Molecular systems biology and genomics: Elucidating the intersection between environmental agents and human disease

There has been significant research on the use of gene expression data to identify people with diseases (disease biomarkers), predict exposure to chemicals (exposure biomarkers) and to predict effects from exposure to chemical agents (effects biomarkers). Much of the recent work on effects biomarkers has focused on the classification of the genes into ontology groups that can then be used to predict a biological effect. Gene sets can be formed in a number of ways. One approach widely used is to define a gene set based upon the membership of proteins, enzymes and transcription factors in signaling or metabolic pathways already described by many years of research. There are numerous collections of pathways that could be used such as the Kyoto Encyclopedia of Genes and Genomes Pathways (KEGG pathways). One example of such an approach is the Structurally Enhanced Pathway Enrichment Algorithm (SEPEA). We used the SEPEA algorithm to build a linkage model between genes associated with human diseases and the KEGG pathways. Starting with the Genetic Association Database (GAD), we classified human diseases and conditions into 208 broad diseases and disease categories (e.g. liver cancer, epilepsy, type II diabetes) which we will refer to as the human diseases. For each human condition, we pulled from the GAD all of the genetic polymorphisms associated with that condition and used SEPEA to determine which pathways are most likely to be associated with the disease. This created a linkage mapping between human diseases and the KEGG pathways.