ES cells have been used as model systems for differentiation studies since they are totipotent cells derived from the inner mass of developing blastocysts. ES cells grown under appropriate conditions maintain their totipotent capacity and are able to generate all cell lineages after being introduced into host blastocysts. ES embryoid bodies reflect those found in the embryo indicating that this model system provides access to early therapeutic approaches. Undifferentiated and differentiated ES cells offer a multitude of therapeutic uses such as treatment of Parkinson's disease, diabetes and heart failure. Several scientific questions remain to be answered using stem cells for treatment of diseases. One of the caveats of direct transplantation of stem cells is the formation of ES-cell derived tumors. Thus it is desirable to differentiate the stem cells into terminally mature differentiated derivatives before transplanting them into the recipient host. Results in a limited number of animal studies have shown that differentiated stem cells, after implantation in adult animals, do not cause significant tumor formation. Additionally, it is highly desirable to have pure differentiated cell types. For example, various methods for ES cell differentiation exist that result in the generation of hematopoietic and neuronal lineages of cells; however, no method exists to be able to generate either a hematopoietic or a neuronal pure cell type from ES cells. The technology proposed in this grant offers several advantages over what is currently available for stem cell research such as establishment of genetically modified ES cells that allow selection of specific cell types from mixed populations of differentiating ES cells.