Natural Fertility in Northeastern Mexico

Archives of Medical Research 31 (2000) 520–525

ORIGINAL ARTICLE

Natural Fertility in Northeastern Mexico: Genetic Structure by Year of Birth and Birthplace Ricardo M. Cerda-Flores*,** and Martha I. Dávila-Rodríguez* *Departamento de Genética de Poblaciones, Centro de Investigación Biomédica del Noreste, Instituto Mexicano del Seguro Social, Monterrey, Nuevo León, Mexico **Human Genetics Center, The University of Texas School of Public Health, Houston, TX, USA Received for publication September 24, 1999; accepted April 14, 2000 (99/173a).

Background. The aims of this population genetics study were 1) to ascertain whether 417 Mexican women with natural fertility (45 years of age, married, not using any family planning methods, residing in the state of Nuevo León) were genetically homogeneous, and 2) to compare the genetic structure of this selected population with the previously reported data of random populations of northeastern Mexico. Methods. A sample of 417 women was interviewed and selected in seven medical units of the Mexican Social Security Institute. They were grouped by their year of birth (1896– 1925 and 1926–1955) and birthplace [persons whose four grandparents were born in the northeastern states (NE) and outside the northeastern states (Not-NE) of Mexico]. Eight genetic marker systems were analyzed. Results. Gene diversity analysis suggests that more than 99.1% of the total gene diversity can be attributed to variation between individuals within the population. Genetic admixture analysis suggests that this selected population, stratified by year of birth and birthplace, have received a predominantly Spanish contribution followed by a lesser Mexican Indian contribution. Conclusions. The genetic structure of this selected population was homogeneous and similar to the random populations of northeastern Mexico. This finding corroborates the utility of this selected population for genetic and epidemiological studies. © 2001 IMSS. Published by Elsevier Science Inc. Key Words: Fertility, Homogeneity, Population genetics.

Introduction Nuevo León, one of the five states of northeastern Mexico (that also include Coahuila, Tamaulipas, San Luis Potosí, and Zacatecas) has an area of 64,555 km2 and in 1990 had a population of 4,492,500 inhabitants (Figure 1). It is a young population; 83.8% is less than 44 years of age (1). Nearly 82% of the state’s population is concentrated in the Monterrey Metropolitan Area (MMA) in the central western section of the Mexican state of Nuevo León (1,2). In 1960, Nuevo León had 1,078,848 inhabitants and nearly 50% migrated to the MMA, principally from the adjoining states of San Luis Potosí and Zacatecas (3–5). Address reprint requests to: Ricardo M. Cerda-Flores, División de Genética, CIBIN, IMSS, Administración de Correos #4, Apdo. Postal #20, 64720 Monterrey, Nuevo León, México. Tel.: (⫹52) (8) 190-4036; Fax: (⫹52) (8) 190-40 35; E-mail: [email protected]

The random populations of northeastern Mexico (grouped by the birthplace and year of birth of the four grandparents) are similar in terms of the contribution of Spanish (78.5%) and Mexican Indian genes (21.5%), and in that more than 96% of total genetic diversity (GST) could be attributed to individual variation within the populations defined by birthplace and/or year of birth. There is no nonrandom association of alleles among the genetic marker systems despite the mestizo origin of the population. These results suggest genetic homogeneity (no genetic variation) of these Mexicans and indicate the utility of this population for genetic and epidemiological studies (6). The aims of this population genetics study were as follows: 1) to study genetic variation in a selected female population residing in the state of Nuevo León; 2) to compute the total and individual contribution of the ancestral populations to this population; 3) to demonstrate whether there is a residual

0188-4409/00 $–see front matter. Copyright © 2001 IMSS. Published by Elsevier Science Inc. PII S0188-4409(00)00 0 9 8 - 9

Cerda-Flores and Dávila-Rodríguez / Archives of Medical Research 31 (2000) 520–525

521

Figure 1. Location of the northeastern (NE) states of Mexico.

effect of population mixture of the nonrandom association of alleles, and 4) to compare the observed distribution of the number of heterozygous loci in this population with data reported in the literature for the northeastern Mexican population.

