Resource specialization and host-associated genetic differentiation (HAD) are frequently invoked as an explanation for the high diversity of plant-feeding insects and other organisms with a parasitic lifestyle. While genetic studies have demonstrated numerous examples of HAD in insect herbivores, the general rarity of comparative studies means that we still lack an understanding of how deterministic HAD is, and whether patterns of host shifts can be predicted over evolutionary time scales. We applied genome-wide SNP data obtained through low-coverage genome resequencing to define species limits and to compare host-plant use in population samples of leaf- and bud-galling sawflies collected from seven shared willow (Salix) host species. To infer the repeatability of long-term cophylogenetic patterns, we also contrasted the phylogenies of the two galler groups with each other as well as with the phylogeny of their Salix hosts estimated based on RADseq data. We found clear evidence for host specialization and HAD in both of the focal galler groups, but also that leaf gallers are more specialized to single host species than are most bud gallers. In contrast to bud gallers, leaf gallers also exhibit statistically significant cophylogenetic signal with their Salix hosts. The observed discordant patterns of resource specialization and long-term host use in two related galler groups that have radiated in parallel across a shared resource base indicate a general lack of evolutionary repeatability and suggest that short- and long-term host use and ecological diversification in plant-feeding insects are dominated by stochasticity and/or lineage-specific effects.

Contained in this dataset are the three filtered VCF files, two tree files for the cophylogenetic analysis, an example script using the combined VCF file and the metadata.

• Bud_galler_Vs_own = Bud galler samples mapped to the Euura lappo reference genome followed by variant calling.
• Leaf_galler_Vs_own = Leaf galler samples mapped to the Eupontania aestiva reference genome followed by variant calling.
• Bud_and_Leaf_gallers_Vs_bud_reference = Bud and leaf galler samples mapped to the Euura lappo reference genome followed by variant calling.
• Salix_concatRADloci.phy = phylip file containing the concatenated Salix RAD loci
• Galler_raxml_tree = The galler raxml tree used for the cophylogenetic analysis.
• Salix_raxml_tree = Salix tree used for the cophylogenetic analysis.
• ALL_Galler_MT_Concat.phy = Phylip alignment of concatendated mitochondrial genes.
• Galler_mtDNA_data_for_Arlequin.arp = Mitochondrial gene data for use in Arlequin.
• Statistical-analysis-Clustering_and_cophylogeny = Script used to perform the DAPC and sNMF analysis. Additionally, contains the script for the PACo cophylogenetic analysis based on the Bud_and_Leaf_gallers_Vs_bud_reference.vcf file.
• Metadata.txt = A comma-delimited file with the metadata for the combined galler dataset.