Phylogeny and patterns of diversity of goat mtDNA haplogroup A revealed by resequencing complete mitogenomes

Abstract

We sequenced to near completion the entire mtDNA of 28 Sardinian goats, selected to represent the widest possible diversity of the most widespread mitochondrial evolutionary lineage, haplogroup (Hg) A. These specimens were reporters of the diversity in the island but also elsewhere, as inferred from their affiliation to each of 11 clades defined by D-loop variation. Two reference sequences completed the dataset. Overall, 206 variations were found in the full set of 30 sequences, of which 23 were protein-coding non-synonymous single nucleotide substitutions. Many polymorphic sites within Hg A were informative for the reconstruction of its internal phylogeny. Bayesian and network clustering revealed a general similarity over the entire molecule of sequences previously assigned to the same D-loop clade, indicating evolutionarily meaningful lineages. Two major sister groupings emerged within Hg A, which parallel distinct geographical distributions of D-loop clades in extant stocks. The pattern of variation in protein-coding genes revealed an overwhelming role of purifying selection, with the quota of surviving variants approaching neutrality. However, a simple model of relaxation of selection for the bulk of variants here reported should be rejected. Non-synonymous diversity of Hg's A, B and C denoted that a proportion of variants not greater than that allowed in the wild was given the opportunity to spread into domesticated stocks. Our results also confirmed that a remarkable proportion of pre-existing Hg A diversity became incorporated into domestic stocks. Our results confirm clade A11 as a well differentiated and ancient lineage peculiar of Sardinia.

Figure 1. Median joining network of 30 sequences showing the occurrence of 23 non-synonymous substitutions.

Each circle represents a sequence. Branch length is proportional to the number of mutations. Note that two A11 and two uncharacterized sequences are lumped into single (larger) nodes (see Materials and Methods). The haplogroup or D-loop clade (for Hg A) affiliation of clusters of sequences is shown in large empty characters. Aminoacid replacements are shown beside each branch in their polarized form. The Hg A portion of the network is magnified while the original network is in the inset. A large lozenge indicates the likely position of the root and was added in the drawing to partition three variants peculiar of Hg C. Their polarization was obtained by comparison with C. ibex.

Figure 2. Scatterplot of number of variants recorded as a function of length of each of the 13 protein-coding mtDNA genes.

Filled dots: variants scored among all 30 sequences considered. Empty dots: variants scored among the 28 Hg A sequences. Panel A: all positions. Panel B: only 3^rd codon positions. Panel C: only non-synonymous variants (note that the three black dots at 0 variants overlap and obscure empty dots). coxIII is indicated by an arrow.