Biomedical Engineering Reference
In-Depth Information
datasets in the data example, the minimum probe rank percentile in one
group is compared to the maximum in the other group. This measure is an
alternative to mean or median dierence, and is potentially useful in basic
medical research, especially for in vitro studies where the study is often
well-controlled, and the sample sizes are quite small (usually 2-4). For a
larger sample size, the dierence of probe rank percentile could become
median dierence. The probe level threshold could be a pre-specied cuto
based on the percentile dierence or a p-value (e.g., p < 0:05) from the
Wilcoxon Mann-Whitney test (a test for median dierence between two
groups).
In the data analysis of the prolactin study, the use of the probe weighted
rank approach leads us to identify a subset of genes strongly associated with
cell proliferation. Quantitative RT-PCR conrmed most genes. Moreover,
an alternative splicing form of the prolactin gene was identied. By taking
these observations together, the probe weighted rank approach provides an
alternative for analyzing oligonucleotide gene array data.
4.3. Integrated Bioinformatics Tool
The integrated bioinformatics tool can be used to extract relevant biological
functions associated with gene expression changes eciently, and generate a
simplied web-based output for investigators to expedite their research. The
tool has the following unique features: (1) Integration of genomic database.
The database is sucient for researchers to study the association of reg-
ulated genes with biological functions and pathways. Classication based
on GO annotations and KEGG pathway in the database lists all biological
processes, cellular components, molecular functions, and pathways involved
in the regulated genes. Utilizing the corresponding frequency tables to indi-
cate the likelihood of particular biological functional activities, investigators
can easily identify pathways associated with regulated-genes. In addition,
the integrated database includes useful variables, such as gene alias, gene
name, gene description, KEGG pathway, and RIF, to identify genes associ-
ated with keywords of interest; (2) Eective presentation of analysis results.
The results of data analysis are presented in an easily readable format. The
outputs are self-tutorial and easy to operate with very basic knowledge
of using internet web browsers. Starting with a simple main HTML le,
users can easily browse the results from the whole set of regulated-genes
to a single dierentially expressed gene. For all regulated-genes, the tool
will group them into subgroups based on their dierential expressions, GO
Search WWH ::
Custom Search