Biomedical Engineering Reference
In-Depth Information
Figure 16.10
The tumor necrosis factor (TNF) signature comprising signifi cantly dif-
ferentially expressed genes from two sets of samples from human umbilical vein endo-
thelial cells (HUVECs) treated with TNF vs. nontreated samples was generated. This
signature then was uploaded into tranSMART and used in an enrichment search using
the target enrichment analysis (TEA) algorithm. Here a statistically signifi cant hit is
presented— GSE3365 — comparing Crohn ' s disease patients ' peripheral blood mono-
nuclear cells (PBMCs) with normals. All the genes from the signature which have
corresponding genes in this contrast are shown in the table with the
p
- values and fold
changes calculated from GSE3365 as well as the so-called TEA scores. The total score
is shown on the top.
state-of-the art methodology and strict QC based on the requirements of
Johnson and Johnson informatics scientists.
Finally, we deployed a statistical method to measure the enrichment of a
signature across disease comparisons. The resulting statistically signifi cant hits
then comprise the disease indication hypothesis for the drug candidate. Such
an enrichment analysis is shown in Figure 16.10.
16.7.4
Federation
We ultimately chose data warehousing as a solution for slowly changing or
internal data. For more dynamic content we followed the data federation
approach. Thus a set of external data sources are federated when searching
for genes, drugs, and pathways and the results are rendered in the application.
Primarily, gene-centric information is linked from public sources using Entrez
gene id—direct link to Entrez Gene, Entrez Global search, and Google Scholar;
licensed sources — GeneCards ( http://www.genecards.org ); and internal
sources—a gene index application called Hydra and a pathway integration
application called Pictor. This latter application integrates multiple-pathway
Search WWH ::
Custom Search