Genetic diversity of Creole hens of the Colombian southwest

In Colombia, the Creole hen is a socially important species due to its nutritional value, easy reproduction, adaptability and hardiness. Therefore, it was necessary to assess its diversity and genetic relationships and structure. The aim of this study was to contribute to the biodiversity knowledge in the country and to the agricultural sustainable development through the study of the genetic diversity in Creole birds. A total of 224 Creole hen samples taken from the departments of Cauca, Caldas, Chocó, Nariño, Valle and from 20 commercial lines were analysed using 17 microsatellite markers. DNA was extracted using the Salting Out method and was further amplified by the polymerase chain reaction (PCR). The statistical analysis for the evaluation of genetic diversity was performed by using ARLEQUIN program Version 3.5, GENALEX Version 6.5, Microsatellite Toolkit and FSTAT software. The population structure was determined by using STRUCTURE software, version 2.3.4. A total of 79 alleles were identified for the populations; the average number of alleles per hen was 4.65 ± 1.66. The expected heterozygosity was higher than what was observed and varied from 0.59 in Chocó to 0.62 in Valle del Cauca. The FIS value was 0.40 (p<0.001). The genetic difference between Creole birds and commercial-line birds estimated by the FST value was 0.10 (p<0.001). The highest genetic distances were found between Caldas and the rest of populations, and the lowest distances were found between Nariño and Chocó. A high genetic diversity was estimated. However, a deficit in heterozygotes was detected, indicating mating between related individuals, which may favour the expression of undesired genes and genetic diseases, compromising viability and decreasing reproduction. Two groups were found: the first was formed by birds from Chocó, Nariño and Valle del Cauca and the second by Cauca and Caldas birds.


INTRODUCTION
Creole hens are widely distributed in rural areas, where they are maintained mainly by the local inhabitants and significantly contribute to both the food supply and the economic sustainability of rural hou-seholds.These birds are resistant, adapted to the rural environment, survive with few inputs and tend to adapt to the food availability fluctuations (Zaragoza, 2012).Their broodiness is one of the main advantages when compared to commercial birds (Valencia, 2009).Local birds are characterized by low or medium re-turns and are kept in small populations.These animal populations face genetic erosion, which leads to the loss of genetic variability.This variability is important for the retention of specific features because these local species possess genes and characteristics relevant to their adaptation to specific environments and to meeting local demands (Romanov et al., 1996).According to the United Nations Food and Agriculture Organization (FAO, 2007), the main threats for animal genetic resources are massive production, importation of foreign birds and indiscriminate breeding.With the decrease in peasant populations, several cultural traditions and knowledge for breeding, management and production have been lost (Valencia, 2009).
The study hypothesis is as follows: there is high diversity and genetic differentiation between and within Creole hen populations in the Colombian southwest.The aim of the study was to evaluate the diversity, differentiation and genetic relationships in populations from Chocó, Valle del Cauca, Cauca, Nariño and Caldas using 17 microsatellites.

Sampling
Two hundred twenty-four Creole hends were sampled in the departments of Caldas, Cauca, Chocó, Nariño and Valle del Cauca, of which 113 were males and 111 females.A group of 20 commercial birds was also included (Hy Line, Lohmann, Ross and Cobb 500) (table I).The Creole birds were sampled according to the types of hen described by Valencia, (2008).This characterization included phenotypic features such as naked neck, feathered tarsus, and curly plumage, among others.

Dna extraction anD pcr amplification
DNA was extracted from 3-4 µl of blood following the Salting Out extraction protocol (Miller et al., 1988).DNA quantification was determined using known concentrations of lambda bacteriophage DNA (Thermo Scientific, Waltham, Massachusetts, USA) in 0.8 % agarose gels stained with ethidium bromide (J.T. Baker, Warren, New Jersey, USA).The 17 microsatellite markers were PCR-amplified.Each PCR contained a total volume of 25 µL, from: 1X Taq buffer (Thermo Scientific, Waltham, Massachusetts, USA), 4 mM MgCl 2 (Thermo Scientific, Waltham, Massachusetts, USA), 0.2 mM dNTPs (Thermo Scientific, Waltham, Massachusetts, USA), 0.4 mg/mL bovine serum albumin (BSA) (Amresco, Solon, Ohio, USA), 10 ng DNA, 0.4 µM of each primer (Bioneer, Korea), and 0.03 U Taq polymerase (Thermo Scientific, Waltham, Massachusetts, USA); the final volume was adjusted with distilled water.The amplification was performed as follows: initial denaturation at 94 ºC for 10 minutes, followed by 35 cycles of denaturation at 94 ºC for 0.75 min, annealing at 57-64 ºC depending on the microsatellite for 1 min, extension at 72 ºC for 1.5 min and a final extension at 72 ºC for 10 minutes.The amplified product was subsequently visualized in a 1 % agarose gel stained with ethidium bromide.

