The application of demographic history modelling and inference to the study of divergence between species has become a cornerstone of speciation genomics. Speciation histories are usually reconstructed by analysing single populations from each species, assuming that the inferred population history represents the actual speciation history. However, this assumption may not be met when species diverge with gene flow, e.g., when secondary contact may be confined to specific geographic regions. Here, we tested whether divergence histories inferred from heterospecific populations may vary depending on their geographic locations, using the two wood ant species Formica polyctena and F. aquilonia. We performed whole-genome resequencing of 20 individuals sampled in multiple locations across the European ranges of both species. Then, we reconstructed the histories of distinct heterospecific population pairs using a coalescent-based approach. Our analyses always supported a scenario of divergence with gene flow, suggesting that divergence started in the Pleistocene (ca. 500 kya) and occurred with continuous asymmetrical gene flow from F. aquilonia to F. polyctena until a recent time, when migration became negligible (2-19 kya). However, we found support for contemporary gene flow in a sympatric pair from Finland, where the species hybridise, but no signature of recent bidirectional gene flow elsewhere. Overall, our results suggest that divergence histories reconstructed from a few individuals may be applicable at the species level. Nonetheless, the geographical context of populations chosen to represent their species should be taken into account, as it may affect estimates of migration rates between species when gene flow is spatially heterogeneous.

Whole-genome data of 10 Formica polyctena and 10 F. aquilonia wood ants.

Questions should be directed to the corresponding authors, Beatriz Portinha (bcportinha@gmail.com) and Dr. Pierre Nouhaud (pierr3.nouhaud@gmail.com)

This Dryad directory contains Variant Call Format (VCF) files, Site Frequency Spectra (SFS) and model input files (.tpl and .est) used for fastsimcoal2 analyses in the Portinha et al. manuscript to be published in Molecular Ecology.

In this directory, you will find the following folders:

- VCF_files: this folder contains two compressed VCF files, one of which includes all of the called sites (whole_genome.vcf.gz) and the other, which had SNPs in genic regions and within 10kbs of genes removed (non_genic.vcf.gz); this folder also includes a README.txt file with further details.

- VCF_script: this folder contains a bash script used to produce the VCF files present in the "VCF_files" folder; this folder also includes a README.txt file with further details.

- SFS_files: this folder contains the SFSs used for the demographic modelling analyses performed with fastsimcoal2, built from the VCF files provided in this directory; in this folder, you will find the datasets built for each population comparison, as well as a README.txt file with instructions on how to compute these SFSs.

- fastsimcoal2_input_files: this folder contains the input files (.tpl and .est) of all models tested with fastsimcoal2; you will also find a README.txt file in this folder with detailed explanations of the model name tags.

- fastsimcoal2_scripts: this folder contains the bash scripts we used to run each demographic model 100 times; you will also find a README.txt file in this folder with further details.