Evolutionary Biology

Elevated heterozygosity clued atypical reproduction: evidence consistent with hybridogenesis mediating insular occurrence of a narrowly endemic spiny frog

Yuchi Zheng author has email address
Bin Wang
Yue Cao
Xiaomao Zeng
Yun Xia author has email address

Chengdu Institute of Biology, Chinese Academy of Sciences, Chengdu, China
BeiGene (Beijing) Co., Ltd., Beijing, China

https://doi.org/10.7554/eLife.109069.1

Open access
Copyright information

Figures and data

Several types of different signatures of reproductive systems where hybrids are F₁ types.

Sampling sites.
Symbols indicate genotype compositions of samples. Site number colors correspond to clusters BB1 and BR1 (black), BB2 and BR2 (purple), BB3 and BR3 (blue), BB4 and BR4 (white), and RR (red) in Fig. 2.

Principal coordinate plot (A), heterozygosity pattern (B), ADMIXTURE ancestry proportions (C), and neighbor-joining tree (D) based on the 45,523-SNP main dataset.
Italic numbers on the neighborjoining tree indicate sampling sites in Fig. 1. Numbers beside nodes are bootstrap support values ≥70, and circles on nodes indicate support values of 100. The sample 28B, a putative hybrid of different Quasipaa boulengeri lineages, is marked by an asterisk. In the neighbor-joining tree, the two individuals showing Q. robertingeri genotypes found far from the known species range are highlighted in purple. BB: Q. boulengeri; RR: Q. robertingeri; BR: putative Q. boulengeri × Q. robertingeri hybridogen; Qe: four Q. exilispinosa individuals.

Dataset summaries.
Unique hit: loci with unique BLAST hits on homologous chromosome-level scaffolds of two genomes of the related Quasipaa spinosa and Nanorana parkeri; BB: Q. boulengeri; RR: Q. robertingeri; BR: putative Q. boulengeri × Q. robertingeri hybridogen.

Heterozygous SNP numbers of twelve individuals in subsets of the 45,523-SNP main dataset.
The 25,046-site subset contained no missing data and 12,183 sites each corresponding to a locus with unique BLAST hits on homologous scaffolds of two published genomes of Quasipaa spinosa and Nanorana parkeri. ID: individual ID; BB: Q. boulengeri; RR: Q. robertingeri; BR: putative BB x RR hybridogen. SNPs being heterozygous for BR while showing one allele in BB and the other in RR were considered (H) or excluded (H′).

Statistical parsimony network for haplotypes of a single exemplary locus, CLocus_761519 of 144 bp, from individuals of Quasipaa boulengeri, Q. robertingeri, and putative Q. boulengeri × Q. robertingeri hybridogen genotypes.
This locus was heterozygous in putative hybridogens and had unique BLAST hits on the first scaffolds of published Q. spinosa and Nanorana parkeri genomes. Lines represent single substitutions. Filled circles are haplotypes not encountered. The other circles’ sizes are proportional to haplotype frequency. Putative hybridogens’ sample IDs are shown. Samples of Q. robertingeri exhibited haplotypes marked by asterisks.

Maximum likelihood population tree of Quasipaa robertingeri with three migration edges (arrows) inferred using TreeMix (A) and RAxML mitochondrial gene tree (B).
The TreeMix tree was based on the 8,535-SNP nuclear dataset of 100% completeness for ingroups. The number below an arrow is migration weight that represents the fraction of ancestry derived from the migration edge. s.e.: average standard error of the sample covariance matrix of allele frequencies. The RAxML tree was based on a 10,568-site dataset of 52.7% completeness and six partitions. Italic numbers beside nodes are bootstrap support values ≥70, and circles on nodes indicate support values of 100 (closed) and 98–99 (open). The characters BB, RR, or BR following the sample ID indicate Q. boulengeri, Q. robertingeri, or putative hybridogen nuclear genotypes, respectively. Lineage names within the Qb clade correspond to cluster names in Fig. 2. Outgroups are not shown for the RAxML tree.

ADMIXTURE cross-validation plot and ancestry proportions for 55 Quasipaa boulengeri individuals based on the 45,523-SNP dataset.
*: Sample 28B.

Statistical parsimony networks for haplotypes of 26 loci.
Corresponding scaffolds of published genomes of Quasipaa spinosa (Qs) and Nanorana parkeri (Np) are shown. The 0a and 1a in the sequence name indicate allele 0 and allele 1, respectively. The last character in a sequence name, b, r, or H, indicates Q. boulengeri, Q. robertingeri, or putative hybridogenetic individuals, respectively. Between them, the sample ID is given. Lines represent single substitutions, and numbers beside lines are site numbers. Circles are haplotypes not encountered.

Statistical parsimony networks for haplotypes of 26 loci.
Corresponding scaffolds of published genomes of Quasipaa spinosa (Qs) and Nanorana parkeri (Np) are shown. The 0a and 1a in the sequence name indicate allele 0 and allele 1, respectively. The last character in a sequence name, b, r, or H, indicates Q. boulengeri, Q. robertingeri, or putative hybridogenetic individuals, respectively. Between them, the sample ID is given. Lines represent single substitutions, and numbers beside lines are site numbers. Circles are haplotypes not encountered.

Neighbor-joining tree based on the SNP identity-by-state distance matrix for individuals.
Numbers at nodes are bootstrap support values ≥70, and circles indicate support values of 100. BB: Qu-asipaa boulengeri; RR: Q. robertingeri; BR: putative hybridogenetic hybrids between Q. boulengeri and Q. robertingeri.

Sign up for email alerts