Comparison of Structures from Frequencies of Genes and Surnames in the Population of Ferrara

The genetic structure based on isonymy and on gene frequencies of 7 enzyme systems was studied in a sample of 1361 individuals residing in the Ferrara Province in the Po delta (northern Italy). The sample was divided into two subsamples. The first, or indigenous, sample ( n = 885) was composed of persons born and residing in the same commune of the province; the second, or migrant, sample ( n = 476) was composed of persons who immigrated from a different commune. The study of the seven polymorphic genetic systems shows that there is no significant difference in gene and genotype frequencies between the two subsamples. On the other hand, the migration indicator derived from isonymy of family names is significantly larger in the migrant group than in the indigenous group. Isonymy techniques permit the detection of recent migration even under equality of gene frequencies. In 1965 Crow and Mange derived the quantitative relationship between isonymy of family names, or surnames, and inbreeding. Since their seminal work an important series of investigations on population structure using isonymy has appeared (Yasuda and Morton 1967; Yasuda and Furusho 1971; Yasuda et al. 1974; Lasker 1968, 1969, 1977, 1983; Lasker and Roberts 1982; Mascie-Taylor and Lasker 1984, 1985; Mascie-Taylor et al. 1985). Gottlieb (1983) and Lasker (1985) have published comprehensive reviews and references on the use of isonymy in the study of population structure. Piazza et al. (1987) proposed a method for the estimation of migration rates that makes use of the distribution of surnames and of the isonymy between them. Barrai et al. (1987) observed that the distributions of surnames in several small and large towns are almost exactly linearized by a log2-log2 transformation, and this observation was considered an indication that most surname distributions fit the Kar1 Department of Evolutionary Biology, University of Ferrara, Via Borsari 46, 44100 Ferrara, Italy. 2USL (Local Health Unit) 31, Ferrara, Italy. 3 Present address: Instituto Venezolano de Investigaciones Cientificas, PO Box 21827, Caracas 1020/A, Venezuela. Please address all correspondence to I. Barrai. Human Biology, April 1993, v. 65, no. 2, pp. 225-235. Copyright © 1993 Wayne State University Press, Detroit, Michigan 48202 This content downloaded from 157.55.39.111 on Wed, 03 Aug 2016 06:14:25 UTC All use subject to http://about.jstor.org/terms 226 / BERETTA ET AL. lin and McGregor (1967) model, which predicts the distribution of neutral alleles in a population of N haploid individuals, each carrying one of k different alleles. Although there are a few studies of kinship in which the distribution of surnames and therefore isonymy is related to gene and genotype frequencies of neutral markers (Azevedo et al. 1969; Morton and Hussels 1970; Morton et al. 1973), the relative power of gene and genotype frequencies versus isonymy in describing different components of the genetic structure of populations is not yet well known. Here, we study the gene frequencies from the phenotypes of seven presumably neutral genetic systems (Barrai, Canella et al. 1991) and the characteristics of the surname distributions in two samples from the Ferrara Province in northern Italy. The first sample is composed of persons born and residing in the same commune of the province (hereafter called the indigenous sample), and the second sample is composed of persons who migrated from another commune (hereafter called the migrant sample). By dividing the samples in this way, we expect a higher inbreeding level in the indigenous sample, which might be indicated by higher homozygosity and isonymy, compared with the migrant sample, in which the genetic structure should mirror that of the general population if we assume random migration to the commune of Ferrara. The joint sample is in HardyWeinberg equilibrium (Beretta et al. 1989), but this is not incompatible with deviations from equilibrium resulting from inbreeding in the indigenous sample; thus our hypothesis can be tested on this material. Materials and Methods Two groups of persons residing in the province of Ferrara in the Po delta area in Italy were classified according to their origin: those born in the commune where they reside (885) and those born elsewhere who immigrated to the commune of residence (476). For all individuals we had available the surname and the phenotype for ACPI, ESD, GLOl, GPT, PGD, PGM1, and PGP. In these systems alleles are codominant, so that genotypes correspond to phenotypes. HardyWeinberg equilibrium was tested by chi-square analysis. For PGD only two classes of genotypes were found; thus it was not possible to test for equilibrium. Independence of genotypic frequencies from sample classification was also tested by chi-square analysis. The surname distributions were studied by fitting a regression to the log2log2 transformation of the number of surnames (S) that were represented k times (Barrai et al. 1987). The goodness of fit of the regression indicates that surnames satisfy the expectations of the neutral theory of evolution under random genetic drift, mutation, and migration (Zei et al. 1983). This content downloaded from 157.55.39.111 on Wed, 03 Aug 2016 06:14:25 UTC All use subject to http://about.jstor.org/terms Genetic Structure of Ferrara / 227 Random isonymy was calculated within samples and between samples. Random isonymy within samples is defined by