SaaS Clouds Supporting Biology and Medicine - Cloud Computing with e-Science Applications - page 261

TABLE 11.2
List of Mutated Genes Indicative of Each
Breast Cancer Subtype

Luminal A Luminal B Basal-like HER2E

TP53 TP53 TP53 TP53

PIK3CA PIK3CA RB1 PIK3CA

MAP3K1 MAP3K1 BRCA1 PIK3R1

GATA3 PTEN

FOXA1

TABLE 11.3
List of Highly Expressed Genes Indicative of
Each Breast Cancer Subtype

Luminal A/B Basal-like HER2E

ESR1 PIK3CA FGFR4

XBP1 KRAS EGFR

MYB EGFR HER2

RB1 FGFR1 GRB7

FGFR2 GATA3

KIT BCL2

MET ESR1

PDGFRA

differs from other subtypes by having a high level of PIK3CA mutations and
lower frequency of phosphatase and tensin homolog (PTEN) mutations.
As there are significant overlaps in the gene signatures of these cancer sub-
types, to improve the accuracy of analysis, nonmutated genes can be utilized.
The genes in Table 11.3 are known to be expressed in a nonmutated form in
cancer tumors. By utilizing the list of mutated and highly expressed genes, it
is possible to improve the accuracy of diagnosis.
The mutated gene list (see Table 11.2) and highly expressed gene list
(see Table 11.3) were loaded into EXP-PAC. Using these lists, queries were
performed on the breast cancer data set (see Figure 11.11). For each breast
cancer subtype, mutated genes were identified by performing a keyword
search for genes in the mutated gene list and looking for results with high
mutation scores. Highly expressed genes were identified through a keyword
search of gene symbols, this time looking for genes that appeared more than
once. To ensure that displayed genes were present in the tumor data, the
intensity filter was set to return genes with an intensity greater than 0.
Results (see Table 11.4) showed that, out of the genes known to undergo
mutation during breast cancer, only TP53 was expressed in a mutated
form. Mutated genes common to other subtypes were not present, which is
indicative of a basal-like breast cancer tumor. Examining the expression of