Search

The method of estimation of parameters of monolocus diallele model (MDM) of qualitative trait on relative group data is described, these parameters being the number of affected and normal proband relatives of the arbitrary family degree and siblings in different matings types. For the case of single ascertainment, the expressions of corresponding probabilities have been taken as functions of MDM parameters using ITO matrixes and genetic transition matrix. Estimation of parameters was obtained by the maximum likelyhood method. Hardy-Weinberg equilibrium is not always necessary for this method; some weak requirement of stationarity is quite enough.

Based on ITO matrices, a method for parameter estimation of the monolocus diallele model (MDM) of qualitative trait is described, taking account of non-genetic (environmental) factors. The model parameters, probabilities of relatives' affection, constructing the likelihood function and testing hypotheses of the effect of environmental factors on the penetrations of MDM genotypes are outlined. Examples are given, concerning estimation of epilepsy MDM parameters, taking account of two factors-harmfulness of antenatal ontogenesis period and harmfulness which provokes paroxysms.

The initial data for the analysis have resulted in epidemiologic (547 patients) and genetic-epidemiologic (365 patients) study of patients with diagnosis of epilepsy living in five districts of the Khabarovsk Territory. The population frequency of epilepsy was equal to 0.288%. With the use of data on numbers of sick and healthy first-third-degree relatives, and the method of maximum likelihood, the monolocus diallelic model (MDM) parameters of the family and population epilepsy prevalence were estimated. For each of 9 MDM variants two decisions were obtained, depending on the use of population probability of the feature. The analysis of 8 sets of initial data allowed to ascertain the influence of information about the first-third-degree relatives on parameter estimates. The calculation results are presented for one of initial data sets (set A). Three MDM variants were shown to predict the values of the relative affection probability, these being rather close to frequencies observed. The arguments are presented in favour of quasi-dominant variant with following parameters: frequency of mutant allele in a population - 5.28%, homozygote penetrance - 21.5% and that for heterozygote - 2.6%. According to parameter estimates within this model, the probability of offspring disease in the family with the known number of sick and healthy parents was calculated.