One mother for two species via obligate cross-species cloning in ants

Although clonality is the most straightforward mode of reproduction, most animal species take a more complex route⁶. In sexual species, for instance, reproduction requires the interaction of males and females, which typically means that two different morphs have to be produced⁷. Such complexity is further amplified in some species, in which females produce distinct morphs depending on seasonal conditions, population density or social caste^8,9,10,11. Even in these extreme cases, a seemingly universal constraint persists: regardless of their morphological variation, phenotypes produced by a female invariably belong to the same species. Here, we report that this rule has been transgressed by Messor ibericus ants, with females producing individuals from two different species.

Previous studies on Messor genus ants have reported conflicting results, suggesting widespread hybridizations between species that rarely co-occur in Europe^12,13. Here, a combination of field work, population genomic analyses and laboratory experiments provide the resolution of this paradox: females of one of the species (M. ibericus) clone males of the other (Messor structor), as they need their sperm to produce the worker caste. We discuss the evolutionary history of this natural case of cross-species cloning, which suggests a domestication-like process for exploiting another species’ gametes.

Population genetic analyses revealed that M. ibericus queens are unable to produce workers without mating with males of another species. To reach this conclusion, we analysed genome-wide data in 390 individuals (Supplementary Table 1) from five European species of the Messor genus (phylogenetic tree in Fig. 1a and Extended Data Figs. 1 and 2). In ants, workers and queens of the same species are diploid individuals expected to be genetically similar¹⁴. Our data showed that this is not the case in one out of the five species analysed. In M. ibericus, all worker genomes (n = 164) featured a 15 times higher heterozygosity than their queens or queens and workers of the four other species (n = 127; average of 0.797 versus 0.047 on 43,084 polymorphic sites, two-sided Wilcoxon rank-sum test P < 2.2 × 10⁻¹⁶; Fig. 1a). Such high heterozygosity levels suggest that M. ibericus workers are hybrids. We confirmed this hypothesis by conducting an analysis specifically designed to detect first-generation hybrids¹⁵, which identified all M. ibericus workers as such (Methods and Supplementary Table 1). With the exception of one Messor ponticus worker, queens and individuals of the other four species were identified as non-hybrids (Supplementary Table 1).

a, Proportion of heterozygous positions on the total number of polymorphic sites (SNPs, n = 43,084) for queens and workers of M. ibericus (n = 220), M. ponticus (n = 12), Messor mcarthuri (n = 6), Messor muticus (n = 8) and M. structor (n = 45). Species individuals are arranged vertically according to their phylogenetic relationships (tree was built from one representative individual of each species; Extended Data Fig. 1). Each hybrid worker from M. ibericus colonies (n = 164) displays a pie chart representing its respective population ancestry proportion estimated from the fastStructure software¹⁶, with blue and red representing, respectively, M. ibericus (maternal) and M. structor (paternal) genome proportions. Average hybrid worker heterozygosity (n = 164) is significantly higher than average heterozygosity of M. structor queens or queens and workers of the four other species (n = 127; average of 0.797 versus 0.047, two-sided Wilcoxon rank-sum test, P < 2.2 × 10⁻¹⁶). b, Map representing the distribution of sequenced hybrid workers (n = 164). The distribution areas of each parental species have been estimated from our sampling and reports from the literature^13,23. Hybrid workers localized in areas where both parental species co-occur are highlighted by a picture representing an M. ibericus queen (blue) with an M. structor male (red). Hybrid workers localized in areas without the paternal species are highlighted with the same picture but with a question mark instead of the father. SNP, single nucleotide polymorphism.

Full size image

To identify the maternal origin of hybrid workers, we conducted a phylogenetic analysis on the maternally inherited mitochondrial genome. The resulting tree suggests an M. ibericus maternal ancestry, as all hybrid workers share the mitochondrial genome of M. ibericus sexual individuals (Extended Data Fig. 2). To identify the paternal species, we conducted a phylogenetic analysis of nuclear DNA after separating the maternal and paternal alleles of the hybrid genomes (Methods). The resulting phylogenetic tree showed that hybrid workers have an M. structor paternal ancestry, as all paternal alleles (n = 164) formed a well-supported clade with individuals of this species (Extended Data Fig. 3). Finally, a population structure analysis¹⁶ on 5,856 genes (44,191 variants) revealed that workers in M. ibericus colonies had virtually equal population ancestry proportions from M. ibericus and M. structor (averaging 0.49 and 0.51, respectively; Fig. 1 and Supplementary Table 1), which confirms further that they are first-generation hybrids.

These results imply that M. ibericus depends on hybridization for worker production, as already observed in cases of sperm parasitism¹⁷, in which queens exploit sperm from another lineage or species to produce workers^{12,18,19,20,21}. Here, M. ibericus queens strictly depend on males of M. structor, which is a well-differentiated, non-sister species (Fig. 1a). This finding is particularly surprising because these two species do not share the exact same distribution area^22,23. This paradox is clearly illustrated by hybrid workers being found across Southern Europe in spite of the total absence of their paternal species (Fig. 1b; 69 Mediterranean populations with confirmed M. ibericus but no M. structor colonies found). As even more compelling evidence, first-generation hybrid workers from the Italian island of Sicily are found more than a thousand kilometres away from the closest known occurrence of their paternal species. This raises the question of how queens can hybridize in such an isolated area (Fig. 1b). To solve this conundrum, we examined males from M. ibericus colonies more closely.

Morphological and molecular analyses showed that M. ibericus queens lay the M. structor males they require for worker production. By sampling 132 males from 26 M. ibericus colonies, we observed a sharp morphological dimorphism: 44% of sampled males displayed a dense pilosity (Fig. 2a), whereas the other 56% were nearly hairless (Fig. 2b). By conducting phylogenetic analyses including 62 hairy versus 24 hairless male nuclear genomes, we showed that the two morphs perfectly correspond to two different species (Extended Data Fig. 2). Whereas all hairy males group with M. ibericus, all hairless ones group with M. structor, which are two non-sister species that we estimated to have split more than 5 million years ago (Ma) (Methods, Fig. 2c and Extended Data Figs. 1 and 4). Multiple lines of evidence point to the production of males of both species by M. ibericus queens.

