Marker assisted breeding
Traditional methods of making better rice hybrids have depended upon combining strains with favorable traits. The limitation of this process is that the strains need to have an observable trait (also known as a phenotype) for this to occur. Different versions of genes don't always have an obvious phenotype and many potentially favorable combinations can be missed.
Marker assisted breeding uses recently obtained rice DNA sequences to circumvent this problem. Because the DNA sequence of the gene of interest is now known, breeders can always identify which version is present in a rice strain even if the phenotype is subtle or hidden.
The Rice Genome
Now that the entire rice genome has been sequenced, the question shifts to identifying genes that are involved in increased yield, disease resistance and nutritional value. This problem is made more difficult because very few cereal plants have been sequenced, and therefore, many of the rice genes do not resemble any genes of known function.
Protein Structure Prediction
Genes act through the proteins that they encode. Proteins are molecules that fold into a specific shape, or atomic structure. This structure often gives insight about its function. Experimentally determining protein structure is very laborious — it can take years for a single protein. Fortunately, accurate predictions of protein structures can be made from the DNA of genes using computer simulations. The Computational Biology Research Group at the University of Washington has developed Protinfo, which can produce protein structures at a fraction of the cost and time.
Protinfo & World Community Grid
Protinfo is being used to create three-dimensional models of the tens of thousands of rice proteins. These models are then used to predict the function of each protein and to understand the role of the gene that encodes it. The models, and any analysis resulting from examining them, will be housed at the Bioverse database and webserver, which is a comprehensive framework to relate molecules such as proteins and DNA to an organism's pathways and systems.
Volunteers' computers on World Community Grid will run the Protinfo software to create models of all proteins encoded by the rice genome whose structure can be predicted reliably. These models will be analyzed to choose the best ones. From the resulting structures, prediction tools will determine the function of each protein and the role of the gene that encodes it. Using the power of Protinfo, World Community Grid will initially examine over 10,000 genes, and produce 100,000 models per gene. Generating these one billion models on the 320 CPU cluster at the Computational Biology Research Group would take about 30 years to accomplish. Using World Community Grid, this task will take only about a year.
This knowledge will hopefully lead to the development of improved hybrids of rice strains with higher yield, greater disease and pest resistance, and a full range of bioavailable nutrients. This knowledge can also be extended to other food crops such as wheat and maize.