Computational protocol: How do introgression events shape the partitioning of diversity among breeds: a case study in sheep

Similar protocols

Protocol publication

[…] The presence of null alleles was tested using FreeNA [] i.e. loci with an estimated frequency of null alleles (r) higher than 0.2 were considered as potentially problematic for calculations []. Allele frequencies, numbers of alleles, observed heterozygosity (Ho), non-biased expected heterozygosity (He), effective number of alleles (Ae) and F-statistics [] were estimated with GENETIX 4.05.2 []. GENEPOP 4.07 [] was used to evaluate departure from Hardy-Weinberg equilibrium and pairwise genic differentiation among breeds []. Allelic richness (Ar) was computed with the rarefaction method using FSTAT 2.9.3.2 []. Significance levels of the tests were corrected with sequential Bonferroni correction on loci. Potential hierarchical genetic structure was investigated with the AMOVA procedure implemented in ARLEQUIN 3.5.1.2 []. Breeds were divided into different groups according to: (i) type (dairy, meat, hardy meat, and high prolificacy; see Table ) or (ii) original geographic location (Massif Central, South-West, South-East, North-West, Plain from Center/Northern part of France, and the United-Kingdom (UK); see Table ). The Mérinos de Rambouillet breed (MeRa) was excluded from all the AMOVA analyses because this breed is the unique representative of its type (patrimonial). Romanov (Roov) and Texel (Texe) breeds were also excluded because they were the only representatives of their geographical groups (respectively, Russia and The Netherlands). Significance levels were determined after 16 000 permutations.The matrix of Reynolds distances (DR []) was computed using PHYLIP 3.69 [] and used to draw a Neighbor-Net [] network with SPLITSTREE 4.5 []. A principal component analysis was also performed using PCAGEN []. The significance of the axis was evaluated using permutation tests (1000 randomizations of the genotypes).Clustering approaches were performed on the 51 breeds using a Bayesian clustering procedure implemented in STRUCTURE [] with the number of K clusters ranging from 1 to 10 and then equal to 15, 20, 25, 30, 35, 40, 45, 48, 51, and 55. For each value of K, 50 runs were performed with 1 000 000 iterations following a burn-in period of 100 000, under the admixture and correlated allele frequency model. Since consistency across runs seems to be an informative method for assessing species structure across breeds [, ], we used CLUMPP [] to estimate the similarity function G’ over runs for the different values of K, using the LARGEKGREEDY algorithm. We selected a subset of runs that included the run with the highest number of similar runs (symmetric similarity coefficients (SSC) greater than 0.90) grouped with the corresponding similar runs. We used this subset to compute a mean Q-matrix. Breed assignment was performed as in Leroy et al. []. Animals were considered as correctly assigned to their breed if they were primarily associated to the cluster that included the largest number of animals belonging to the breed, using results for K = 51. For clusters that comprised two breeds, runs were performed for K = 2 using only the breeds that were associated within the sub-cluster.The contribution of each breed to the diversity of the whole set of breeds was computed according to the method of Caballero and Toro []. Let pki be the average frequency of allele k in breed i, then, the average coancestry between breeds i and j is:1fij=ΣkpkipkjWhen several markers are used, coancestry is averaged over loci. The total genetic diversity (GDT) is assumed to be the sum of the within-breed genetic diversity (GDWS) and the between-breed genetic diversity (GDBS):2GDT=1‐ΣiΣjfij/n2,3GDWS=1−Σifii/n,4GDBS=ΣiΣiDij/n2In these equations, n is the number of breeds and Dij is Nei’s minimum distance between breeds i and j. Contribution of a breed to the diversity of the whole set of breeds was computed by the loss or gain of diversity ∆GD when the breed is removed. […]

Pipeline specifications

Software tools Genepop, Arlequin, PHYLIP, SplitsTree, CLUMPP
Applications Phylogenetics, Population genetic analysis
Organisms Ovis aries