Computational protocol: Estimates of effective population size and inbreeding in South African indigenous chicken populations: implications for the conservation of unique genetic resources

Similar protocols

Protocol publication

[…] Population representative samples were collected from four indigenous chicken breeds, kept for conservation purposes at the Poultry Breeding Resource Unit of the Agricultural Research Council: Venda (VD_C, n = 30), Ovambo (OV_C, n = 26), Naked Neck (NN_C, n = 29) and Potchefstroom Koekoek (PK_C, n = 29). For comparison, a random sample of village chickens from three South African provincial regions was also taken: Venda (Limpopo Province, VD_F, n = 30), Ovambo (Northern Cape Province, OV_F, n = 42) and the Eastern Cape Province (EC_F, n = 26). Village chicken populations were sampled from farming areas from which the current conservation flocks originated, which are Limpopo Province (VD_F chickens) and Northern Cape Province along the border with Namibia (OV_F chickens) as well as the Eastern Cape Province (EC_F chickens). Ninety-eight households were randomly selected from 23 villages of Vhembe and Mopani Districts in the Limpopo Province, Kgalagadi and Namaqua Districts of the Northern Cape Province and Alfred Nzo and OR Tambo Districts of the Eastern Cape Province. For each district, 2–5 villages were selected. The distance between villages within district ranged from 20 to 40 km, 100 to 500 km between districts within a province and over 1000 km between provinces. One chicken was sampled per household. Blood was collected from the wing vein onto FTA micro-cards (Whatman Bio Science, Kent, UK) for each individual. A standard phenol/chloroform extraction protocol was followed for DNA isolation (Sambrook and Russell ). Each individual animal was genotyped for a set of 29 autosomal microsatellite markers as previously described in Mtileni et al. ().Micro-checker v.2.2.3 (Van Oosterhout et al. ) was used to detect genotyping errors due to allelic dropout, stuttering and null alleles (null allele estimates as per the method of Brookfield ). Departures from Hardy-Weinberg equilibrium were evaluated by means of the exact probability test (500 batches, 10,000 iterations) in Genepop v.4.0 (Rousset ). Markers were also tested for neutrality using an Fst outlier test as implemented in Lositan v.1.44 (10,000 permutations assuming the infinite alleles model, with a correction for false discovery rate at 0.01 and statistical significance at the 5 % nominal level) (Antao et al. ). The following genetic diversity estimates were calculated in Genalex v.6.4 (Peakall and Smouse ): Observed (Ho) and expected heterozygosity (He), number of alleles (An), effective number of alleles (Ae) and the information (Shannon-Weaver) index (I). To test whether there were significant differences in diversity between the village (field) chickens and the conservation flocks a Kruskal-Wallis test was performed (nominal level of 5 % for statistical significance) in XLStatistics v.10.05.03 (Carr ). Pairwise Fst estimates (significance testing: 10,000 permutations, at a nominal level of 5 %) and a locus-by-locus hierarchical analysis of molecular variance (AMOVA) (significance testing: 10,000 permutations, at a nominal level of 5 %) was calculated in Arlequin v.3.5.1.2 (Excoffier and Lischer ). Where appropriate, the conservation flocks were grouped with the relevant geographically correlated field populations for the hierarchical AMOVA. To further elucidate the relationship between the various populations a dendrogram was constructed using Nei’s genetic distance, Da (Nei ) and the neighbour joining clustering method (significance testing: 1000 bootstrap replicates) in Treefit v.1.0 (Kalinowski ). A principle coordinate analysis (PCoA) was also conducted per population and sample in Genalex. Effective population sizes were calculated using the heterozygous excess test in NeEstimator v.1.3 (Peel et al. ) as well as the linkage disequilibrium (LD) test (minimum allele frequency, 0.02) in LDNe v.1.0 (Waples ). The occurrence of recent bottlenecks was evaluated by means of the Wilcoxon signed rank test [assuming the infinite alleles (IAM) and the two-phase mutation models (TPM), 10,000 replicates at 5 % nominal level] and the mode-shift test, in Bottleneck v.1.2.02 (Piry et al. ). Furthermore, to evaluate the extent of inbreeding, mean relatedness was calculated for each population using the method of Queller and Goodnight () (significance testing by 1000 bootstrap replicates) as well as mean Fis estimated in Genalex. […]

Pipeline specifications

Software tools Genepop, GenAlEx, Arlequin, NeEstimator
Application Population genetic analysis
Organisms Gallus gallus
Diseases Genetic Diseases, Inborn