ABSTRACT
This study aims to test the predictive power of gene expression data derived from NIH's database dbEST, which collects gene expression results from a large number and variety of DNA array experiments. The motivation of this study is to make comparable experimental studies, which are usually performed only for one or a few tissues or organs, with a wide variety of other tissues. Confirmation of a good predictive power of dbEST would put a number of interesting and partially surprising recent findings, solely based on data mining, on a more solid basis than available so far. The expression of nine genes (eIF4E, DDX6, HAT1, USP28, HSP90(beta, PKM2, PLK1, COX2 and OPN) plus two calibration genes in paired normal and cancer colon tissues of eight individual patients was investigated by quantitative RT-PCR and compared with the predictions made by the data-base. GUS and beta-actin reveal only little variation among different patients, making them good internal calibration standards. In normal colon tissue, data mining correctly predicts the expression of all nine genes, which covers two orders of magnitude. In cancer, dbEST is somewhat less precise, but still valuable for the comparison with clinical results.
Subject(s)
Biomarkers, Tumor/analysis , Colonic Neoplasms/diagnosis , Colonic Neoplasms/metabolism , Databases, Protein , Diagnosis, Computer-Assisted/methods , Gene Expression Profiling/methods , Neoplasm Proteins/analysis , Gene Expression Regulation, Neoplastic , Humans , Reproducibility of Results , Sensitivity and SpecificityABSTRACT
The expression patterns of 62 genes interacting with p53 have been investigated in 24 normal and cancerous tissues using NIH's dbEST library. The expression levels of individual genes, such as the TTP53 gene itself, but also other genes, vary up to 33-fold among the 24 different tissues and no consistent pattern can be recognized. However, when expression levels for all 63 genes are summed, these "cumulated levels" are surprisingly constant over the 24 investigated normal tissues. In cancers, the variation is further reduced. Essentially, the cumulated expression levels in cancer are independent of those in normal tissue. We furthermore constructed a linear statistical classifier, i.e., a weighted sum of gene expression levels, which robustly distinguishes normal from cancer tissue independent of the particular kind of tissue. Thus, despite very large differences for individual genes and considerable changes during carcinogenesis, the cumulated expressions have narrowly defined levels.
Subject(s)
Genes, p53 , Neoplasms/genetics , Databases, Genetic , Female , Gene Expression Profiling , Genes, Tumor Suppressor , Humans , Male , Oligonucleotide Array Sequence Analysis , Oncogenes , Pregnancy , Tissue DistributionABSTRACT
Using NIH's public database dbEST for expression of genes and ESTs, genes of the glycolysis pathway have been found to be overexpressed in a set of 24 cancers representing more than 70% of human cancer cases worldwide. Genes can be classified as those that are almost ubiquitously overexpressed, particularly glyceraldehyde-3-phosphate dehydrogenase, enolase 1, and also pyruvate kinase, and those that are overexpressed in less than 50% of the investigated cancers. Cancers can be classified as those with overexpression of the majority of the glycolysis genes, particularly lymph node, prostate, and brain cancer, in which essentially all glycolysis genes are overexpressed, and those with only sporadic overexpression, particularly cancers of the cartilage or bone marrow. This classification may be useful when cancer therapies aimed at the Warburg effect are designed.
