摘要:The 2001 report of the United Nations Scientific Committee on the Effects of Atomic Radiation (UNSCEAR) on `Hereditary effects of radiation' incorporates two important concepts that have emerged from advances in radiation genetics and molecular biology: (a) most radiation-induced mutations are DNA deletions, often encompassing multiple genes; however, because of structural and functional constraints, only a proportion of induced deletions may be compatible with viability and hence recoverable in the progeny and (b) viability-compatible DNA deletions induced in human germ cells are more likely to cause multi-system developmental abnormalities rather than single-gene diseases. The work reported in this paper pursues these concepts further: it examines how mechanistic insights gained from studies of repair of radiation-induced DNA double-strand breaks (DSBs) in mammalian somatic cells and from those on the origin of deletions in human genomic disorders can be extended to germ cells the aim being the development of a framework to predict regions of the human genome that may be susceptible to radiation-induced deletions. A critical analysis of the available information permits the hypothesis that in stem cell spermatogonia, most induced deletions may arise via the non-homologous end joining (NHEJ) mechanism of DSB repair whereas in irradiated oocytes, the main mechanism is likely to be non-allelic homologous recombination (NAHR) between misaligned region-specific segmental duplications that are present in the genome (NAHR is an error-prone form of homologous recombination repair). Should this hypothesis turn out to be valid, then it is possible to build on the structural and functional aspects of genomic knowledge to devise strategies to predict where in the genome deletions may be induced by radiation, their extent and their potential phenotypes.