Sticky Genomes: Using NGS Evidence to Test Hybrid Speciation Hypotheses

Abstract

Hypotheses of hybrid origin are common. Here we use next generation sequencing to test a hybrid hypothesis for a non-model insect with a large genome. We compared a putative hybrid triploid stick insect species (Acanthoxyla geisovii) with its putative paternal diploid taxon (Clitarchus hookeri), a relationship that provides clear predictions for the relative genetic diversity within each genome. The parental taxon is expected to have comparatively low allelic diversity that is nested within the diversity of the hybrid daughter genome. The scale of genome sequencing required was conveniently achieved by extracting mRNA and sequencing cDNA to examine expressed allelic diversity. This allowed us to test hybrid-progenitor relationships among non-model organisms with large genomes and different ploidy levels. Examination of thousands of independent loci avoids potential problems produced by the silencing of parts of one or other of the parental genomes, a phenomenon sometimes associated with the process of stabilisation of a hybrid genome. Transcript assembles were assessed for evidence of paralogs and/or alternative splice variants before proceeding. Comparison of transcript assemblies was not an appropriate measure of genetic variability, but by mapping reads back to clusters derived from each species we determined levels of allelic diversity. We found greater cDNA sequence diversity among alleles in the putative hybrid species (Acanthoxyla geisovii) than the non-hybrid. The allelic diversity within the putative paternal species (Clitachus hookeri) nested within the hybrid-daughter genome, supports the current view of a hybrid-progenitor relationship for these stick insect species. Next generation sequencing technology provides opportunities for testing evolutionary hypotheses with non-model organisms, including, as here, genomes that are large due to polyploidy.

Fig 1. New Zealand stick insects illustrate hybrid speciation hypotheses that arise from evolutionary studies.

(A) Two lineages of stick insects have been sampled across their range in New Zealand and by contrasting maternal relationships from mitochondrial DNA sequences with bi-parental multicopy nuclear markers a role for hybridisation has been inferred. Diploid Clitarchus hookeri (orange squares) has both sexual and asexual populations. No males of any of the Acanthoxyla forms are known (purple squares diploid females, purple circles triploid females). (B) Hybrid species are the product of interspecific mating resulting in genomes that are a mix of the two parental species but are reproductively isolated from both these parent taxa. The resulting allelic diversity is illustrated and compared to the diversity expected within non-hybrids and autopolyploids. When short DNA sequence reads are mapped to parents, related and non-related species, allelic similarities can be used to infer origins.

Fig 2. Bioinformatic pipeline for analysis of cDNA from non-model organisms with large polyploid genomes to test hybrid origin hypotheses.

Fig 3. Next Generation sequencing results from mRNA of two New Zealand stick insects (A).

Length distributions of transcript assemblies produced from the cDNA sequence of two stick insects were similar. A log length frequency distribution plot used values rounded to 1 decimal place for the longest consensus sequence generated from each cluster. (B) Sequence divergence (measured by SNP density per nucleotide) observed when reads were mapped to ~2,600 loci (transcript assemblies). Loci without variation (SNP-free) were removed. The putative parental Clitarchus hookeri genome contains many loci with low allelic diversity. SNPs detected in less than 10% of the short reads were ignored but reads were included whether or not they passed the strand bias filter within VarScan. Only the longest assembled transcripts generated per cluster were included. (C) SNPs detectable on all transcript assemblies by BWA mapping to ~2,600 Acanthoxyla and Clitarchus transcript assemblies using VarScan with minimum variant frequency of 10% irrespective of strand filter results. The first violin of each color comprises all data, and the second excludes transcript assemblies with no sequence variation (SNP-free). Purple–Acanthoxyla reads mapped onto Acanthoxyla transcript assemblies; Pale green–Clitarchus reads mapped onto Acanthoxyla transcript assemblies; Pale purple–Acanthoxyla reads mapped onto Clitarchus transcript assemblies; Green–Clitarchus reads mapped onto Clitarchus transcript assemblies.

Fig 4. A similar level of genetic divergence within a hybrid stick insect genome as is observed between parental alleles when alleles are compared.

Heat maps of sequence similarity between short-read cDNA sequences mapped to transcript assemblies. (A) Acanthoxyla transcript assemblies with Clitarchus reads (upper right) and Acanthoxyla reads (lower left). (B) Clitarchus transcript assemblies with Clitarchus reads (upper right) and Acanthoxyla reads (lower left).