Computational Systems Biology Methods and Protocols.7z

(nextflipdebug5) #1
study-specific analyses which depend on the experiment design.
Take microarrays from Affymetrix as examples; SNP genotype
calls and CNVs (for Affymetrix SNP 6.0 only) were determined
by using BRLMM algorithm for Affymetrix GeneChip Human
Mapping 500K Array [56] and Birdseed algorithm for its updated
version of Affymetrix Human SNP Array 6.0 [57], while DGE
algorithm was used for analyzing expression in Affymetrix Gene-
Chips [58] for the first step; Plink software is used in SNP-based
GWAS as for the second step of analyses. For genome-wide associ-
ation analyses, signals with low quality should be filtered as well.
Normally, samples with low SNP call rate (e.g., 95%) are probably
due to the low sample quality and should be excluded. In another
hand, SNP quality control procedures were performed on the basis
of call rate, minor allele frequency, and Hardy–Weinberg equilib-
rium. After the filter steps, the rest SNPs can be used for association
analyses [59]. On each single sites, regression model was applied to
conduct GWAS [60].

3.3 Transcriptome
Profiling in ALL


Due to the limitation of the genomic determinants or contributions
to leukemogenesis as well as the relapse through low-throughput
techniques, intensive effort has been taken to use high-resolution
genomic profiling to identify novel genetic alterations and may be
translated to the clinical practice as new prognostic tools or thera-
peutic targets ultimately [9]. Several scientific centers have per-
formed genomic profiling of B-progenitor ALL and identified
novel genetic alterations associated with high-risk disease
[61–68]. In general, most of these studies used microarray-based
approaches to investigate gene expression profiling, structural
genetic alterations, or individual target genes by profiling leukemic
cells obtained at diagnosis or relapse [9].
For expression microarrays, unlike the usage of them in solid
cancers, it is impossible to get the cancer samples together with
their matched control samples for hematopoietic malignancy.
However, transcriptome screenings are conducted to further sepa-
rate the ALL subtypes or find out the special cases, thus investigat-
ing their clinical relevance, revealing the possible related affected
pathways, and taking into consideration individualized therapy.
Not surprisingly, the well-established subtypes (e.g., hyperdiploidy,
TEL-AML1) can be easily separated by expression signatures of a
series of genes with high sensitivity and specificity [68]. Of the most
importance, a new subtype (i.e., BCR-ABL1-like or Ph-like) can be
separated from the previous uncharacterized groups (B-others)
according to two independent studies [68, 69]. Patients in this
group don’t have unique chromosomal changes but can cluster
with BCR-ABL1 translation positive cases according to the
transcriptome-based hierarchical clustering analyses. Interestingly,
BCR-ABL1-like patients also have a significant lower sensitivity for
chemotherapy response (e.g., mostly resistant toL-asparaginase and

Insights of Acute Lymphoblastic Leukemia with Development of Genomic... 395
Free download pdf