GRAIL will increasingly provide clues about gene regulation by generating comparative information for closely related genomes such as human and mouse.

The human genome contains information that could be used to prevent birth defects and treat or cure devastating diseases, but it is written in a language that scientists are only beginning to understand. To help decipher the code, Ed Uberbacher and colleagues at Oak Ridge National Laboratory combined cutting-edge computer technology with their knowledge of human biology to develop GRAIL (for Gene Recognition Analysis Internet Link), a "thinking" computer program that imitates the human learning process as it searches for genetic meaning. GRAIL and successor software programs can rapidly identify key instructions in genes from within vast stretches of DNA that appear to be meaningless—a critical contribution to the massive, international effort to record and understand the billions of bits and pieces of DNA that make up the human genome, which contains an estimated 30,000 genes. Like humans, GRAIL learns by observing. To train the program, the scientists developed seven statistical rules differentiating genes from other parts of DNA. As it examines more and more bits of DNA, the program learns when and how well the rules work and adjusts them accordingly.

Scientific Impact: About 1,000 biotechnology companies and laboratories now use GRAIL to track down genes that play a role in human disease. The program has become increasingly productive over time and greatly accelerated the process of locating human genes.

Social Impact: In a recent application, GRAIL located the gene that causes adrenoleukodystrophy, an often-fatal disease of the nervous system that affects young boys. Identification and understanding of such genes could lead to treatments or cures for conditions ranging from sickle-cell anemia to muscular dystrophy and cystic fibrosis.

Reference: E.C. Uberbacher and R.J. Mural, "Locating Protein Coding Regions in Human DNA Sequences Using a Multiple Sensor-Neural Network Approach," Proc. Natl. Acad. Sci. USA, Vol. 88, pp. 11261-11265 (1991).

Edward C. Uberbacher, Ying Xu, and Richard J. Mural, "Discovering and Understanding Genes in Human DNA Sequence Using GRAIL," Comput. Methods Macromol. Sequence Anal., Vol. 266, pp. 259-281, 1996.

Ying Xu and Edward C. Uberbacher, "Automated Gene Identification in Large-Scale Genomic Sequences," Journal of Computational Biology, Vol. 4.3, pp. 325-338, 1997.

URL: http://compbio.ornl.gov

Technical Contact: Dr. Marvin Stodolsky, Life Sciences Division, Office of Biological and Environmental Research, 301-903-4742

Press Contact: Jeff Sherwood, DOE Office of Public Affairs, 202-586-5806

SC-Funding Office: Office of Biological and Environmental Research