The very large cDNA microarray they constructed has more than 42,000 detectors built into two chips using verified cDNA libraries upon which many other researchers currently rely. Analyzing the organ tissue with this tool allowed Khan and his team to identify 18,927 genes that constitute their database. "We found that each organ had a unique expression level profile," said Khan, "and, remarkably, any truly random subset of 1,000 genes could distinguish one organ from another."
Each organ revealed a very distinct profile of active genes, different from all others. However, the gene profiles from different organs that share similar biological functions also showed patterns of expression. For example, though the cerebrum and the cerebellum are two distinct parts of the brain, located apart from each other and doing very different jobs, their gene expression profiles reflected their commonality as part of the nervous system. Similarly, "muscle contraction" genes were found in skeletal muscle, smooth muscle tissue, and the heart -- all organs that share a common way of functioning.
To illustrate the kind of useful data that can emerge from using this tool, Khan's team analyzed 100 samples of the most common pediatric solid tumor cancer, neuroblastoma (NB), which accounts for 7 percent to 10 percent of all childhood cancers. Even though the tumor samples were taken from a variety of patients with different stages of cancer, the database kicked out a list of 19 genes that were consistently overexpressed compared to normal brain tissue.
"All of these genes are involved in one way or another with the kinds of activities associated with the development of cancer -