M. ibericus queens lay males belonging to different species that differ morphologically (symbolized by male symbols in blue and red for M. ibericus an M. structor, respectively) and genetically. M. ibericus and M. structor males produce sperm for producing either new queens or workers, respectively. All share the same mitochondria (corresponding to the M. ibericus mitochondria, depicted here in blue; Extended Data Fig. 2). a, M. ibericus male photo (hairy). b, M. structor male photo (hairless). c, Phylogenetic tree of 223 non-hybrid individuals. Based on 5,656 nuclear genes (2,780,573 bp) and simplified from Extended Data Fig. 1. All represented nodes have maximal bootstrap support (100). Triangle widths are relative to the number of individuals. Branch lengths are relative to divergence time estimated from Fig. 1 and Extended Data Fig. 4 (see Methods for details). Scale bars, 1 mm. Credit: The top picture of an ant is adapted with permission from a photo from Flickr (https://www.flickr.com) taken by M. Kukla. bp, base pairs.

Full size image

First, M. structor males share the same mitochondria as their M. ibericus nestmates, pointing to common M. ibericus mothers for the whole colony (n = 24; Fig. 2, Extended Data Fig. 2 and Supplementary Table 1). This nuclear–mitochondrial genome mismatch is unique to males found in M. ibericus colonies, as it has not been observed in any other M. structor individual when found in their own species colonies (n = 53; Extended Data Fig. 2 and Supplementary Table 1).

Second, genotyping 286 eggs or larvae from 5 M. ibericus laboratory colonies showed that 11.5% exclusively contained M. structor nuclear genome (Supplementary Note 1, Supplementary Table 2 and Supplementary Figs. 1 and 2). To confirm that such M. structor eggs were laid by M. ibericus queens and not workers, we isolated 16 queens and genotyped their newly produced eggs after 24 h. Again, we found that 9% of these eggs exclusively contained M. structor DNA (Supplementary Note 1, Supplementary Fig. 3 and Supplementary Table 3), which was not the case for broods produced by workers (see Supplementary Note 2 for details).

Third, beyond genetic evidence, direct observations confirmed the emergence of adult males of both species from a single queen colony. We monitored a laboratory colony headed by a single M. ibericus queen for 18 months, checking broods weekly. Among seven eggs that developed into reproductive adults, two were identified as M. structor (hairless) males, and three as M. ibericus (hairy) males. Genomic analyses confirmed their morphological identification, with their whole nuclear genome matching solely either M. ibericus or M. structor (individuals ORT3M1 to ORT5M5; Extended Data Fig. 1 and Supplementary Table 1). Despite those M. structor births, we confirmed that the whole genome of the mother queen solely matches M. ibericus (ORT3Q1; Extended Data Fig. 1 and Supplementary Table 1). Other adult male emergences of both species (one of each) have been observed in another laboratory colony after 19 months of brood monitoring (Extended Data Fig. 5 for a picture of live individuals).

Whereas male Hymenoptera typically inherit their nuclear genome from their mother through unfertilized eggs²⁴, our results demonstrate that M. ibericus queens can produce males without transmitting their nuclear genome. This observation points to androgenesis (that is, male clonality), whereby a male provides the sole source of nuclear genetic material for the embryo²⁵. Embryos devoid of maternal DNA have been observed in other groups, with the fertilization of non-nucleate ovules²⁶ or the elimination of the maternal genome after fertilization²⁷. In ants, both should spontaneously lead to males genetically identical to the sperm, as males are typically produced from haploid embryos through haplodiploidy²⁴. At the intraspecific level, several cases of ants cloning males from their own species’ sperm have been observed^28,29,30,31. Here, our results imply that this phenomenon has crossed species barriers, with male cloning from allospecific sperm stored in the spermatheca. Consistent with this explanation, M. ibericus queens are polyandrous and mate with both species’ males, as we retrieved sperm of both M. ibericus and M. structor when sequencing the spermatheca content of a queen that gave birth to both species (ORT3QS1 in Supplementary Table 1 and Extended Data Fig. 3; see also the BAN1QS spermatheca, which again contains spermatozoa of both species).

The combination of obligate hybridization for worker production (Fig. 1) and cross-species cloning (Fig. 2) points to the following scenario: M. ibericus queens first stored sperm from another species, then began to clone males from this sperm. This pathway is consistent with the widespread observation of facultative or obligate sperm parasitism¹⁷, a well-described phenomenon in which queens use sperm from a co-occurring lineage or species to produce their workers^{15,18,19,20,21,28,29,30,32}. This strategy may have been selected either to benefit from potential worker hybrid vigour¹⁷ or to prevent queen-only production due to the fixation of a caste-biasing genotype^18,32. In the ancestral state of this scenario, M. ibericus exploits sperm from co-occurring M. structor colonies (Fig. 3a), as has been observed in other Messor species^12,33. In the derived state, M. ibericus queens directly produce the species they depend on, resulting in a clonal lineage of M. structor males they maintain in their colonies (Fig. 3b).

a, Ancestral state of the M. ibericus reproductive system; n = 20 colonies deduced to correspond to this state have been sampled (Supplementary Table 1). b, Derived state of the M. ibericus reproductive system; n = 130 colonies deduced to correspond to this state have been sampled (Supplementary Table 1). Note that M. structor males have an M. ibericus mitochondrial genome, which is indicated with a red chromosome and a blue mitochondrion. c, Phylogenetic tree simplified from Extended Data Fig. 1 (as in Fig. 2c). Links to a and b are based on Extended Data Fig. 3, in which hybrid workers have been separated into paternal and maternal genomes. M. structor ‘clonal’ lineage stands for a clade composed of males from M. ibericus nests and the paternal genome of their worker daughters (derived state). M. structor ‘wild-type’ lineage stands for a clade composed of all castes from normal M. structor nests and the paternal genome of some hybrid workers found in M. ibericus co-occurring nests (ancestral state). d, Photo of M. structor males from M. structor colonies (hairy). e, Photo of M. structor males from M. ibericus colonies (hairless). Scale bars, 1 mm.

