Estimating the Age of Mutant Disease Alleles Based on Linkage Disequilibrium

Abstract

With more and more disease genes being mapped and/or cloned, there is a growing interest in dating the age of underlying mutations. The knowledge of the age of mutation is important to finely map disease genes by linkage disequilibrium mapping. It would also help us understand the origin, evolution, and dispersion of the mutant disease genes. Despite increasing interests in dating disease mutations, the development of appropriate statistical methods is largely fragmentary, and there is a lack of systematic treatment of the topic. We propose two classes of methods for estimating the age of mutant allele at the disease locus based on linked marker data. Our methods can not handle only single-locus marker data, but also multi-locus marker data as well. Moreover, our methods can be used even when the location of the disease locus is unknown, and/or when there are mutations at the marker or disease locus. We show that some previous results are special cases of our methods. We also derive a recursive equation previously obtained by Serre et al. [Hum Genet 1990;84:449-454] and provide an explicit solution to the equation. To illustrate our methods, we applied them to two groups of data sets, one is cystic fibrosis data collected from several European populations, and the other is data on several genetic diseases (diastrophic dysplasia, progressive myoclonus epilepsy, congenital chloride diarrhea, and Batten disease) all collected from the Finnish population. The former data set allows us to trace the origin and dispersion of the most common mutation for cystic fibrosis. The latter provides an opportunity to examine whether all mutations for these diseases have the same age.

Linkage disequilibrium — understanding the evolutionary past and mapping the medical future

“…This approach is straightforward and leads to reasonable estimates of allele age, but it does not take account of the stochastic nature of recombination and genetic drift, and hence exaggerates the accuracy of the resulting estimates. Various statistical methods have been developed that provide more realistic confidence intervals of estimated ages [103][104][105][106] .…”

Age of the intronic GAA triplet repeat expansion mutation in Friedreich ataxia

“…Even in the case of haplotype data from the same study, the age estimate varies considerably from marker to marker. This disappointing feature of the results is probably due to sampling variation in allele frequencies at different loci, and is common to mutation dating by LD analysis (Guo and Xiong 1997). Furthermore, different FRDA haplotypes could have evolved from an ancestral haplotype not only as a consequence of recombination events, but also because of microsatellite instability.…”

“…Recently, using haplotypes from 29 German FRDA families, Zühlke et al (1999) tried to estimate the age of the FRDA mutation by the method of Risch et al (1995). Unfortunately, due to a typographical error in the original paper (pointed out by Guo andXiong 1997 andColombo 2000), their calculations are based on a wrong version of Risch et al's for-…”

Is the Ala12 variant of the PPARG gene an "unthrifty allele"?

“…Allele age based on linkage disequilibrium By using a first order approximation of the likelihood 9 and the single linked C1431T marker locus, a maximum likelihood estimate of the age (in generations) of the Ala12 allele is where h is the recombination rate per generation between both polymorphisms, k 1 is the number of chromosomes carrying both the T1431 and the Ala12 alleles, k 2 is the number of chromosomes having both the C1431 and the Ala12 alleles, p 1 is the frequency of the T1431 allele on Pro12 chromosomes, and p 2 is the frequency of the C1431 allele on Pro12 chromosomes. Because in all the surveyed populations the Ala12 allele is in high linkage disequilibrium with the T1431 variant (D9.0.500), it is assumed that the mutation originating the Ala12 allele first arose in a chromosome carrying the T1431 allele.…”

“…Three different recombination rates were used: h = 0.0008 which corresponds to an average rate of 1 Mb-1 cM, h = 0.0016 which is equal to double the average rate, and h = 0.0004 which corresponds to one half of the average rate. 95% Confidence intervals (95% CI) were estimated based on the composite likelihood given by Guo and Xiong 9 (see equation 12 in their paper).…”