StatiStical analySiS
The following diversity indices were estimated: total number of alleles (TNA), mean number of alleles per locus (MNA), mean effective number of alleles (MENA), observed (Ho) and expected heterozygosity (He), Hardy-Weinberg equilibrium (HWE) and Wright's F-statistics: F IS , F ST and F IT were calculated using Arlequin (Excoffier et al., 2005) and GenAlex (Peakall and Smouse, 2006) software.The polymorphic information content (PIC) was calculated using the Microsatellite Toolkit software for Excel (Park, 2001).The allelic richness (Rt) was calculated using the FSTAT software (Goudet, 2001).An analysis of molecular variance (AMOVA) was performed with different hierarchical levels: between Creole and improved (224 Creole individuals and 20 commercial-line individuals); between communities: native (101 individuals sampled from indigenous territories of Cauca and Nariño), afro-Colombian (51 individuals from Chocó) and rural (72 individuals from Valle del Cauca and Caldas).Fixation indices were calculated, and the population structure was determined based on the F ST .Likewise, the FST was calculated based on population pairs, and the Reynolds' genetic distance was also calculated (Reynolds et al., 1983) using Arlequin software, version 3.5 (Excoffier et al., 2005).Population structure was determined using the cluster analysis method based on models implemented in Structure software, version 2.3.4 (Pritchard et al., 2000).The clustering of individuals was tested assuming an increased number of groups (K) and using a mixture model with the correlated allelic frequencies.The procedure was implemented with 500,000 iterations after a 100,000 burn-in period for each (K).To identify the most probable (K) value, between 2 and 6 independent simulations were performed using a ∆K modal distribution (Evanno et al., 2005).

microSatellite marker polymorphiSmS
In the analysis obtained from 224 Creole and 20 commercial birds, a total of 79 alleles were detected, ranging from three alleles (ADL278 and MCW78) to 10 (MCW69), with a mean of 4.65 ± 1.66 alleles per locus (table II).
The expected heterozygosity (He) was higher than the observed heterozygosity (Ho).This last parameter ranged from 0.16 ± 0.03 to 0.65 ± 0.06; the lowest value was for microsatellite ADL268, and the highest value was for MCW183, with a mean of 0.35 ± 0.06.He values ranged from 0.42 ± 0.06 to 0.77 ± 0.02, with a mean of 0.60 ± 0.03; the markers with the highest values were MCW69, MCW34 and MCW206.
A total of 76 % of the microsatellites presented PIC values over 0.5, which were characterized as highly informative; 24 % were reasonably informative; the three markers with the highest PIC values were MCW69, MCW34 and MCW183.
The F IS ranged from 0.04 (MCW183) to 0.74 (ADL268) except for MCW183 was not significant (Ns) and MCW98, which was significant (p<0.05).The rest of the F IS values were highly significant (p<0.001), with a mean value of 0.41 ± 0.05, indicating an excess of homozygotes.Most microsatellites (65 %) showed deviations from HWE (p<0.05).

genetic DiverSity within populationS
The estimation of genetic diversity within populations is shown in table III.
The MNA for Creole populations was 4.24 ± 1.18, while the MENA was 2.72 ± 0.23.The mean He value (0.60 ± 0.14) was higher than the mean Ho value (0.35 ± 0.20) in all populations.The Ho ranged from 0.33 ± 0.18 to 0.42 ± 0.17, while the He ranged between 0.59 ± 0.13 and 0.62 ± 0.10, indicating that there is high genetic diversity in the Creole hens from the departments included in the study.The average F IS between populations was very high, positive and highly significant (F IS = 0.40), indicating that all five populations had heterozygote deficiency.

genetic relationShipS between populationS
Population structure analysis with different hierarchical levels revealed that greater genetic differentiation was obtained when comparing the Creole and commercial-line birds (F ST = 0.10, p<0.001).The diffe-  rences were lower when Creole birds were compared (F ST = 0.03; p<0.001) and when the afro-descendant, native and rural communities were compared (F ST = 0.02; p<0.001).
The Reynolds' genetic distance (1983) and the F ST pairwise estimations (table IV) revealed that the populations of Caldas/Choco were the most distant and that the Nariño/Chocó group had the lowest genetic distance.However, the commercial lines showed lower distance and genetic differences compared to the populations from Cauca and higher distance and differences compared to the population from Caldas.