Full size image

To confirm the advent of such a clonal lineage of males, we examined the two primary subdivisions of the M. structor nuclear phylogeny (Fig. 3c). As expected, one subdivision corresponds to a clonal lineage, consisting exclusively of nearly identical M. structor males, all found within M. ibericus colonies and carrying M. ibericus mitochondria (n = 24; Fig. 3b,c and Extended Data Fig. 1). By contrast, we retrieved a ‘wild-type’ lineage, which grouped all M. structor castes when found in their own species’ colonies (n = 53; Fig. 3a,c and Extended Data Fig. 1). To further confirm our scenario, we tracked the exact parental origin of each hybrid worker (n = 164; Methods). Consistent with occurrences of both the ancestral and derived states (Fig. 3a,b), we found that the paternal genome can belong to either the ‘wild-type’ or ‘clonal’ lineage (Fig. 3c and Extended Data Fig. 3). Although most hybrid workers were fathered by clonal males (144 out of 164), the fact that some (20 out of 164) were fathered by wild-type males confirms the recent occurrence of our ancestral state hypothesis (Fig. 3a). Consistent with our scenario, ancestral state cases were restricted to a limited geographical area where both species still co-occur (for example, eastern France; Fig. 1b, Extended Data Fig. 3 and Supplementary Table 1). By contrast, derived state cases were widespread across Europe, as maintaining a clonal lineage of males is likely to have allowed rapid expansion of M. ibericus beyond the natural range of M. structor (for example, Mediterranean Europe; Fig. 1b). This pathway seems analogous to domestication³⁴, as M. ibericus co-opted M. structor males into its life cycle, maintaining them as a clonal lineage rather than exploiting them from the wild.

Supporting this view, the clonal lineage exhibited extremely low genetic diversity with high genetic load compared with the wild-type lineage (average synonymous nucleotide diversity π_s of 0.00027 versus 0.0014, average ratio of non-synonymous to synonymous nucleotide diversity π_n/π_s of 0.43 versus 0.21; Supplementary Table 4). This pattern is typically observed in clonal species^35,36, after rapid range expansions^37,38 or in domesticated lineages maintained by humans^39,40. Interestingly, clonal males also differ morphologically: in a similar way that they differ from their M. ibericus nestmates (Fig. 2), they also seemed hairless compared with their wild-type counterparts (Fig. 3d,e). More generally, this clonal morph differs on several other criteria, standing out as the most divergent compared with the wild-type and M. ibericus males (Supplementary Note 3 for details and Supplementary Figs. 4–6), akin to the morphological divergence of domesticated species compared with their wild relatives⁴¹. Such a stark morphological difference does not necessarily result from a selection process. Instead, this difference may have been randomly retained from ancestral polymorphism, or may be due to incompatibilities between the nuclear and mitochondrial genomes of the two species (Fig. 3b) or plasticity due to different rearing conditions when born and kept within M. ibericus nests.

To assess whether clonal males can escape their ‘domesticated’ situation by mating with their wild female counterparts, we conducted a detailed analysis on 45 M. structor genomes to detect potential hybrids (Supplementary Note 4). Our findings confirmed that such events are at present non-existent or extremely rare, as we did not identify any hybrid between clonal and wild-type lineages (Supplementary Fig. 7). Similarly to typical cases of domestication, this raises the question of whether recent genetic isolation from wild populations warrants a different species classification⁴². Further analyses therefore support the idea that clonal males still belong to M. structor, as phylogenetic conflict (Supplementary Fig. 8a), population genetic structure (Supplementary Fig. 8b), species delimitation inferences (Supplementary Fig. 8c,d), low F_st fixation index (Supplementary Fig. 9), low genetic divergence (Supplementary Fig. 10a) and high historical gene flow (Supplementary Fig. 10b) are all consistent to support clonal and wild-type lineages as part of the same species (see Supplementary Note 4 for details). Taken together, these results further support the idea that clonal males should be characterized as a domesticated lineage of M. structor. All in all, this means that M. ibericus females interact with up to three males that are morphologically and genetically distinct (M. ibericus, ‘domesticated’ M. structor and ‘wild’ M. structor males; Extended Data Fig. 6), laying two of them (Fig. 2) and mating with the three (Fig. 3).

To our knowledge, females needing to clone members of another species have not previously been observed. Although cross-species cloning has been reported in hermaphrodite conifers and clams²⁵, these are instances of male parasites occasionally using other species’ eggs. In such cases, producing males of another species is not in the interest of females, as they are incidental victims of parasitism. This contrasts with the system reported here, for which producing another species’ male is not an accident, but a female life cycle requirement. We suggest defining such females as xenoparous, meaning they need to produce individuals of another species as part of their life cycle. This shows the evolution of xenoparity (xeno-, meaning ‘foreign, strange, different’, and -parity, meaning ‘produce, bring forth, give birth’), which is the need to propagate another species’ genome by means of its own eggs.

Transition towards xenoparity seems to result from sexual evolution along a parasitism–mutualism continuum. Similar to several other harvester ant species, M. ibericus first transitioned into obligate sperm parasitism^12,17 (Fig. 3a), a situation in which they lost the ability to produce workers by themselves due to epistatic incompatibilities^18,43 or selfish caste-biasing genotypes³². Although not the most straightforward path towards xenoparity, this situation might have evolved towards reciprocal sperm parasitism, a form of sperm mutualism seen in other harvester ants in which two lineages depend on each other’s sperm for worker production^12,18,21. Whether it be in the case of simple or reciprocal parasitism, dependence on males from another species is sub-optimal for queens, as it requires them to mate with two different male partners and restricts their colonies to the geographic range of their host. By producing the required species’ males in their own colonies (Fig. 3b), M. ibericus has gained a clear advantage, as it maintains obligate hybridization while minimizing the inherent constraints (Extended Data Fig. 7). Investigating the male cloning mechanism will help to determine whether this developmental innovation is analogous to male parasitism²⁵ or unique to the M. ibericus reproductive system.

While trapped in the life cycle of a species exploiting their sperm, clonal males propagate their genome through the reproductive efforts and parental care of M. ibericus. In a sense, clonal males can be viewed as a perfected form of male parasites, as they are essential to their female hosts but reproduce at the expense of their ova. By depending on each other’s gametes, both species have intertwined their life cycles, evolving from sexual parasitism³ to sexual co-dependency (Extended Data Fig. 8). In spite of this, females seem to control the terms of the relationship, as our data on brood genotyping suggest that they impose the timing of male eggs’ development and maturity (Supplementary Note 1). Such a situation seems akin to a sexual domestication, as M. ibericus controls the reproduction of a species it first exploited from the wild.

