How precisely can data from transgenic mouse mutation-detection systems be extrapolated to humans?: lesions from the human factor IX gene.

Transgenic mutation-detection systems have been pioneered in mice, but the approach is applicable to any species in which transgenic animals can be generated. The observed mutations seen in mutation-detection systems are influenced by the underlying pattern of mutation, i.e., the mutational pattern that occurs in wild-type organisms in endogenous segments of DNA that are not under selective pressure. Unfortunately, the biology of most genes and assays markedly skew the underlying pattern of mutation. Herein, we raise multiple issues that must be addressed in order to estimate the underlying pattern of spontaneous mutation from transgenic mouse mutation-detection systems. If these issues can be addressed, the underlying pattern of spontaneous mutation can then be deduced for multiple cell types and for transgenes integrated into different parts of the genome. Even though transgenic methodology cannot be applied directly to humans, it is likely that comparable data on the underlying pattern of spontaneous mutation will be available in humans. Such data are currently available for germline mutations in the factor IX gene. These data are reviewed because of their relevance to two of the multiple issues that must be addressed in transgenic mouse mutation-detection systems: (1) How can the underlying pattern of mutation be deduced from the observed pattern? and (2) How similar are the underlying patterns of mutation in humans and in mice? The analysis of recent germ-line mutation in the factor IX gene yield estimates of the mutation rates per base pair per generation. In brief, the mutation rates vary from 0.037 x 10(-10) for deletions (> 20 bp) to 360 x 10(-10) for transitions at the dinucleotide CpG. If these mutation rates are extrapolated to the entire genome, the aggregate mutation rate is estimated to be 36 x 10(-10). This implies that the diploid genome of each person contains about 21 de novo mutations. In the future, the underlying pattern of spontaneous mutation will be deduced for multiple human genes and these will serve as benchmarks to assess the similarity of the mutational process in humans and in mice.