Materials and Methods Sample description and loci studied. Genetic data from this population were collected as part of a larger investigation of the epidemiology of diabetes disease and fertility. The frequencies of the phenotypes of the blood group systems ABO, Rh, MNSs, Duffy (Fy), Kidd (Jk), Lewis (Le), Kell (K), and Kp were determined in 417 women (⭐45 years of age) residing in the state of Nuevo León. The women were interviewed and selected when receiving medical attention at seven medical units of the Mexican Social Security Institute (IMSS) from 1990–1992. All clinics served patients with general medical problems; none were specialty clinics, such as gynecologic clinics. The medical problems seen at these clinics did not bias the sample as far as the authors are aware. The study was approved by the Ethical Committee of the Centro de Investigación Biomédica del Noreste, IMSS, in Mexico (February 1, 1990). Blood groups were determined using a commercial antisera microplate method (7,8). The study was based on self-reports that included 417 women aged 45–89 years who reported that they had not used contraceptives, were still married to the same spouse, had completed fertility, were Roman Catholic, and had lived for more than 30 years in the state of Nuevo León.

A structured questionnaire was used to obtain data on age (in years), age at marriage, years of education, zone of residence, occupation, birthplace of the woman and her spouse, birthplace of the four grandparents, total number of pregnancies, number of male live births, number of female live births, and number of spontaneous abortions. The following two categories were considered for birthplace: 1) persons whose four grandparents were born in the northeastern Mexican states (NE), and 2) persons whose four grandparents were born outside of the northeastern area (Not-NE). The data were subdivided by year of birth and number of cohorts (assuming a period of 30 years per generation). Accordingly, 14 human cohorts have elapsed since this population was colonized in 1596 (9). Two categories were considered for the year of birth: 1) persons born between 1896 and 1925 (cohort 11), and 2) persons born between 1926 and 1955 (cohort 12). Statistical methods. Phenotypic data on eight blood group markers and all questionnaires were entered into a database for statistical analyses, which were conducted in six parts. First, because no consanguinity was reported among the 417 women, the observed phenotype distribution for each marker was used to calculate allele frequencies, using the maximum likelihood method (10). The homogeneity of allele frequency distributions in the sample of women was tested (by a contingency table analysis) for all markers. Second, the extent of genetic variation between the year of birth

522

Cerda-Flores and Dávila-Rodríguez/ Archives of Medical Research 31 (2000) 520–525

and birthplace was assessed using the nested gene diversity computer program (NEGST) developed by Chakraborty et al. (11). Third, the percentage contribution of ancestral populations to the hybrid population was calculated by the method of Chakraborty (12,13), each group considered the product of the admixture of two parental populations: Spanish and Mexican Indian. Fourth, the study population grouped by year of birth and birthplace was compared using heterogeneity of admixture proportions by a ␹2 test (14). Fifth, individual genetic admixtures were estimated by the method proposed by Hanis et al. (15). Finally, computation for nonrandom association of alleles among different genetic loci was conducted according to the methods of Brown et al. (16) and Chakraborty (17,18) to examine whether any residual effects of admixture remained in the women studied. Gene frequency data on the ancestral populations were obtained from the compilation of Mourant (19), the exact sources of which are available in Hanis et al. (20) (see Appendix 1). A value of p ⬍0.05 was considered significant.

Table 1. Allele frequencies among Mexican women by birthplace and year of birth Birthplace System

Allele

ABO

A1 A2 B O

Prob.a Rh

Prob.a MNSs

Prob.a Duffy

Results Allele frequency. The allele frequency estimates for the eight loci (Table 1) were used for a goodness-of-fit ␹2 test to determine whether the phenotype and genotype frequencies in this population grouped by birthplace and birth year depart from Hardy-Weinberg proportions. The phenotype (genotype) frequencies for the majority of the loci are in reasonable agreement with their Hardy-Weinberg expectations. Only the Rh system showed statistical significance in the female population grouped by birthplace (p ⬍0.05).

Gene diversity analysis. The majority (99.13%) of the total average gene diversity (HT ⫽ 47.4%) can be attributed to individual variation within the population (Table 2). Only a small contribution to the total variability (0.67%) comes from the between-birthplaces level of subdivision. The birth year difference is even smaller (0.20%).