Although matching all criteria of domestication³⁴, the relationship we describe is both more intimate and integrated than the most remarkable examples known so far, from human-driven domestication⁴⁰ to lichen symbiosis⁴⁴. Contrary to such examples, both partners are obligate mating partners, as the domesticating species is directly cloning the domesticated one by means of its own egg cytoplasms. Such replication of an alien genome within one’s own cytoplasm echoes the endosymbiotic domestication of organelles (for example, mitochondria) within eukaryotic cells^45,46. Clonal males may thus be regarded as organelles at the superorganism level^47,48, resulting from the integration of this alien genome into a colony that directly replicates it. This leads to colonies producing the greatest diversity of individuals, differing in terms of sexes, castes and species, each with a dedicated role within a cohesive reproductive unit. Besides revealing a reproductive mode under which one species needs to clone another, such a ‘two-species superorganism’ challenges the usual boundaries of individuality. Major evolutionary transition in individuality occurs when distinct entities evolve into an integrated, higher-level unit^49,50,51. As two species have become sexually interdependent in such an integrated entity, evolution towards xenoparity exemplifies how such transitions can occur through a sexual domestication process.

Sampling

To better understand hybridization patterns between M. ibericus and M. structor, we gradually sampled individuals of both species, along with their respective closest relative species across Europe (M. ponticus, M. muticus and M. mcarthuri)¹³. In total, we sequenced 377 individuals from 125 different populations (280 M. ibericus, 8 M. ponticus, 6 M. mcarthuri, 75 M. structor and 8 M. muticus; Supplementary Table 1). From these individuals, we sequenced one reference genome of M. ibericus with long-read sequencing, 327 genomes with short-read sequencing and 51 transcriptomes. The previously published transcriptomes of seven M. ibericus, one M. structor and five M. ponticus¹² were added to the final dataset. Short-read sequencing of a Messor wasmanni worker was also added and subsequently used as an outgroup. We also dissected the spermatheca of two M. ibericus queens and sequenced their contents by short-read sequencing. We kept and monitored 65 colonies of M. ibericus in artificial nests. Colonies were kept in a room at 25 °C and 40% humidity and were fed with grass seeds.

Reference genome assembly

To obtain a reference genome for our population genomic dataset, high-molecular-weight DNA extraction of an M. ibericus queen from the Passa population was sequenced using PacBio long-read sequencing⁵². Illumina short-read sequencing from the same individual was also produced to polish the genome assembly. The genome assembly was performed using the Wtdbg2 assembler v.2.5 (ref. ⁵³). The PacBio Sequel reads were processed assuming a genome size of around 300 megabases and using preset2 settings which are more appropriate for genome sizes lower than 1 gigabase. The initial assembly was then polished using POLCA⁵⁴, a tool incorporated in the MaSuRCA assembler (v.3.4.1), leveraging its ability to correct discrepancies using Illumina short reads. The process was enhanced further by using NextPolish v.1.3.1 (ref. ⁵⁵) for error correction, using both the short and long reads for correction with default settings. To improve the assembly’s contiguity, we used RagTag v.1.0.2 with the scaffold command⁵⁶, using a Messor capitatus genome (GAGA-0413_Messor_capitatus.fasta)⁵⁷ as a reference. Following the scaffolding, we applied another round of polishing with POLCA before proceeding to fill the gaps in the assembly with TGS-GapCloser (v.1.1.1)⁵⁸. This program uses raw long reads and Racon⁵⁹ to polish the filled gaps. Finally, this was followed up with a final round of polishing with POLCA and NextPolish to ensure the accuracy of the assembly.

The resulting assembly was evaluated using QUAST (v.5.0)⁶⁰, with a total assembly length of 310,325,892 bp divided in 618 contigs, GC% of 36.82 and N50 of 12,028,351 bp. We then ran a BUSCO (v.4.0.5)⁶¹ analysis to evaluate the completeness of the genome. We used default parameters, database hymenoptera_odb10 and lineage dataset Camponotus floridanus. We retrieved complete sequences of 97.7% of the 5,991 Hymenoptera single-copy orthologues.

DNA and RNA extraction and library preparation

To sequence a population genomic dataset for each species, we performed Illumina short-read sequencing of either whole genomes or transcriptomes on 390 samples. DNA extractions were performed using Macherey-Nagel NucleoMag Tissue kit with an extra RNase step. Library preparation for whole-genome sequencing was performed using customized Illumina protocols^62,63.

For RNA extraction, we killed individuals in liquid nitrogen and conserved them at −80 °C. RNA extractions were performed using the following protocol. First, the sample was homogenized with ceramic beads in 1 ml of TRIzol solution (3 × 30 min at 6.5 °C). The homogenized samples were incubated in TRIzol for 5 min at room temperature. Next, 200 µl of chloroform was added and the mixture was vortexed vigorously for 15 s, followed by incubation for 5 min at room temperature. The mixture was centrifuged for 25 min at 12,000 rpm (4 °C). The upper aqueous layer (~500 µl) was transferred to a new tube. Next, 1 µl of glycogen blue (RNAse-free, Invitrogen, 15 mg ml⁻¹, catalogue no. AM9516) was added. The mixture was then vortexed and incubated overnight at −20 °C. The sample was centrifuged for 30 min at 12,000 rpm (4 °C). After centrifugation, the supernatant was discarded and 1 ml of 80% EtOH was added. The mixture was vortexed briefly and centrifuged for 5 min at 12,000 rpm (4 °C). All supernatants were removed and the pellet was allowed to air dry for 15–20 min at room temperature. Library preparations for transcriptome sequencing were made using the Roche KAPA mRNA HyperPrep Kit (catalogue no. 08098115702).

Read mapping

To map the short-read sequencing data onto the reference genome, Illumina reads of the 390 samples (326 whole-genome sequencing, 64 RNA sequencing) were trimmed and filtered using fastp v.0.23.2 (ref. ⁶⁴), requiring a minimum quality score of 20 (-q 20), discarding reads with more than 70% unqualified bases (-u 70) or 40 unknown bases (-n 40) and retaining only reads longer than 40 bases after trimming (-l 40). We then mapped all the filtered reads to the reference genome of M. ibericus using BWA-MEM2 (v.2.2.1)⁶⁵ with default parameters. Unmapped reads and secondary alignments were discarded using SAMtools (v.1.15.1)⁶⁶ with the view command and option -F 260.

Coding sequence search

To ensure reliable population genetic analyses, we focused on coding sequences of highly conserved single-copy orthologue genes. For this, we produced a sequence alignment map file (bam file) for single-copy orthologue genes of the nuclear genome using the following approach. From bam file alignments, we isolated the reads that overlapped coding regions of the reference genome BUSCO genes, using samtools view together with option -L and a bed file obtained from the BUSCO output. Retrieved reads were then realigned to the 5,856 BUSCO genes of the reference genome using BWA-MEM2 with default parameters. Most of the following analyses were conducted on these 5,856 genes.

