Fig 1.
Geography and genetics of colour pattern.
(A) Approximate ranges of the four subspecies of D. chrysippus, with the contact zone outlined. Sampling locations for each of the subspecies and the contact zone are indicated. Cartoon chromosomes show the genotypes of each subspecies at the A (white hindwing patch), B (brown background colour), and C (forewing tip) colour patterning loci, based on previous crosses [20]. Note the linkage of B and C, putatively forming a ‘BC supergene’ [19]. Two examples of heterozygotes that can be found in the contact zone are shown. Note that Cc heterozygotes can exhibit the transiens phenotype with white markings on the forewing, with approximately 50% penetrance. (B) Model showing how fusion of the BC autosome to the W chromosome has produced a neo-W (purple) in contact zone females (top left), while males have two autosomal copies of the BC chromosome (top right). Daughters inherit the neo-W, while sons inherit the other BC chromosome haplotype from their mother. The latter allele is then lost due to male-killing by Spiroplasma.
Fig 2.
Identification of the BC supergene on Chromosome 15.
(A) Pairwise genetic differentiation (FST), plotted in 100-kb sliding windows with a step size of 20 kb across all chromosomes. Three subspecies pairs with sample sizes ≥6 are shown (see legend in panel B), revealing strong overlap in patterns of differentiation. Locations of SNPs most strongly associated with the A, B, and C loci (Wald test, 99.99% quantile) are plotted above in grey, brown, and black, respectively. See S2 Fig for a more detailed plot. (B) Expanded plot of FST across a 10-Mb portion of Chromosome 15 (chr15). Note that the first approximately 6 Mb of the chromosome is not included in this plot due to complex structural variation (see main text). Scaffolds are indicated below the plot in alternating shades. The locations of our most likely candidate genes for B (yellow) and C (arrow) are indicated. (C) Allelic clustering on chr15 in six representative individuals: three homozygotes and three heterozygotes (see S4 Fig for all individuals, and see panel B for chromosome positions). Coloured blocks indicate 20-kb windows, in which sequence haplotypes could be assigned to one of three clusters based on pairwise genetic distances (see Methods for details). Windows in grey indicate insufficient relative divergence to be assigned to a cluster, and white indicates missing data. Alleles are named according to the form in which they occur. In heterozygotes, the name of the dominant allele is in bold. Data deposited in the Dryad repository [36].
Fig 3.
Recent sweep of a young neo-W.
(A) The number of high-frequency female-specific mutations (>20% in females and absent in males) in 100-kb sliding windows (20-kb step size). (B) Distance-based phylogenetic network for the distal colinear region of chr15 outside of the BC supergene reveals that most contact zone females carry the conserved neo-W haplotype. Cartoons show how the colinear region of chr15 is outside of the BC supergene but would still capture reduced divergence among individuals carrying a shared non-recombining neo-W. (C) Box plot comparing the density of heterozygous sites in 100-kb windows in the colinear region of chr15 between wild-type individuals from the contact zone (CZ) and those carrying the neo-W. Cartoon chromosomes above the plot match those shown in panel B. A relative lack of elevated heterozygosity in the neo-W lineage indicates a lack of divergence of the fused neo-W haplotype, consistent with the fusion being recent. (D) Box plot of nucleotide diversity (π) within each population for the same colinear region of chr15. On the far right, π is shown for the haploid neo-W haplotype specifically, based on partial sequences isolated from this haplotype (see Methods and S10 Fig for details). The near absence of genetic diversity implies a very rapid spread of the neo-W through the population. Data deposited in the Dryad repository [36].
Fig 4.
Matrilineal inheritance causes coupling between neo-W and Spiroplasma.
(A) Maximum likelihood phylogeny for the whole mitochondrial genome. Individuals are coloured according to population of origin (see Fig 1A), and those carrying the neo-W (‘W’) and Spiroplasma (‘S’) are indicated (including one cured individual in which Spiroplasma was eliminated). Females are indicated by circles and males by squares. (B) Maximum likelihood phylogenies for the neo-W haplotype and Spiroplasma genome isolated from infected females. Corresponding clades are shaded to indicate congruence. Note that two samples are excluded in panel B: the cured sample, which lacked Spiroplasma because of tetracycline treatment, and one infected female found to lack the neo-W. Whether the latter represents an ancestral state or secondary loss requires further investigation. In all trees, nodes supported by more than 70 of 100 bootstrap replicates are indicated by circles. Data deposited in the Dryad repository [36].