Sampling

Samples were collected from January to December 2015 using a two-stage cluster sampling⁸. The seven districts comprised the primary units, and the production units (PU) within the districts were considered secondary units. A sample of three primary units (n) was selected; 17 secondary units were randomly selected within these primary units. From each secondary unit (PU), six chickens were selected, resulting in a total sample size of 109 adult chickens: 93 adult creole hens, in the early laying stage and weighing 2.0 kg on average, and 16 adult roosters. The estimated inventory of adult chickens was 2,004 animals in the Valles Centrales region; thus, 6 % of the population was sampled. The sample size was obtained with a precision of 10 % and reliability of 90 %. Sampling was distributed in six geographic cores (ITVO: Nazareno and ITVO; Cuilápam: San Antonio, San Lucas, and Cuilápam; Etla: San Juan and Suchilquitongo; Huixtepec: Huixtepec 1, 2, 3, and 4; Ocotlán: Chichicapam, San Antonino 1 and 2; and Tlacolula: Teotitlán, Totolapam 1 and 2, plus a Plymouth Rock chicken control. Blood samples (2 mL) were drawn from the cubital vein of the wing of each animal and collected into Vacutainer tubes with EDTA. Samples were stored at -20 °C until further processing.

DNA extraction, microsatellite markers, and PCR procedure

DNA was extracted using the ChargeSwitch® gDNA Plant Kit (Invitrogen), following the protocol provided by the manufacturer. The DNA was then quantified using an ultra-low volume spectrophotometer (NanoDrop 2000, Thermo Scientific, Wilmington, DE, USA); DNA concentrations were adjusted to 10 ng μl^-1. To evaluate genetic diversity, it was used ten pairs of previously reported^9-11 microsatellite primers (Table 1). Primers were fluorescently labeled (6-FAM, HEX, or ROX) at their 5' end for PCR multiplex.

The PCR reaction final volume was 25 µL, this included: 5 µL of 5X Buffer (Promega), 2 µL of 25 mM MgCl₂ (Promega), 0.5 µL of 2 mM dNTPs mix (Promega), 1 µL of each primer (5 pmol), 1 UI of Taq polymerase (Promega), and 1 µL of template DNA (10 ng µl^-1). The amplification was carried out in a thermocycler (Bio-Rad C1000^TM) with the following conditions: initial denaturation at 94 °C for 5 min, followed by 35 denaturation cycles at 94 °C for 45 s, annealing for primer groups at 53.6 °C, 56.6 °C, and 62 °C for 1 min, extension at 72 °C for 1 min, and final extension at 72 °C for 10 min. PCR products were separated by capillary electrophoresis in an automated DNA sequencer (Genetic Analyzer ABI-3130 Applied Biosystems, Foster City, CA, USA) and analyzed with the GeneMapper® software of Applied Biosystems.

Statistical analysis

Based on the allelic profile determined in each individual for each locus, it was calculated the allelic frequencies, observed heterozygosity (Ho), expected heterozygosity (He), and Hardy-Weinberg equilibrium. Wright's F-statistics (F_IS, F_IT, and F_ST) were estimated based on the population's degree of genetic differentiation, considering the geographic cores as a clustering criterion. The POPGENE software (Version 1.3.2) was used to define the genetic diversity parameters. Furthermore, using the allelic frequencies of each locus, a principal component analysis was carried out with SAS V. 9.4¹², as well as a clustering analysis using the UPGMA method, which is based on the Euclidean distances between the creole chicken and reference populations, to determine their pattern of genetic similarity.

Genetic diversity

A total of 109 alleles were identified in the 18 chicken populations. Considering the ten evaluated loci, the number of alleles ranged from six for MCW145 to 16 for MCW158, with an average of 10.9 alleles per locus. All loci were polymorphic (Table 2), coinciding with some of the primers with a chicken mapping⁹. MCW145 was the exception because it showed a lower value compared to previous reports^9,10.

Previous studies of different local creole chicken breeds reported a lower total number of alleles per locus using different microsatellite loci in the same species. In Sweden, Germany, and Poland, studies reported 113, 217, and 62 alleles with 24, 29, and 10 loci, respectively^1,13,14. Studies in China, Thailand, and India reported similar values; 276, 227, and 170 alleles with 29, 20, and 17 loci^6,15,16. Similarly, researchers in Israel and Iran^17,18 observed 211 and 310 alleles in local populations using 22 and 31 microsatellite loci, respectively, akin to what was observed in this study. The average number of reported alleles in the evaluated populations indicates a wide diversity of genes, similar to that in other countries with well-defined local chicken breeds.

The effective number of alleles per locus ranged from 2.1 (MCW95) to 10.9 (MCW135), with an average of 4.9. The observed heterozygosity (Ho) value ranged from 0.08 (MCW145 locus) to 0.92 (MCW134) (Table 2). Meanwhile, the expected heterozygosity (He) ranged from 0.54 (MCW95 and MCW145) to 0.91 (MCW135). Although these results suggest that the analyzed populations present a wide genetic diversity, they also indicate discrepancies between both types of heterozygosity, which indicates deviations from the Hardy-Weinberg equilibrium. These deviations result from migration or genetic drift processes, possibly attributed to the low number of individuals that generally make up each population. Other studies^10,19 have reported He values of 0.69 and 0.75, respectively, for the MCW145 locus. These results were higher than what was observed in this study (0.54) because this locus is associated with a lower number of alleles, significantly influencing the He since each allele represents an additional term to the summation for He calculation. Previous studies^10,20 for the MCW134 locus reported He values of 0.68 and 0.64 in China, respectively. These values are lower than the value obtained in this study (0.805).