To retrieve mitochondrial genes, we first identified and isolated reads of mitochondrial origin using seven mitochondrial genomes of the Myrmicinae sub-family (Myrmica scabrinodis, Cardiocondyla obscurior, Solenopsis invicta, Solenopsis geminata, Solenopsis richteri, Atta laevigata and Wasmannia auropunctata) as reference and mirabait (v.4.9)⁶⁷ with options -D 50 -k 31 -n 2. Corresponding mitochondrial reads were then assembled using megahit v.1.2.9 (ref. ⁶⁸) with kmer size every 10 bp from 31 to 101 bp (--k-min 31 --k-max 101 --k-step 10). We analysed assemblies using mitofinder (v.1.4.2)⁶⁹ with default options to identify and retrieve mitochondrial genes.

Variant calling

We called variants (SNPs) using GATK (v.4.3)⁷⁰. First, we obtained precalling variant files by using gatk HaplotypeCaller with default parameters and option -ERC BP_RESOLUTION on each of 390 individual realignment files (bam files). All individuals were then pooled to produce SNP calls for the whole dataset using gatk MergeVcfs with default options. We filtered SNPs using vcftools (v.0.1.16)⁷¹, keeping only variants with a genotype quality of more than 10 (--minGQ 10). The resulting vcf file is available from the Zenodo repository (https://zenodo.org/records/11506545 (ref. ⁷²)).

Heterozygosity and hybrid detection

To detect hybrids among sequenced individuals, we computed SNP heterozygosity. For this, we filtered further the SNPs using vcftools (--remove-indels --maf 0.05 --max-missing 0.8) before computing the number of heterozygous sites for each individual using a Python home-made script (available from the Zenodo repository via https://zenodo.org/records/11506545 (ref. ⁷²)). Number of heterozygous positions per individual was then divided by the total number of polymorphic sites (n = 43,084; Supplementary Table 1). We expected to observe higher heterozygosity values in hybrid individuals. Our analysis clearly confirmed this pattern for M. ibericus workers compared with other individuals (two-sided Wilcoxon rank-sum test, P < 2.2 × 10⁻¹⁶; Fig. 1a). A single M. ponticus worker showed similar heterozygosity values (0.58) and has been retained with M. ibericus workers to be tested for hybrid status below.

To further confirm the hybrid status of M. ibericus workers, we used a Bayesian approach designed to specifically detect first-generation hybrids¹⁵. To reduce computing time and avoid discrepancies between whole-genome and RNA sequencing data, we restricted the analysis to 833 highly expressed housekeeping genes, that is, keeping only universal genes common to metazoa found in the OrthoDB⁷³ dataset metazoa_odb10. The approach estimated the γ parameter, which is a measure of the heterozygosity acquired during the divergence of two separated populations. We expected the γ hybrid index to be higher in hybrid genomes (164 M. ibericus workers and 1 M. ponticus) and greater than zero in hybrid genomes (all other individuals). We found a clear non-overlapping distribution of higher γ hybrid index in M. ibericus workers and one hybrid M. ponticus worker (average of 0.00186, range from 0.0002 to 0.0024) versus all other individuals which have very close to zero γ values (average of 1.86 × 10⁻⁶, range from 6.598 × 10⁻⁷ to 6.461 × 10⁻⁶), with a highly significant difference (two-sided Wilcoxon rank-sum test, P < 2.2 × 10⁻¹⁶).

Nuclear phylogeny inference

Because hybrid individuals contain genomes from different species, they can bias the inference of phylogenetic relationships. To obtain clear relationships among Messor species, we thus first excluded hybrid individuals and built a phylogeny on the basis of nuclear genes of the 223 non-hybrid individuals. Individual variant calling files (vcf files) were treated separately from this point on. Indels were removed from each individual vcf file using vcftools (--remove-indels). Consensus sequences for 5,856 single-copy orthologue genes (BUSCO genes) were extracted from vcf files using bcftools (v.1.15.1) consensus⁷⁴, with heterozygous position treated as missing data. Positions with depth coverage of less than 3 were replaced by a gap and genes with more than 50% gaps were excluded. Alignments for each of 5,856 single-copy genes were built separately using MAFFT⁷⁵. Outgroups were added using the macse alignTwoProfiles command⁷⁶: M. wasmanni (this study), Messor barbarus⁷⁷, Aphaenogaster floridana⁷⁸ and Acromyrmex echinatior⁷⁹. We concatenated the 5,856 gene alignments then trimmed the obtained supermatrix using trimal by removing sites with more than 5% missing data, resulting in an alignment of 2,780,573 sites. We inferred a phylogenetic tree using IQ-TREE (v.2.07)⁸⁰ with a GTR + I + F + G4 model (general time reversible model with proportion of invariant sites, empirical base frequencies and a gamma distribution with four rate categories) and 1,000 ultrafast bootstraps (-bb 1000). The topology of the resulting tree is available in Extended Data Fig. 1. By removing hybrid individuals, we expected clear parental relationships among species. As expected, all nodes defining the species relationships exhibited a maximal bootstrap support of 100.

Mitochondrial phylogeny inference

To identify the maternal species of hybrid individuals, we built a mitochondrial phylogeny including our 390 individuals (Supplementary Table 1), and we first aligned separately the 15 mitochondrial genes using MAFFT (v.7.490)⁷⁵. We added data of the M. wasmanni genome as an outgroup. We then concatenated the alignments before cleaning the resulting supermatrix using trimal (v.1.4)⁸¹ with the automated1 option. The resulting 2,585-site supermatrix was then used to infer a phylogenetic tree with IQ-TREE (v.2.07)⁸⁰, using the MFP option for automatically selecting the substitution model with 1,000 ultrafast bootstraps (-bb 1000). We expected that all individuals sampled from M. ibericus colonies—including M. ibericus males and females, hybrid workers and M. structor males—were laid by M. ibericus queens. Given that the mitochondrial genome is maternally inherited, we expected these individuals to group within the same clade. As expected, all individuals from M. ibericus colonies grouped in the same clade, regardless of their hybrid status or nuclear genome species origin (Extended Data Fig. 2).