Genetic admixture analysis. Table 3 presents the estimated values of admixture based on seven genetic loci. In the present investigation, we considered this mestizo population as the product of the admixture of two parental populations of Spanish and Mexican-Indian ancestry. The allele frequencies of the ancestral populations are presented in Appendix 1. The women of NE had a predominant Spanish contribution (68.97%) that was different from the women of Not-NE (54.13%) (␹2 ⫽ 15.97, p ⬍0.05). Additionally, a significant difference was found in persons born between 1896 and 1925 (␹2 ⫽ 12.14, p ⬍0.05) and between 1926 and 1955 (␹2 ⫽ 13.90, p ⬍0.05).

Prob.a Kidd

Prob.a Lewis

Prob.a Kp Prob.a Kell Prob.a

DCE DCe DcE Dce dCE dCe dCe dce MS Ms NS Ns Fya Fyb Fy Jka Jkb Jk Lea Leb le Kpa Kpb K k

Year of birth

NE

Not-NE

1896–1925

1926–1955

Total

0.133 0.027 0.071 0.769 ⬎0.05 0.023 0.359 0.200 0.032 0.021 0.041 0.113 0.212 ⬍0.05 0.276 0.363 0.081 0.280 ⬎0.05 0.432 0.394 0.174 ⬎0.05 0.285 0.441 0.274 ⬎0.05 0.129 0.533 0.338 ⬎0.05 0.053 0.947 ⬎0.05 0.014 0.986 ⬎0.05

0.115 0.029 0.057 0.800

0.139 0.034 0.061 0.766 ⬎0.05 0.048 0.398 0.262 0.012 0.019 0.027 0.052 0.182 ⬎0.05 0.264 0.381 0.065 0.290 ⬎0.05 0.451 0.367 0.182 ⬎0.05 0.243 0.487 0.269 ⬎0.05 0.100 0.546 0.354 ⬎0.05 0.050 0.950 ⬎0.05 0.003 0.997 ⬎0.05

0.121 0.023 0.064 0.792

0.129 0.028 0.063 0.780

0.064 0.393 0.202 0.051 0.016 0.024 0.111 0.139

0.059 0.393 0.225 0.034 0.014 0.028 0.088 0.158

0.242 0.394 0.108 0.256

0.251 0.389 0.089 0.271

0.453 0.365 0.182

0.452 0.366 0.182

0.309 0.385 0.305

0.278 0.431 0.290

0.107 0.526 0.367

0.104 0.535 0.361

0.050 0.950

0.050 0.950

0.020 0.980

0.120 0.988

0.096 0.453 0.258 0.075 0.000 0.000 0.064 0.053 0.227 0.415 0.096 0.262 0.473 0.337 0.189 0.271 0.421 0.308 0.078 0.538 0.384 0.048 0.952 0.010 0.990

a

Prob.: Probability for homogeneity among birthplaces and year of birth; NE: Women whose four grandparents were born in the northeastern Mexican states (Nuevo León, Coahuila, San Luis Potosí, Tamaulipas, or Zacatecas); Not-NE: Women whose four grandparents were born outside the northeastern states of Mexico.

Individual admixture analysis. Individual admixture estimates were calculated for each of the 417 females for seven loci. Distribution in percentages of individuals with each admixture proportion (proportion of genes of Spanish descent) ranged from 0.0 (none) to 1.0 (full), as shown in Figure 2. A great proportion of values is closer to 1.0 than 0.0, with a mean of 0.69, indicating a great influence of Spanish ancestry on the genetic constitution of the women of NE. Among Not-NE women the mean was 0.54, indicating less influence from Spanish ancestry. Table 4 presents the values of individual admixture by birthplace and year of birth. A statistically

Cerda-Flores and Dávila-Rodríguez / Archives of Medical Research 31 (2000) 520–525

523

Table 2. Gene diversity analysis of allele frequency data from the selected women population residing in Nuevo León GSTa

HTb

Locus

Within population

Between birthplace

Between year of birth within birthplace

Total gene diversity

ABO Rh MNSs Duffy Lewis Kidd Kell Kp Mean SEc

99.83 97.23 99.59 99.78 99.70 99.19 99.25 99.86 99.13 ⫾0.44

0.10 2.47 0.30 0.22 0.28 0.20 0.05 0.14 0.67 ⫾0.41

0.08 0.30 0.11 0.00 0.02 0.61 0.70 0.00 0.20 ⫾0.09

0.372 0.747 0.704 0.628 0.573 0.650 0.022 0.095 0.474 ⫾0.10

a

Expressed as percentage of total; bAbsolute total gene diversity in the entire sample; cStandard error.