genetic Structure
According to the structure results, out of the five Creole bird populations, the most probable number for K was two ancestral populations (figure 1).

microSatellite marker polymorphiSmS
The majority of the markers used were reliable and informative because 15 out of 17 had numbers of alleles higher or equal to four, as recommended by Nassiri et al., 2007 andNassiry et al., 2009.The microsatellite panel used here was adequate for the genetic evaluation of all five Creole hen populations because, based on the PIC, all systems were determined as polymorphic.According to the classification reported by Botstein et al. (1980), a total of 13 out of 17 were highly informative, while four were reasonably informative.The mean PIC value (0.58) was higher compared to the value (0.46) reported by Osei-Amponsah et al. (2010) in two bird populations (Forest and Savannah) in Ghana.However, the value was lower than the one reported by Clementino et al. (2010) in ecotypes of Brazilian birds (0.73).
The mean number of alleles per locus (4.65 ± 1.66) was higher than the numbers reported with 22 markers by Hillel et al. (2003) in 52 populations, Berthouly et al. (2008) in 14 French native and six Japanese breeds, and Leroy et al. (2007) in African and Asian populations.It was also higher than the number reported by Tadano et al. (2007b) in nine populations of Japanese native populations with a panel of 40 systems.However, Ramadan et al. (2012) obtained a higher average number of alleles per locus in Egyptian birds, as did Tadano et al. (2007a) in commercial lines and Dávila et al. (2009) in Spanish birds.Most markers had a He above 0.5, especially MCW69, MCW34 and MCW206, which suggests the necessity for future studies on genetic diversity in hens.The mean He was 0.60 ± 0.03; according to Menezes (2005), an He value above 0.5 indicates high marker genetic diversity.
The mean F IS value was 0.41 ± 0.05, which was higher than the values reported by other researchers (Dorji et al., 2012, Nasr et al., 2012, Ramadan et al., 2012, Osei-Amponsah et al., 2010, Dávila et al., 2009and Tadano et al., 2007b).This result is an indicator of a heterozygote deficit.Eleven out the 17 microsatellites deviated from HWE (p<0.05).Dorji et al. 2012 andClementino et al. 2010 reported lower percentages of HWE deviation (40 %) and (44 %) in native birds from Bhutan and in ecotypes of Brazilian birds, respectively.The behaviors of both the FIS and the HWE can be attributed to the condition of domestic populations, where mating is not random, the proportion of males is lower compared to females, and roosters remain in the production systems for long periods of time.Moreover, most populations are very small (approximately 14 birds/producer), and subdivisions can occur between populations, including natural selection, leading to endogamy and genetic drift.

genetic DiverSity within populationS
The MNA obtained for all five Creole populations (4.24 ± 1.18) is a recommended value by Barker (1994).According to Barker (1994), the mean number of alleles per locus should be higher than four to reduce the effect of the standard error when calculating the genetic distance between populations.The mean MENA (2.72 ± 0.23) was higher than the one obtained from Balearic island races (2.47) by Mendez (2012) and was similar to the value reported by Osei-Amponsah et al. (2010) in Ghana local birds (2.8 ± 1.3) and by Ramadan et al. (2012) in Egyptian birds (2.8 ± 1.2).
Table IV.Genetic differentiation and distances estimated for six populations.The pairwise population FST is indicated above the diagonal and Reynolds' distance below it (Estimaciones de diferenciación genética y distancias genéticas para seis poblaciones.El F ST por pares de poblaciones se indica por encima de la diagonal y la distancia de Reynolds, debajo de la diagonal).The He value of Colombian Creole birds was high (60.0%), a much higher value than the one reported by Hillel et al. (2003) in 52 populations (He= 47 %), which included red rooster subspecies (Gallus gallus gallus and G. gallus spadiceus), indigenous populations, and commercial and experimental lines.The He measured in this study was the same as the one for the Gallus gallus domesticus subspecies from the Red Jungle Fowl but was lower than the value for Gallus gallus spadiceus (64 %).This difference indicates that the genetic diversity of Colombian Creole birds is comparable to the genetic diversity of ancestral birds.
The mean Ho value (0.35 ± 0.20) was lower than the mean He (0.60 ± 0.14).This was similar to the values reported by Tadano et al. (2007a) between Japanese native races and commercial lines, by Cuc et al. (2010) in Vietnamese local birds and by Dávila et al. (2009) in Spanish birds.
According to Simon and Buchenauer (1993), the mean F IS value (0.40) indicates that the population is in danger due to the high inbreeding, and conservation processes should be implemented.The high inbreeding is most likely due to the relatively small population sizes (approximately 14 birds/producer).Additionally, bird replacement is performed with animals born within the same systems, which suffer from high degrees of relatedness and the use of low numbers of male breeders, which are replaced at an advanced age, predisposing populations to high consanguinity rates.