Divergence time estimation

We estimated the divergence times of our species with MCMCtree from the PAML package (v.4.10.7)⁸². For this, we first built a phylogenetic tree with one individual per species, as recommended by the MCMCtree manual. Representative individuals were selected on the basis of their coverage (Supplementary Table 1): Y15452-1 for M. muticus, Y16370-1 for M. mcarthuri and Y14753-1 for M. ponticus, Y15268-1 for M. structor (wild-type lineage), SH19-04 for M. structor (clonal lineage) and the long-read reference genome for M. ibericus. The same outgroups as for the previous analysis were kept. We trimmed the supermatrix by removing all sites with at least one gap, resulting in an alignment of 6,089,069 sites. We inferred a phylogenetic tree using IQ-TREE (v.2.07)⁸⁰ with a GTR + I + F + G4 model and 1,000 ultrafast bootstraps (-bb 1000). The resulting tree had similar species relationships as the previous one. All nodes had a maximal bootstrap support of 100. We used this tree to constrain the topology of the divergence time estimation. This analysis was run using the same 6,089,069 supermatrix by using MCMCtree rapid approximate likelihood computation⁸³. Based on a time-calibrated phylogeny of the Stenammini tribe⁸⁴, we constrained the root node with soft bounds from −71.2 to −101.8 Ma, corresponding to the lower and upper bound of the 95% highest posterior density of that study’s main analysis. Similarly, we set soft bounds on the common ancestor of M. barbarus and M. wasmanni from −6.4 to −12.5 Ma and the common ancestor of A. floridana and all Messor from −12.7 to −21.1 Ma. We ran two runs of the analysis with the correlated rates model, HKY85 substitution model for 600 million generations. We confirmed convergence and sufficient effective sample sizes (≫200) for all parameters using Tracer v.1.7.2 (ref. ⁸⁵). The resulting tree with confidence intervals of estimated divergence time is available in Extended Data Fig. 4.

We used the average divergence times of this tree’s nodes as secondary constraints for the mitochondrial and nuclear trees with all individuals, using the least squares method in IQ-TREE (v.2.12)⁸⁰ with the –date option to obtain an ultrametric tree as used for Figs. 2 and 3 and Extended Data Figs. 1–3.

Phasing maternal and paternal alleles of hybrid individuals

To identify the parental species of each hybrid individual, we developed a custom phasing approach for separating paternal and maternal alleles of each hybrid individual. SNP calls were first isolated after filtering out indels (vcftools –remove-indels). Previous results indicated that M. ibericus is the maternal species of all hybrid workers, as hybrids are found only in M. ibericus queen colonies (see also Fig. 1 and Extended Data Fig. 1). Consequently, alleles of hybrid individuals that do not match M. ibericus are expected to belong to the paternal species. Given the exceptionally low genetic diversity of M. ibericus (π_s of 0.00045; Supplementary Table 4), the risk of confusing intraspecific polymorphism with paternal alleles is minimized. We exploited this specificity by writing a Python script comparing each SNP of each hybrid worker with variants of a reference maternal genome (M. ibericus queen genome with the highest coverage, SH19-06). The script parses a vcf file position by position, and applies the following approach:

When a site is heterozygous in the hybrid worker (for example, A/G) and one of the nucleotides matches the maternal reference at homozygous state (for example, A/A), this nucleotide is assigned as maternal (for example, A) and the other is assigned as paternal (for example, G). In all other cases (heterozygous position in maternal reference, no matching allele between hybrid and maternal references), an N is assigned to both the paternal and maternal alleles. N is also assigned when the coverage of a site is below 3 reads in the focal hybrid or the reference maternal genome. Once the paternal and maternal nucleotides have been discriminated at all sites, maternal and paternal sequences are reconstructed using bcftools (v.1.15.1) consensus⁷⁴ (default options).

The same approach was used for: (1) spermatheca content of M. ibericus queen; (2) males laid by orphaned M. ibericus/structor workers; and (3) the only hybrid M. ponticus worker (using the M. ponticus queen RDNIPQ).

Phylogenetic analysis of paternal and maternal alleles of hybrids

To identify the paternal species of hybrid individuals, we built a phylogenetic tree including non-hybrid genomes and paternal + maternal haplomes of hybrid genomes inferred from the previous analysis. The maternal and paternal sequences of the 5,856 orthologue genes were aligned with the corresponding data from non-hybrid individuals and then concatenated. We applied the same filtering as for the non-hybrid individual supermatrix (no more than 5% missing data). We inferred a phylogenetic tree from the resulting supermatrix (1,089,038 bp, 559 haplomes) using IQ-TREE (GTR + I + F + G4 model, 1,000 ultrafast bootstraps). We deduced maternal and paternal species of hybrid individuals from the phylogenetic placement of their corresponding maternal and paternal haplomes (Extended Data Fig. 3). Given their hybrid nature, we expected that the paternal haplomes of M. ibericus workers would group with M. structor. As expected, all M. ibericus workers (n = 164) were identified as hybrids, with M. ibericus mothers and M. structor fathers, supported by a maximal bootstrap value of 100 for each haplome grouping with its corresponding species of origin (Extended Data Fig. 3).

Additionally, we expected the spermatheca of M. ibericus queens to contain sperm from both M. ibericus and M. structor males. As expected, the sequences from the spermatheca content grouped with either the M. ibericus clade or the M. structor clade with maximal bootstrap support value of 100 (Extended Data Fig. 3). The hybrid M. ponticus worker was identified as a hybrid between M. ponticus mother and M. structor father with maximal bootstrap support value of 100 (Extended Data Fig. 3).

Population structure analysis

To estimate the population ancestry proportions of hybrids, we selected M. ibericus and M. structor individuals from variant files (same variant filtering as for SNP heterozygosity computing) and then produced a bed file using PLINK⁸⁶ (v.1.90b6.21) before using fastStructure¹⁶ (v.1.0) with k = 2. Because our hybrid detection approach is designed to detect first-generation hybrids, we expected each hybrid worker to exhibit population ancestry close to 50% from each parental species. As expected, we obtained average proportions of 0.49 and 0.51 for M. ibericus and M. structor ancestry, respectively. The resulting population ancestry values are detailed in Supplementary Table 1 and visualized in Fig. 1.

