Genetic diversity and historical demography of underutilised goat breeds in North-Western Europe

Abstract

In the last decade, several studies aimed at dissecting the genetic architecture of local small ruminant breeds to discover which variations are involved in the process of adaptation to environmental conditions, a topic that has acquired priority due to climate change. Considering that traditional breeds are a reservoir of such important genetic variation, improving the current knowledge about their genetic diversity and origin is the first step forward in designing sound conservation guidelines. The genetic composition of North-Western European archetypical goat breeds is still poorly exploited. In this study we aimed to fill this gap investigating goat breeds across Ireland and Scandinavia, including also some other potential continental sources of introgression. The PCA and Admixture analyses suggest a well-defined cluster that includes Norwegian and Swedish breeds, while the crossbred Danish landrace is far apart, and there appears to be a close relationship between the Irish and Saanen goats. In addition, both graph representation of historical relationships among populations and f4-ratio statistics suggest a certain degree of gene flow between the Norse and Atlantic landraces. Furthermore, we identify signs of ancient admixture events of Scandinavian origin in the Irish and in the Icelandic goats. The time when these migrations, and consequently the introgression, of Scandinavian-like alleles occurred, can be traced back to the Viking colonisation of these two isles during the Viking Age (793-1066 CE). The demographic analysis indicates a complicated history of these traditional breeds with signatures of bottleneck, inbreeding and crossbreeding with the improved breeds. Despite these recent demographic changes and the historical genetic background shaped by centuries of human-mediated gene flow, most of them maintained their genetic identity, becoming an irreplaceable genetic resource as well as a cultural heritage.

Figure 1

Principal component analysis (PCA) of allele frequencies that shows the population structure and the relationship between the twenty-four goat breeds analysed in the WHOLE dataset. Breeds acronyms according to Supplementary Table S8: Norwegian landraces (SKO, SEL, NRW); Irish landraces (ARR, BLB, OIG); Icelandic Landrace, ICL; Finnish Landrace, FIN; Swedish Landrace, SWE; Danish Landrace, DNK; Dutch Landrace, NLD; French breeds (Fosseé, FSS; Alpine, ALP_FR; Saanen, SAA_FR), Swiss breeds (Alpine, ALP_CH; Saanen, SAA_CH; Toggenburg, TOG); Spanish breeds (Bermeya, BEY; Malaguena, MLG); Italian breeds (Girgentana, GGT; Ciociara Grigia, CCG; Jonica, JON; Alpine, ALP_IT; Saanen, SAA_IT).

Figure 2

Population structure and relationship. Principal component analysis (PCA) that focuses on the relationship between the NW-EU populations dataset. Breeds acronyms according to Supplementary Table S8: Norwegian landraces (SKO, SEL, NRW); Irish landraces (ARR, BLB, OIG); Icelandic Landrace, ICL; Finnish Landrace, FIN; Swedish Landrace, SWE; Danish Landrace, DNK.

Figure 3

Model-based clustering. Circular representation of Admixture software results for K = 2, 4, 6, 10, 14, 18, 22 and 28, where K22 indicates the best estimate of number of genetic components based on the lowest cross-validation error value. Breeds acronyms according to Supplementary Table S8: Norwegian landraces (SKO, SEL, NRW); Irish landraces (ARR, BLB, OIG); Icelandic Landrace, ICL; Finnish Landrace, FIN; Swedish Landrace, SWE; Danish Landrace, DNK; Dutch Landrace, NLD; French breeds (Fosseé, FSS; Alpine, ALP_FR; Saanen, SAA_FR), Swiss breeds (Alpine, ALP_CH; Saanen, SAA_CH; Toggenburg, TOG); Spanish breeds (Bermeya, BEY; Malaguena, MLG); Italian breeds (Girgentana, GGT; Ciociara Grigia, CCG; Jonica, JON; Alpine, ALP_IT; Saanen, SAA_IT).

Figure 4

Effective population size and past demography. (A) Historical trends of Ne across generation in the past using the NW-EU target dataset. The x-axis indicates the generations, and the Ne estimates are indicated in the y-axis. (B) LD decay curves generated plotting the r2 (Linkage disequilibrium value) values against loci distance. Breeds acronyms according to Supplementary Table S8: Norwegian landraces (SKO, SEL, NRW); Irish landraces (ARR, BLB, OIG); Icelandic Landrace, ICL; Finnish Landrace, FIN; Swedish Landrace, SWE; Danish Landrace, DNK.

Figure 5

Runs Of Homozygosity analysis (ROH). Total proportion of the genome covered by homozygous segments. Each dot represents one individual and each breed coloured according to the colour code illustrated in the legend. Breeds acronyms according to Supplementary Table S8: Norwegian landraces (SKO, SEL, NRW); Irish landraces (ARR, BLB, OIG); Icelandic Landrace, ICL; Finnish Landrace, FIN; Swedish Landrace, SWE; Danish Landrace, DNK.

Figure 6

Inferred unrooted tree from a neighbour-joining analysis of Reynolds’ distance data. The Neighbor-Net graph displays the reticulate relationship among the analysed populations. Each terminal branch represents one breed. Breeds are labelled according to breed grouping in Fig. 1. Breeds acronyms according with Supplementary Table S8: Norwegian landraces (SKO, SEL, NRW); Irish landraces (ARR, BLB, OIG); Icelandic Landrace, ICL; Finnish Landrace, FIN; Swedish Landrace, SWE; Danish Landrace, DNK; Dutch Landrace, NLD; French breeds (Fosseé, FSS; Alpine, ALP_FR; Saanen, SAA_FR), Swiss breeds (Alpine, ALP_CH; Saanen, SAA_CH; Toggenburg, TOG); Spanish breeds (Bermeya, BEY; Malaguena, MLG); Italian breeds (Girgentana, GGT; Ciociara Grigia, CCG; Jonica, JON; Alpine, ALP_IT; Saanen, SAA_IT).

Figure 7

(A,B) Comparison between the two analyses carried out with the TreeMix program on the two datasets. To the left of the figure, (A) refers to the WHOLE dataset for m = 15 and (B) to the NW-EU target dataset for m = 8. The migration edges indicate the direction of the gene flow between two breeds and the colour of the arrow is according to the weight of the migration event. Breeds acronyms according to Supplementary Table S8, and their colours correspond to those in Figs. 1 and 2; Norwegian landraces (SKO, SEL, NRW); Irish landraces (ARR, BLB, OIG); Icelandic Landrace, ICL; Finnish Landrace, FIN; Swedish Landrace, SWE; Danish Landrace, DNK; Dutch Landrace, NLD; French breeds (Fosseé, FSS; Alpine, ALP_FR; Saanen, SAA_FR), Swiss breeds (Alpine, ALP_CH; Saanen, SAA_CH; Toggenburg, TOG); Spanish breeds (Bermeya, BEY; Malaguena, MLG); Italian breeds (Girgentana, GGT; Ciociara Grigia, CCG; Jonica, JON; Alpine, ALP_IT; Saanen, SAA_IT).

Figure 8

Sampling location for the 24 breeds analysed in the WHOLE dataset. Highlighted in turquoise are the countries the breeds are sampled from. The number of individuals is indicated in brackets. Other details in Supplementary Table S8.