• Cross-breeding between species, or hybridization, in vertebrates has generally been considered a rare occurrence.
• Hybridization may be an important evolutionary mechanism for either homogenization (reversing initial divergence between species), speciation (from reproductive isolation of hybrid populations), or adaptation (transfer of adaptive alleles) (Abbott et al., 2013; Barrera-Guzmán, Aleixo, Shawkey, & Weir, 2018; Mallet, 2007).

• Unlike other ranid frog species with broad areas of potential intergradation (Shaffer et al 2004) RASI/RABO rarely occur sympatrically
  • breeding experiments by Zweifel (1955) found high incidences of embryological abnormalities—indicating a post-zygotic barrier between the species (low viability in fertilization).
  • However, these experiments only crossed female R. sierrae with male R. boylii from very different sites (Butte and Nevada County vs. Contra Costa County).

• Hybridization can promote gene flow (introgression) between species
• Important evolutionary mechanism for:
  • homogenization, adaptation, or speciation

Flexible and cost efficient, thousands of samples multiplexed at once (high throughput).

• Many amphibian species cryptic/lack data, Large genome size, assessment and monitoring hard to do
• restriction site-associated DNA sequencing, highly efficient genotyping across thousands of individual samples for targeted loci
• For frogs, can sample with non-invasive methods using mouth swabs, or tadpole tail clips
• (i.e., small pops. with limited connectivity, low diversity)
• RAPTURE/RADSeq is a powerful and effective method

RAD Tags = sequences downstream of restriction sites. In order for samples to be sequenced on an Illumina machine, after cutting fragments must be ligated to adapters that will bind to an Illumina flow cell (RADSeq uses modified Illumina adapters that enable the binding and amplification of restriction site fragments only)

• Digest with Sbf1
  • Add RAD adapters
  • Sonicate/shear
  • purify
  • specify Capture baits/loci of interest (better/more sequencing at locations you want to look at)
  • Different cutters at different frequencies in the genome, this is rougly 300 cut sites per 100,000 bp

PCR duplicates can be identified as fragments that are identical in insert length and sequence composition, because random shearing ensures that fragments at a given locus are unlikely to be of equal length unless they are duplicates (Fig. 1; Davey et al. 2011; Hohenlohe et al. 2013). In contrast, for methods without a random-shearing step, all fragments at a given locus are expected to be of equal length whether or not they are PCR duplicates. Pooled DNA was sheared in a Bioruptor NGS sonicator.

Bean Creek/Spanish Creek largely the only locations where both species have been documented together, however, additional samples from Slate Creek and a few other small tributaries may show both spp (Goldberg/Mallory's MS)

• Used NGSAdmix to assess population structure and individual ancestry from genome-wide SNPs
  • Admixture showed the same two unknown samples from the Feather basin were had approximately 50% ancestry from each species.
  • ancestry in the individuals designated as “unknown” in the field also showed very low levels of mixed ancestry between the species
  • very low or nearly non-existent levels of mixed-ancestry in the American and Yuba basins as compared to the Feather. However, introgression between spp appears asymmetric, with a greater proportion of ancestry from R. boylii occurring in the R. sierrae samples

• Our filtering process (see Methods) yielded 3,062 putative diagnostic SNPs that were homozygous for different alleles in R. sierrae or R. boylii samples and also had successfully called genotypes in the two hybrid individuals

  • We observed extremely high heterozygosity and very low homozygosity (6% genotyped as R. boylii, 4% R. sierrae, and 89% were heterozygous) (Figure 4). This level of heterozygosity is far greater than expected for F2 individuals

• In addition, both hybrid individuals had RASI mitochondrial DNA (Bedwell 2018), indicating the maternal ancestor was from a RASI individual and the paternal ancestor was RABO

  • non Hz SNPs may be expected (at low levels), due to low coverage sequencing data; genotyping from low coverage sequencing will cause a low frequency of erroneous homozygous calls, because only one of the two alleles is sampled, causing heterozygotes to be called as homozygotes

The best model based on maximum likelihood estimated the time since divergence between R. sierrae and R. boylii was 370,856 (370,041–371,670) generations.

Early Pleistocene began ~1.8 mya ("Great Ice Age").

• unlikely that there is currently the potential for major introgression between R. sierrae and R. boylii, particularly as hybridization initially may not be adaptive and is often selected against (Abbott et al., 2013; Streicher et al., 2014).

  • hybridization outcomes that fail to produce successful offspring (sterile F1 hybrids) may have a greater cost on the species with low numbers of effective breeders, affecting both locally adapted populations and negatively impacting the probability of population persistence in a given region

  • under future scenarios (e.g., warming climate, range contraction, population crashes) the loss of even several breeding individuals (via reproduction with R. boylii) may have a significant impact in declining populations