Daniel Casciano, PhD, director of the FDA's National Center for Toxicological Research (NCTR), Jefferson, Ark., isn't afraid to tackle the large problem of creating a database for a variety of information, including that from metabolic profiling, proteomics, genomics, and microarrays. The NCTR wants the combined information, an approach it calls systems toxicology, to check whether data from model systems such as mice can be used to predict the safety of drug candidates in humans. Casciano says the FDA is eventually likely to require metabolite data with new drug applications; the regulatory agency is already examining such data voluntarily submitted by pharmaceutical companies.

The NCTR offers ArrayTrack, a microarray database and tools to manage and probe microarray data from public sources. The NCTR also plans to attach a module with chemical structures and a text-mining module that can be used for predictions. It has coined the term "predictomics" to describe the field of toxicoinformatics and risk assessment. This entire integrated system will be called the Toxicology Integrated System (TIC).

TIC is in the process of being developed, but Casciano predicts that in less than a year the system will have some genomic, proteomic, metabolomic, and pathology data in it, and it will be made publicly available. NCTR already has about 40 people working on the project. TIC will include multiple hardware servers, an Oracle database, sophisticated microarray equipment, robots to handle microarray slides, and mass spectrometry and NMR instruments.

Casciano says the database should be complete within five years. "I would almost bet my neck that [at that time] we may have tools that one could take to the field." For example, he says that if there were a railroad car spill and people were exposed to a noxious gas, on-the-spot evaluation could be done by taking a urine or blood sample.

Valigra is a freelance writer based in Cambridge, Mass.