significant difference was only found between the two groups of women stratified by birthplace (␹2 ⫽ 19.82, p ⬍0.05). Nonrandom association among genetic loci. It is well known that the mixture of populations with disparate allele frequencies can produce nonrandom association of alleles at two or more unlinked loci (21–23). From the genotype data available on each woman, we defined the multilocus genotype for each female for eight loci. The number of loci with respect to which individual was heterozygous was determined. This generated an observed distribution of the number of heterozygous loci across 417 females. Chakraborty (17) provided a numerical algorithm to compute the expected distribution for such observations, assuming random association of alleles at the different loci. Table 5 shows the results for our subdivided sample. For each group, the observed distribution agrees with that expected (p ⬎0.05). Table 6 shows the results for the present study and the previous results reported by Cerda-Flores et al. (6). In this study, the mean number of heterozygous loci is 3.81 and the variance is 1.44. Their expected values (under the random association model) are 3.76 and 1.45, respectively. The 95% confidence limit of the variance is 1.25–1.63. Clearly, these values provide no evidence of nonrandom association of the Table 3. Percentage contribution from Spanish and Mexican Indian gene pools to the Mexican women by year of birth and birthplace Year of birth

Birthplace

1896–1925

NE Not-NE NE Not-NE NE Not-NE

1926–1955 Total

n

Spanish

Mexican Indian

␹2

102 91 111 113 213 204

65.22 ⫾ 0.99 53.35 ⫾ 3.26 72.28 ⫾ 3.51 54.99 ⫾ 3.03 68.97 ⫾ 2.32 54.13 ⫾ 2.90

34.78 ⫾ 0.99 46.65 ⫾ 3.26 27.72 ⫾ 3.51 45.01 ⫾ 3.03 31.03 ⫾ 2.32 45.87 ⫾ 2.90

12.14a 13.90a 15.97a

The computations are carried out with seven polymorphic loci (ABO, Rh, MNSs, Duffy, Lewis, Kidd, and Kell); ap ⬍0.05.

Figure 2. Distribution of estimates of individual admixture based on seven genetic markers for Mexican women whose four grandparents were born in the northeastern (NE) and outside the northeastern (Not-NE) states of Mexico.

alleles among the eight polymorphic loci in the total Mexican female population residing in the state of Nuevo León. When the selected female and random population distributions were compared, no statistical significance was found (␹2 ⫽ 2.33, p ⬎0.05).

Discussion The results of gene differentiation for this selected female population suggest that overall the level of gene diversity is small and that more than 99.1% of total gene diversity is accounted for by individual variation within the population. The results obtained in this nonrandom population (selected in seven IMSS medical units) are similar to the random population study of the NE (6). The dihybrid model showed a greater Spanish contribution. However, the Spanish component is more pronounced

Table 4. Birthplace and year of birth according to degree of admixture Birthplacea

Year of birthb

Admixture

NE

Not-NE

1896–1925

1926–1955

Total

.00–.09 .10–.19 .20–.29 .30–.39 .40–.49 .50–.59 .60–.69 .70–.79 .80–.89 .90–1.0 Total

35 9 15 24 16 23 16 16 12 47 213

57 14 18 21 19 9 16 7 12 31 204

42 11 17 26 14 16 14 10 9 34 193

50 12 16 19 21 16 18 13 15 44 224

92 23 33 45 35 32 32 23 24 78 417

The computations are carried out with seven polymorphic loci (ABO, Rh, MNSs, Duffy, Lewis, Kidd, and Kell); a␹2 ⫽ 19.82; p ⬍0.05; b␹2 ⫽ 4.65; p ⬎0.05.