The number of alleles found in populations per loci was in descending order: Plymouth Rock reference genotype, with 70 alleles, followed by Cuilápam with 65 alleles, contrasting with the Teotitlán population with 41 alleles, this being the lowest value (Table 3). However, the San Antonio 1 population had the higher effective number of alleles (43.8), followed by ITVO with 37.8 alleles; the Suchilquitongo population registered 29 effective alleles. These values imply the potential transmission of a high number of alleles to the following generation of creole chickens.

The observed heterozygosity (Ho) ranged from 0.575 in the San Lucas population to 0.75 in the San Antonino population, with an average of 0.741. Meanwhile, the expected heterozygosity (He) was lower in the reference group (0.625), suggesting that the kinship level between progenitors is lower. Furthermore, Huixtepec 2, San Lucas, and San Antonio 1 (0.733, 0.732, and 0.716, respectively) were the most genetically diverse populations, indicating greater potential in breeding programs. Previous studies in local chickens^2,21-23, carried out in Sweden, Korea, China, and India, reported lower Ho and He values. Moreover, in local Tibetan and Chinese chicken populations, previous studies have reported He values of 0.798 and 0.76^10,24, higher than those observed in this study, probably because of the use of different microsatellite loci.

Genetic differentiation of chicken populations

The Ocotlán and Tlacolula clusters had the lowest F_IS values (-0.072 and -0.079), indicating a higher number of heterozygous individuals within each population. Overall, the less than zero values, close to the Hardy-Weinberg equilibrium, of these clusters indicate that the creole chickens are non-inbred populations. Studies have emphasized the adaptation of local chickens to the different geographical conditions in each country; therefore, the Hardy-Weinberg equilibrium remains constant across generations^2,21.

As for the F_IT, the lowest values were observed at the production unit level in the Ocotlán and control clusters (-0.003 and -0.037). In the Ocotlán cluster, the constant change of male breeders through direct purchase in regional markets results in an important local poultry genetic material flow at a regional level. While in the control group, due to its origin and established reproduction scheme, the production of heterozygous individuals is maximized. The Cuilápam cluster had the highest F_IS (0.106) and F_IT (0.177) values compared to the other clusters. However, statistically, by establishing confidence intervals, its heterozygous individuals are in equilibrium (Table 4); this is explained by the lack of a male substitution process in the production units, increasing the kinship relationship and the proportion of homozygous loci.

The Cuilápam, ITVO, and Huixtepec clusters had a moderate genetic differentiation (F_ST)²⁵ (0.079, 0.077, and 0.076, respectively). These results indicate that these populations are different due to geographic isolation or handling practices, and, therefore, the genetic flow between the individuals from the different poultry regions is reduced. The global F_ST value of 0.059 indicates that 94.1% of the total variation is within the populations, only 5.9% between populations²⁵ (Table 4). This value confirms a moderate genetic diversity within the populations, which is associated with the selection strategies of breeding males and replacement females carried out by the producers and demonstrates the potential for intrapopulation genetic breeding through recurrent selection schemes, taking advantage of the selection and recombination effects in a continuous process. Similar results have been reported^26-28 in local chicken populations in Egypt, Bhutan, Asia, and China. Meanwhile, another study evaluating eight Korean domestic chicken breeds²⁹ reported a differentiation coefficient of 0.180.

Relationships between the populations

The principal component analysis (PC) explained 32.2 % of the total variation with the three first components integrated by the most important alleles of the ten microsatellite loci (Table 5). The three-dimensional plane of the populations (Figure 1), based on the first three principal components, identifies the most distant populations: Nazareno= A, ITVO= B, Suchiquitongo= J, and Huixtepec=N, characterized for not allowing commercial poultry genetic material, suggesting a significant genetic flow within the same populations. Other studies have reported clear differentiation in local chicken populations of Kenya³⁰ and India³¹, using PC and considering the two first components, with genetic variation values of 46.25 % and 23.37 %.

Figure 1 Three-dimensional plane of chicken populations with the principal components PC1 vs. PC2 vs. PC3, generated with the allelic frequencies of ten microsatellite loci 

The groups obtained through the cluster analysis (Figure 2) were similar to those observed in the three-dimensional plane of the principal component analysis; three main groups are distinguished (genetic distance of 0.90). The first (Nazareno, ITVO, San Lucas, and San Antonio) includes isolated and genetically different populations that resulted from the reduced genetic flow between these creole populations and poultry populations with a degree of recombination with known breeds, such as Plymouth Rock. Subgroup 2 includes populations with a higher degree of genetic kinship, evidenced by the proportion of alleles with the reference group (Plymouth Rock) as a result of the introduction of this commercial breed as technological packages, which leads to crosses with chickens adapted to the environmental and handling conditions of the different rural areas, causing a loss and wear of the adaptation and survival characteristics of the creole chickens. The third subgroup included the Huixtepec 4, San Antonino 2, Teotitlán, and Totolapam 1 populations; these populations share similar allelic frequencies, suggesting variability in unique phenotypic traits characteristic of creole chickens.

Figure 2 Dendrogram of the chicken populations, constructed with the UPGMA method based on the Euclidean distances obtained from the frequencies of 109 microsatellite alleles