Landscape genomics and biased FST approaches reveal single nucleotide polymorphisms under selection in goat breeds of North-East Mediterranean

Abstract

Background: In this study we compare outlier loci detected using a FST based method with those identified by a recently described method based on spatial analysis (SAM). We tested a panel of single nucleotide polymorphisms (SNPs) previously genotyped in individuals of goat breeds of southern areas of the Mediterranean basin (Italy, Greece and Albania). We evaluate how the SAM method performs with SNPs, which are increasingly employed due to their high number, low cost and easy of scoring.

Results: The combined use of the two outlier detection approaches, never tested before using SNP polymorphisms, resulted in the identification of the same three loci involved in milk and meat quality data by using the two methods, while the FST based method identified 3 more loci as under selection sweep in the breeds examined.

Conclusion: Data appear congruent by using the two methods for FST values exceeding the 99% confidence limits. The methods of FST and SAM can independently detect signatures of selection and therefore can reduce the probability of finding false positives if employed together. The outlier loci identified in this study could indicate adaptive variation in the analysed species, characterized by a large range of climatic conditions in the rearing areas and by a history of intense trade, that implies plasticity in adapting to new environments.

Figure 1

Plot of F_ST against heterozygosity for the 27 SNPs analysed. Distribution of F_ST values as a function of the within-population heterozygosity (H_S) based on the 27 goat SNPs analysed. The envelope of values corresponds to neutral expectations (with F_ST = 0.078) in the infinite-allele model constructed according to the method of Beaumont and Nichols [21], with a confidence level set to 95%.

Figure 2

Plot of significant association models between alleles and environmental parameters. Histogram of the number of significant association models between genotypes and environmental parameters according to the 6 highest significance levels (Bonferroni correction included) in the analysis. From 1.01E-13 or a confidence level of 99.999999% only CSN1S1 and LIPE are significantly present. Names of environmental variables are listed in the text. Names of loci in the legend are followed by the genotype [25] associated to the models.