genetic relationShipS between populationS
According to Tejedor et al. (1999), the F ST value of bird species is approximately 0.1 or even lower.As expected, the F ST value was higher for the cluster between Creole and improved birds (0.10).Several researchers have reported higher values than the ones measured here (Berthouly et al., 2008, Dávila et al., 2009and Zanetti et al., 2011).However, Osei-Amponsah et al. (2010) reported a lower value (0.04).
The clustering for all five Creole populations showed a genetic differentiation of 0.03, while between communities, the value was 0.02.Both values were low but highly significant, indicating that there is genetic structure in all groups.The similarity of the genetic diversity parameters between Creole birds clustered according to department indicated low genetic differentiation between them, as confirmed by the F ST values shown in table III.
The low genetic structure can possibly be attributed to a short evolutionary period, considering that the introduction of European hens to America has been accredited to the Spaniards.Five hundred years have passed since the birds' arrival in the 15 th century, during which they have essentially adapted to the climatic, food and parasite conditions.It is assumed that the birds have had a short evolutionary period compared to the domestication process, which, according to West and Zhou (1988), occurred in the Indus Valley 5000 years ago and in western China most likely 7500 or 8000 years ago.Another reason for the low genetic structure is that the Creole populations sampled do not belong to pure cores because they show high pheno-typic diversity in terms of colors, feather distribution (naked neck, feathered tarsus, ears and chins), eye and tarsus color, skeleton variants (tailless, dwarves) and comb types, among others.It is probable that the genetic flow has contributed to the similarities between birds of different departments, given the easy transport of fertile eggs and live birds and the sea connection between the departments.This connectedness allows the exchange of genes, increasing the mixture rate between populations.
Birds from Caldas had higher values for Reynold's genetic distance (1983) and the pairwise F ST estimation, indicating that this population is genetically different compared to others, even commercial populations, the above can be attributed presumably to the peasant tradition of the area and to be less mixed commercial birds.
genetic Structure The results obtained with Structure software indicated that two ancestral populations (K= 2), one formed by birds from the departments of Chocó, Nariño and Valle del Cauca, and the second formed by birds from Cauca and Caldas, maintain their own identities.

CONCLUSIONS
Creole hens from the Colombian southwestern region have high genetic diversity.However, an important heterozygote deficit was detected, indicating mating between related individuals, which may favor the expression of undesired genes and genetic diseases, compromise viability and decrease reproduction.Two clusters were identified: the first formed by birds from Chocó, Nariño and Valle del Cauca, and the second by birds from Cauca and Caldas.

Figure 1 .1
Figure 1.Genome distribution for each individual in both inferred clusters (K= 2).Each animal is represented by a vertical bar.The length of the bar colour in the vertical axis defines the membership proportion (Q), to either the first cluster (green) or the second (red) based on Bayesian analysis (structure) (Distribución del genoma de cada individuo en los dos clúster inferidos (K= 2).Cada animal está representado por una barra vertical.La longitud del color de las barras en el eje vertical define la proporción de membresía (Q) al clúster 1 (verde) o clúster 2 (rojo), basado en análisis Bayesiano (structure)).

Table I .
Number of samples (N) of Creole birds, per department, and municipality (Número de muestras (N) de aves Criollas por departamento y municipio).

Table II .
Genetic diversity parameters determined for 17 microsatellites in five Creole bird populations and one commercial bird population (Parámetros de di- versidad genética estimados para 17 microsatélites en cinco poblaciones de aves Criollas y una población de aves comerciales).
N= Populations, number of individuals per population; MNA= mean number of alleles per locus; MENA= mean of effective number of alleles per locus; AR= allelic richness with a minimum value N of 18 individuals; H e = mean expected heterozygosity; H o = mean observed heterozygosity and their respective standard deviations; F IS = endogamy coefficient within populations; HWE= Hardy-Weinberg equilibrium deviations of the number of loci estimated with 17 microsatellites in five Creole hen populations (p< 0.05); **p˂0.001.