Probability distributions and compartment boundaries in the development of Drosophila.

Clone analysis and fate mapping probe several properties of development. Here it is shown that data on fate mapping support a probabilistic model of cell commitment in Drosophila blastoderms. Adult cells have a distribution of possible ancestors, as W. Baker (1978b) inferred from the theory of compartment-boundary development (Garcia-Bellido, Ripoll and Morata 1973). Fate-map data are used here to describe quantitatively the ancestry distributions on the blastoderm fate map. The properties of the distributions are sensitive to, and probes of, developmental events, such as relative time of cellularization and time of commitment. The theory of this analysis shows first how the meaningful interpretation of the stage represented by a fate map depends on the assumptions made in mapping. A general mapping model described below makes it possible to evaluate several interpretations. Interestingly, the data require a 3-dimensional map, and it is argued that this must be due to an effect of the preblastoderm nuclear synctial stage. Second, the theory shows how compartment boundaries affect ancestry distribution and why they have no observable effect on mapping. Third, the variability implied by ancestry variance does not create too much "noise" to make meaningful maps of small areas; rather, oddly enough, it tends to magnify the apparent distances within small areas to make them more resolvable. Empirical results include probabilistic maps of the Drosophila blastoderm. These results argue that time of commitment varies even for cells in the same compartment, demonstrating the need for a more complex model of early development than that proposed in the compartment model. The results also help to evaluate the significance of compartment boundaries in respect to developmental commitment.