Model-based prediction of diabetes incidence in BB/OK rats.

In the BB/OK rat strain, the incidence of diabetes (in the context of animal diabetes defined as the percentage of animals developing diabetes up to the age of 200 days) is related to the sex distribution within litters and to the phenotype of the parents. On the basis of these observations, an algorithm was developed to predict the incidence of diabetes of an inbred generation as the percentage of diabetic animals of a given inbred generation. Thus, the incidence of diabetes was analysed in 25 inbred generations of our BB/OK colony according to sex ratio and the phenotype of parents, which is thought to be an important genetic factor for diabetes onset in the offspring. On the basis of a single gene model, the Maximum-Likelihood method was used to predict the incidence of diabetes (average and standard deviation). The practicability of this approach was checked by a retrospective calculation of Maximum-Likelihood estimates and their comparison with the diabetes incidence as observed in the individual inbred generation. The average difference between the estimated and observed values was 5.4% which shows the practicability of the proposed algorithm and confirms the dominance of genetic factors in the development of diabetes in this BB rat subline.