Synonymous and non-synonymous polymorphism

To estimate population size and the associated genetic load of each species or lineage, we computed the synonymous polymorphism, π_s, and the ratio of non-synonymous over synonymous polymorphism, π_n/π_s (ref. ⁸⁷). The π_n/π_s ratio estimations being very sensitive to SNP call errors, we selected only the genomes and transcriptomes with more than 15× coverage for these analyses. We used the program dNdSpiNpiS (v.1.0) available from this link (https://kimura.univ-montp2.fr/PopPhyl/index.php?section=tools), with options -allow_internal_bc=1 -compute_distances=1 -gapN_site=4 -gapN_seq=0.2. Results are available in Supplementary Table 4 with their respective confidence intervals. As expected in the case of low effective population size and high genetic load for a clonal lineage, we retrieved very low values of π_s (0.00027, confidence interval 0.00021–0.00033) and high values of π_n/π_s (0.427, confidence interval 0.378–0.485) in the M. structor clonal male lineage.

PCR tests for species identification

We developed a PCR test designed to quickly identify M. ibericus (queens or males), M. structor (all castes) and hybrid individuals (M. ibericus workers). On the basis of our genomes, we designed a combination of two primers (namely, CL0001: CCACTGTGGCGTACCTACC; and CL0002: CTACACGTACACGCGACAC) to amplify different microsatellite fragment lengths depending on the species: a 247-bp fragment for M. ibericus, a 467-bp fragment for M. structor and both fragment lengths in hybrid individuals (M. ibericus workers). DNA extractions were conducted using the Phire Tissue Direct kit (Fisher), following the two-step protocol with various tissue types crunched with a pillar: eggs, partial larva, finely cut adult leg or finely cut adult wing. Amplifications were carried out in a 10-μl reaction volume comprising 5 μl of Phire MasterMix 2X (ref K0171), 3 μl of water, 0.5 μl of each primer (10 μM) and 1 μl of DNA extract. The PCR conditions were 5 min at 98 °C, followed by 35 cycles with 10 s at 98 °C, 10 s at 66 °C and 15 s at 72 °C, and a final extension of 1 min at 72 °C. PCR products were run on a 1% agarose gel, resulting in a pattern of short, long or short/long fragments depending on the species/caste combination. We validated the approach on adult individuals of various caste and colonies with previously sequenced genomes: M. ibericus queens (n = 10), M. ibericus/structor hybrid workers (n = 14), M. structor clonal males (n = 9), M. structor queens (n = 3) and M. structor workers (n = 8). As expected, short fragments were observed for M. ibericus queens, long fragments for M. structor (queens, workers, clonal males) and long/short fragments for M. ibericus/structor hybrid workers. Three different test runs confirmed the reliability of the approach (Supplementary Fig. 1).

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Raw reads of genetic data are deposited at the NCBI under project ID PRJNA1145159, with SRA IDs and all data supporting the results of the study indicated for each sample in Supplementary Table 1. Reference genomes, genetic variation data and phylogenetic analyses used for producing the results of the study are available at Zenodo (https://doi.org/10.5281/zenodo.11506545)⁷². We also used the following datasets from the orthoDB database (https://www.orthodb.org/): hymenoptera_odb10 (https://busco-data.ezlab.org/v5/data/lineages/hymenoptera_odb10.2024-01-08.tar.gz) and metazoa_odb10 (https://busco-data.ezlab.org/v5/data/lineages/metazoa_odb10.2024-01-08.tar.gz).

Scripts used for producing the results of the study are available at Zenodo (https://doi.org/10.5281/zenodo.11506545)⁷².

1. Darwin, C. On the Origin of Species by Means of Natural Selection, or the Preservation of Favoured Races in the Struggle for Life (J. Murray, 1859).

2. Mayr, E. Systematics and the Origin of Species, from the Viewpoint of a Zoologist (Columbia Univ. Press, 1942).

3. Lehtonen, J., Schmidt, D. J., Heubel, K. & Kokko, H. Evolutionary and ecological implications of sexual parasitism. Trends Ecol. Evol. 28, 297–306 (2013).

   Article  PubMed  Google Scholar 

4. Loi, P. et al. Genetic rescue of an endangered mammal by cross-species nuclear transfer using post-mortem somatic cells. Nat. Biotechnol. 19, 962–964 (2001).

   Article  CAS  PubMed  Google Scholar 

5. Bolton, R. L. et al. Resurrecting biodiversity: advanced assisted reproductive technologies and biobanking. Reprod. Fertil. 3, R121–R146 (2022).

   Article  PubMed  PubMed Central  Google Scholar 

6. Avise, J. Clonality: The Genetics, Ecology, and Evolution of Sexual Abstinence in Vertebrate Animals (Oxford Univ. Press, 2008).

7. Bachtrog, D. et al. Sex determination: why so many ways of doing it? PLoS Biol. 12, e1001899 (2014).

   Article  PubMed  PubMed Central  Google Scholar 

8. Simpson, S. J., Sword, G. A. & Lo, N. Polyphenism in insects. Curr. Biol. 22, 352 (2012).

   Article  CAS  Google Scholar 

9. Schwander, T., Lo, N., Beekman, M., Oldroyd, B. P. & Keller, L. Nature versus nurture in social insect caste differentiation. Trends Ecol. Evol. 25, 275–282 (2010).

   Article  PubMed  Google Scholar 

10. Shapiro, A. M. in Evolutionary Biology (eds Hecht, M. K., Steere, W. C. & Wallace, B.) 259–333 (Springer US, 1976).

11. Applebaum, S. W. & Heifetz, Y. Density-dependent physiological phase in insects. Annu. Rev. Entomol. 44, 317–341 (1999).

    Article  CAS  PubMed  Google Scholar 

12. Romiguier, J., Fournier, A., Yek, S. H. & Keller, L. Convergent evolution of social hybridogenesis in Messor harvester ants. Mol. Ecol. 26, 1108–1117 (2017).

    Article  PubMed  Google Scholar 

13. Steiner, F. M. et al. Turning one into five: integrative taxonomy uncovers complex evolution of cryptic species in the harvester ant Messor ‘structor’. Mol. Phylogenet. Evol. 127, 387–404 (2018).

    Article  PubMed  Google Scholar 

14. Smith, C. R., Toth, A. L., Suarez, A. V. & Robinson, G. E. Genetic and genomic analyses of the division of labour in insect societies. Nat. Rev. Genet. 9, 735–748 (2008).

