, Mukherjee S, Ebert BL, Gillette MA, Paulovich A, Pomeroy
SL, Golub TR, Lander ES, Mesirov JP.
Gene set enrichment analysis: a knowledge
based approach for interpreting genome
Proc Natl Acad Sci U S A. 2005 Oct 25;102(43):15545
Wu W, Dave N, Tseng GC, Richards T, Xing EP, Kaminski N.
Comparison of normalization methods for CodeLink Bioarray data.
. 2005 Dec 28;6:309.