524

Cerda-Flores and Dávila-Rodríguez/ Archives of Medical Research 31 (2000) 520–525

Table 5. Observed and expected distribution of the number of heterozygous loci in Mexican women grouped by year of birth and birthplace Year of birth 1896–1925 Number of heterozygous loci 0–1 2 3 4 5 6–8 Total Mean Variance 95% CI for variance ␹2 (df ⫽ 5) Probability

Birthplace 1926–1955

NE

Not-NE

Obs.

Exp.

Obs.

Exp.

Obs.

Exp.

Obs.

Exp.

5 12 47 68 46 15 193 3.96 1.36 (1.17,1.72) 7.20 ⬎0.05

5.73 22.53 50.31 61.75 40.08 12.60

6 16 58 78 51 15 224 3.89 1.34 (1.18,1.69) 3.08 ⬎0.05

5.13 23.55 56.07 72.67 49.40 16.77

5 10 47 80 51 20 213 4.06 1.31 (1.18,1.70) 9.70 ⬎0.05

5.11 21.89 52.55 69.00 47.85 16.60

6 18 58 66 46 10 204 3.79 1.36 (1.25,1.63) 3.53 ⬎0.05

6.63 25.24 54.69 64.87 40.38 12.18

3.76 1.45

3.84 1.44

3.86 1.44

3.71 1.45

The computations are carried out with eight polymorphic loci (ABO, Rh, MNSs, Duffy, Lewis, Kidd, Kell, and Kp); Obs. ⫽ observed; Exp. ⫽ expected.

in women in each birth year whose four grandparents were born inside rather than outside of the northeastern states. These differences are well represented in the computation of the individual admixture shown in Figure 2 for NE and Not-NE women, suggesting that the two populations have different ancestral contributions. This finding can be explained by the immigration of persons from South and Central Mexico, who have a greater Mexican Indian contribution. These results are consistent with results previously published on genetic admixture in random population studies in the northeastern states of Mexico (4,6). The expected and observed distributions of the number of heterozygous loci by birth year and birthplace indicate that there is no residual effect of such admixtures on the nonrandomness of allelic associations at the polymorphic loci (Table 5). These results suggest that the admixture occurred sufficiently long enough ago that at present mestizo women are a genetically homogeneous population, at least with respect to the eight loci measured. Table 6. Comparison of the distribution of the number of heterozygous loci between present study results and the previous results Present study

Previous resultsa

Number of heterozygous loci

Obs.

Exp.

Obs.

Exp.

0–1 2 3 4 5 6–8 Total Mean Variance

11 28 105 146 97 30 417 3.81 1.44

11.05 45.63 106.07 134.69 89.97 29.60 417.00 3.76 1.45

13 38 125 168 117 50 511.00 3.97 1.59

13.69 53.63 118.47 150.16 113.01 62.05 511 3.97 1.75

The computations are carried out with eight polymorphic loci (ABO, Rh, MNSs, Duffy, Lewis, Kidd, Kell, and Kp); aCerda-Flores et al., 1991; Goodness-of-fit ␹2: (present study vs. previous results) ⫽ 2.33, df ⫽ 5, p ⬎0.05.

There is no statistical significance found when comparing the homogeneous distributions (Table 6) of the selected women and the random population from northeastern Mexico. This finding corroborates the utility of this selected population for future genetic and epidemiological studies. Acknowledgments The authors are grateful to Medical Units 2, 5, 26, 28, 31, 32, and 37 of the IMSS in Monterrey, Nuevo León, Mexico for use of their Appendix 1: Allele frequencies for seven genetic loci System A B O DCE DCe DcE Dce dCE dCe dcE dce MS Ms NS Ns Fya Fyb ⫹ Fy Jka Jkb ⫹ Jk Le le K k

Spanish

Mexican Indian

0.310 0.067 0.623 0.048 0.418 0.090 0.049 0.002 0.011 0.001 0.380 0.243 0.311 0.057 0.389 0.365 0.635 0.537 0.463 0.699 0.301 0.038 0.962

0.063 0.003 0.934 0.022 0.626 0.330 0.000 0.000 0.000 0.000 0.022 0.346 0.444 0.080 0.130 0.820 0.180 0.360 0.640 0.616 0.384 0.000 1.000

The allele frequencies for the ancestral populations are compiled from the literature (see Hanis et al., 1991 for exact sources from the 1976 compilation of Mourant et al.).

