NCI Creates Gene Expression Database of
Normal Human Organ Tissue
Researchers at the National Cancer Institute (NCI) Center for Cancer
Research (CCR) today unveiled a publicly available Web site that
provides a detailed catalogue of the genes that are actually expressed
in most of the body's major organs. The database, also discussed in the
March Genome Research, offers a one-of-a-kind tool that all cancer
researchers can use to better define potential drug targets and
anticipate their impact elsewhere in the human biosystem.
"The Normal Organ Database democratizes access to information that many,
until recently, considered esoteric data for geneticists only," says Dr.
Javed Khan, leader of CCR's Pediatric Oncology Branch oncogenomics team
that developed the database. Today gene expression profiles are becoming
widely available and widely used, Dr. Khan continues, in part because
microarray technology now lets researchers run high-throughput assays
for thousands of genes at once. "The challenge now is to isolate
meaningful results for small numbers of specific genes within these
large datasets," he adds. "More intuitively, one needs a true working
definition of 'normal' against which to measure disease. This tool makes
this far easier."
Use of the database
is not limited to
cancer biologists, but is also open to those involved in developing new
drugs for a wide range of diseases such as heart disease and autoimmune
disorders. It may elucidate the pathological processes in these diseases
Each of the 19 organs tested in the Genome Research study revealed a
distinctive expression pattern or genetic fingerprint, even though the
sources of human postmortem tissue samples used in the study were
biologically diverse. By using so many samples (158) from many different
regions of each organ, Dr. Khan's team engineered a generic database
that is statistically sound. "Remarkably," says Dr. Khan, "any truly
random subset of 1,000 genes will distinguish one organ from another."
With this baseline of normal gene expression for a given organ now
accessible on the Web site, researchers should be able to more
effectively analyze tissue samples from their own work. For example, Dr.
Khan's work has largely focused on neuroblastoma (NB). Using the organ
database, the research team detected and identified 19 highly expressed
genes from their own NB samples that seem to be crucial to the biology
of the NB carcinogenic process. This information now can inform
clinicians of potential "druggable" targets.
The CCR team then took that same data one step further, trying to
actually predict outcomes for particular patients based only on gene
expression. "Using a sophisticated computer program that relies on
artificial neural networks, the team analyzed NB expression profiles
from the database and developed a patentable prognostic tool that
improves on the current prognosis standard in the field, developed by
NCI and the Children's Oncology Group," Dr. Khan said.
"I expect this new searchable database to be heavily used by the
scientific community," predicts Dr. Paul Meltzer, Molecular Genetics
section chief of the Cancer Genetics Branch at the National Human Genome
Research Institute. "Because now any user can extract information
relevant to their own scientific interests without having to consult an
expert in microarray data analysis."
The normal organ database used cDNA microarrays to profile genes. For
those researchers whose data was derived from Affymetrix chips or
oligonucleotide arrays, Dr. Khan intends soon to provide transformation
matrices that will make the datasets fully compatible.