3.1 NextRAD and mitochondrial data
The nextRAD dataset consisted, on average, of 1.6 million single-end reads per individual, with an average read length of 123 bp. The final assembled genome contained 250,856 fragments with a total length of 240.6 Mb. Considering a set consisting of 173 samples, we identified 14,446 loci. However, we had to remove 32 individuals from further analyses because of high numbers of missing data (<10% missing genotypes). From the remaining 141 individuals, we recovered a total of 5,383 loci. After filtering loci that were present in at least 90% of the individuals, we ended up with a final dataset containing 1,707 SNPs (average coverage 78X) used for further analyses. Numbers of loci removed in each filtering step are detailed in Table S2. Since the 10% of missing data criterium used for filtering the SNPs dataset is the strictest threshold, we evaluated other filtering thresholds that yielded essentially the same results in population genetic structure and species delimitation analyses.
The mtDNA dataset consisted of 233 sequences, 143 collected in the native range and 90 in the non-native range. Additionally, we includedCOI sequences of five individuals from California, United States on Cactaceae, but SNPs data at this location could not be generated. The length of the amplified COI fragment was 413 bp, with 371 conserved (89.83%) and 42 (10.16%) variable sites, distributed in 27 haplotypes. The mtDNA sequences are available in GenBank under accession numbers MT138921 - MT138931.