Cerda-Flores and Dávila-Rodríguez / Archives of Medical Research 31 (2000) 520–525

facilities to sample and interview study participants. The authors would additionally like to acknowledge Diego González-Ramírez, Miguel Angel Zúñiga-Charles, and Sara Ann Barton for all their help, and the anonymous reviewer for his/her constructive criticisms and suggestions on an earlier draft.

References 1. Instituto Nacional de Estadística Geografía e Informática. XI Censo General de Población y Vivienda. Vol. I. Nuevo León. México, D.F.: Instituto Nacional de Estadística Geografía e Informática;1991. 2. Dirección de Estadística y Procesamiento de Datos del Gobierno de Nuevo León. Monterrey, Nuevo León, Mexico: Cifras de Nuevo León;1978. 3. Montemayor-Hernández A. Historia de Monterrey. México: Asociación de Editores de Monterrey, A.C.;1971. 4. Cerda-Flores RM, Ramírez-Fernández E, Garza-Chapa R. Genetics admixture and distances between populations from Monterrey, Nuevo León, Mexico and their putative ancestral populations. Hum Biol 1987;59:31. 5. Cerda-Flores RM, Garza-Chapa R. Variation in the gene frequencies of three generations of human from Monterrey, Nuevo León, Mexico. Hum Biol 1989;61:249. 6. Cerda-Flores RM, Kshatriya GK, Barton SA, Leal-Garza CH, GarzaChapa R, Schull WJ, et al. Genetic structure of the population migrating from San Luis Potosí and Zacatecas to Nuevo León in Mexico. Hum Biol 1991;63:309. 7. Crawford MA, Gottman FE, Gottman CA. Microplate system for routine use in blood bank laboratories. Transfusion 1970;10:258. 8. Lapinski FJ, Crowley KM, Merrit CA, Henry JB. Use of microplate methods in paternity testing. Am J Clin Pathol 1978;70:766. 9. Del Hoyo H. Historia del Nuevo Reino de León (1577–1723). 2nd ed. México: Editorial Libros de México;1979. 10. Reed TE, Schull WJ. A general maximum likelihood estimation program. Am J Hum Genet 1968;20:579.

525

11. Chakraborty R, Haag M, Ryman N, Stahl G. Hierarchical gene diversity analysis and its implication to brown trout population data. Heredity 1982;97:17. 12. Chakraborty R. Gene identity in racial hybrids and estimation of admixture rates. In: Ahuja YR, Neel JV, editors. Genetic microdifferentiation: human and other populations. New Dehli, India: Indian Anthropological Association;1985. p. 171. 13. Chakraborty R, Ferrell RE, Barton SA, Schull WJ. Genetic polymorphism and fertility parameters in the Aymara of Chile and Bolivia. Am Hum Genet 1986;50:69. 14. Rao CR. Linear statistical inference and its applications. 2nd ed. New York: John Wiley;1979. 15. Hanis CL, Chakraborty R, Ferrell RE, Schull WJ. Individual admixture estimates: disease associations and individual risk of diabetes and gallbladder disease among Mexican-Americans in Starr County, Texas. Am J Phys Anthropol 1986;70:433. 16. Brown AHD, Feldman MW, Nevo E. Multilocus structure of natural populations of Hordeum spontaneum. Genetics 1980;96:523. 17. Chakraborty R. The distribution of the number of heterozygous loci in natural populations. Genetics 1981;98:461. 18. Chakraborty R. Detection of non-random association of alleles from the distribution of the number of heterozygous loci in a sample. Genetics 1984;108:719. 19. Mourant AE, Kopec AC, Domaniewska-Sobczak K. The distribution of the human blood groups and other polymorphisms. London: Oxford University Press;1976. 20. Hanis CL, Hewett-Emmett D, Bertin TK, Schull WJ. The origins of US Hispanics: implications for diabetes. Diabetes Care 1991;14:618. 21. Li CC. Population genetics. Chicago, IL, USA: University of Chicago Press;1955. 22. Nei M, Li WH. Linkage disequilibrium in subdivided populations. Genetics 1973;75:213. 23. Chakraborty R, Weiss KM. Admixture as a tool for finding linked genes and detecting that difference from allelic association between loci. Proc Natl Acad Sci USA 1988;85:9119.