Introduced populations of invasive organisms have to cope with novel environmental challenges, while having reduced genetic variation caused by founder effects. The mechanisms associated with this "genetic paradox of invasive species" has received considerable attention, yet few studies have examined the genomic architecture of invasive species. Populations of the heart node ant Cardiocondyla obscurior belong to two distinct lineages, a New World lineage so far only found in Latin America and a more globally distributed Old World lineage. In the present study, we use population genomic approaches to compare populations of the two lineages with apparent divergent invasive potential. We find that the strong genetic differentiation of the two lineages began at least 40,000 generations ago and that activity of transposable elements (TEs) has contributed significantly to the divergence of both lineages, possibly linked to the very unusual genomic distribution of TEs in this species. Furthermore, we show that introgression from the Old World lineage is a dominant source of genetic diversity in the New World lineage, despite the lineages' strong genetic differentiation. Our study uncovers mechanisms underlying novel genetic variation in introduced populations of C. obscurior that could contribute to the species' adaptive potential.

The harvester ant genus Pogonomyrmex is endemic to arid and semiarid habitats and deserts of North and South America. The California harvester ant Pogonomyrmex californicus is the most widely distributed Pogonomyrmex species in North America. Pogonomyrmex californicus colonies are usually monogynous, i.e. a colony has one queen. However, in a few populations in California, primary polygyny evolved, i.e. several queens cooperate in colony founding after their mating flights and continue to coexist in mature colonies. Here, we present a genome assembly and annotation of P. californicus. The size of the assembly is 241 Mb, which is in agreement with the previously estimated genome size. We were able to annotate 17,889 genes in total, including 15,688 protein-coding ones with BUSCO (Benchmarking Universal Single-Copy Orthologs) completeness at a 95% level. The presented P. californicus genome assembly will pave the way for investigations of the genomic underpinnings of social polymorphism in the number of queens, regulation of aggression, and the evolution of adaptations to dry habitats.

Evolution is change over time. Although neutral changes promoted by drift effects are most reliable for phylogenetic reconstructions, selection-relevant changes are of only limited use to reconstruct phylogenies. On the other hand, comparative analyses of neutral and selected changes of protein-coding DNA sequences (CDS) retrospectively tell us about episodic constrained, relaxed, and adaptive incidences. The ratio of sites with nonsynonymous (amino acid altering) versus synonymous (not altering) mutations directly measures selection pressure and can be analysed by using the Phylogenetic Analysis by Maximum Likelihood (PAML) software package. We developed a CDS extractor for compiling protein-coding sequences (CDS-extractor) and parallel PAML (paPAML) to simplify, amplify, and accelerate selection analyses via parallel processing, including detection of negatively selected sites. paPAML compiles results of site, branch-site, and branch models and detects site-specific negative selection with the output of a codon list labelling significance values. The tool simplifies selection analyses for casual and inexperienced users and accelerates computing speeds up to the number of allocated computer threads. We then applied paPAML to examine the evolutionary impact on a new GINS Complex Subunit 3 exon, and neutrophil-associated as well as lysin and apolipoprotein genes. Compared with codeml (PAML version 4.9j) and HyPhy (HyPhy FEL version 2.5.26), all paPAML test runs performed with 10 computing threads led to identical selection pressure results, whereas the total selection analysis via paPAML, including all model comparisons, was about 3 to 5 times faster than the longest running codeml model and about 7 to 15 times faster than the entire processing time of these codeml runs.