    Article  CAS  PubMed  Google Scholar 

15. Weyna, A., Bourouina, L., Galtier, N. & Romiguier, J. Detection of F1 hybrids from single-genome data reveals frequent hybridization in hymenoptera and particularly ants. Mol. Biol. Evol. 39, msac071 (2022).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

16. Raj, A., Stephens, M. & Pritchard, J. K. fastSTRUCTURE: variational inference of population structure in large SNP data sets. Genetics 197, 573–589 (2014).

    Article  PubMed  PubMed Central  Google Scholar 

17. Umphrey, G. J. Sperm parasitism in ants: selection for interspecific mating and hybridization. Ecology 87, 2148–2159 (2006).

    Article  PubMed  Google Scholar 

18. Helms Cahan, S. & Keller, L. Complex hybrid origin of genetic caste determination in harvester ants. Nature 424, 306–309 (2003).

    Article  ADS  CAS  PubMed  Google Scholar 

19. Darras, H. et al. Obligate chimerism in male yellow crazy ants. Science 380, 55–58 (2023).

    Article  ADS  CAS  PubMed  Google Scholar 

20. Lacy, K. D., Shoemaker, D. & Ross, K. G. Joint evolution of asexuality and queen number in an ant. Curr. Biol. 29, 1394–1400 (2019).

    Article  CAS  PubMed  Google Scholar 

21. Kuhn, A., Darras, H., Paknia, O. & Aron, S. Repeated evolution of queen parthenogenesis and social hybridogenesis in Cataglyphis desert ants. Mol. Ecol. 29, 549–564 (2020).

    Article  PubMed  Google Scholar 

22. Seifert, B. The Ants of Central and North Europe (lutra Verlags- und Vertriebsgesellschaft, 2018).

23. Lebas, C. Influence des activités humaines sur la répartition des fourmis du genre Messor dans les Pyrénées-Orientales (Hymenoptera: Formicidae: Myrmicinae). Osmia https://doi.org/10.47446/osmia9.9 (2021).

24. Heimpel, G. E. & de Boer, J. G. Sex determination in the hymenoptera. Annu. Rev. Entomol. 53, 209–230 (2008).

    Article  CAS  PubMed  Google Scholar 

25. Schwander, T. & Oldroyd, B. P. Androgenesis: where males hijack eggs to clone themselves. Philos. Trans. R. Soc. Lond. B 371, 20150534 (2016).

    Article  Google Scholar 

26. Pichot, C., Borrut, A. & El Maâtaoui, M. Unexpected DNA content in the endosperm of Cupressus dupreziana A. Camus seeds and its implications in the reproductive process. Sex. Plant Reprod. 11, 148–152 (1998).

    Article  CAS  Google Scholar 

27. Komaru, A., Ookubo, K. & Kiyomoto, M. All meiotic chromosomes and both centrosomes at spindle pole in the zygotes discarded as two polar bodies in clam Corbicula leana: unusual polar body formation observed by antitubulin immunofluorescence. Dev. Genes Evol. 210, 263–269 (2000).

    Article  CAS  PubMed  Google Scholar 

28. Fournier, D. et al. Clonal reproduction by males and females in the little fire ant. Nature 435, 1230–1234 (2005).

    Article  ADS  CAS  PubMed  Google Scholar 

29. Ohkawara, K., Nakayama, M., Satoh, A., Trindl, A. & Heinze, J. Clonal reproduction and genetic caste differences in a queen-polymorphic ant, Vollenhovia emeryi. Biol. Lett. 2, 359–363 (2006).

    Article  PubMed  PubMed Central  Google Scholar 

30. Pearcy, M., Goodisman, M. A. D. & Keller, L. Sib mating without inbreeding in the longhorn crazy ant. Proc. Biol. Sci. 278, 2677–2681 (2011).

    PubMed  PubMed Central  Google Scholar 

31. Okita, I. & Tsuchida, K. Clonal reproduction with androgenesis and somatic recombination: the case of the ant Cardiocondyla kagutsuchi. Naturwissenschaften 103, 22 (2016).

    Article  PubMed  Google Scholar 

32. Weyna, A., Romiguier, J. & Mullon, C. Hybridization enables the fixation of selfish queen genotypes in eusocial colonies. Evol. Lett. 5, 582–594 (2021).

    Article  PubMed  PubMed Central  Google Scholar 

33. Norman, V., Darras, H., Tranter, C., Aron, S. & Hughes, W. O. H. Cryptic lineages hybridize for worker production in the harvester ant Messor barbarus. Biol. Lett. 12, 20160542 (2016).

    Article  PubMed  PubMed Central  Google Scholar 

34. Purugganan, M. D. What is domestication? Trends Ecol. Evol. 37, 663–671 (2022).

    Article  PubMed  Google Scholar 

35. Jaron, K. S. et al. Convergent consequences of parthenogenesis on stick insect genomes. Sci. Adv. 8, eabg3842 (2022).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

36. Glémin, S., François, C. M. & Galtier, N. Genome evolution in outcrossing vs. selfing vs. asexual species. Methods Mol. Biol. 1910, 331–369 (2019).

    Article  PubMed  Google Scholar 

37. Excoffier, L., Foll, M. & Petit, R. J. Genetic consequences of range expansions. Annu. Rev. Ecol. Evol. Syst. 40, 481–501 (2009).

    Article  Google Scholar 

38. de Pedro, M. et al. Demography, genetic diversity and expansion load in the colonizing species Leontodon longirostris (Asteraceae) throughout its native range. Mol. Ecol. 30, 1190–1205 (2021).

    Article  PubMed  Google Scholar 

39. Glémin, S. & Bataillon, T. A comparative view of the evolution of grasses under domestication. New Phytol. 183, 273–290 (2009).

    Article  PubMed  Google Scholar 

40. Frantz, L. A. F., Bradley, D. G., Larson, G. & Orlando, L. Animal domestication in the era of ancient genomics. Nat. Rev. Genet. 21, 449–460 (2020).

    Article  CAS  PubMed  Google Scholar 

41. Darwin, C. The Variation of Animals and Plants under Domestication (John Murray, 1868).

42. Gentry, A., Clutton-Brock, J. & Groves, C. P. The naming of wild animal species and their domestic derivatives. J. Archaeol. Sci. 31, 645–651 (2004).

    Article  Google Scholar 

43. Anderson, K. E. et al. Distribution and evolution of genetic caste determination in Pogonomyrmex seed-harvester ants. Ecology 87, 2171–2184 (2006).

    Article  PubMed  Google Scholar 

44. Spribille, T., Resl, P., Stanton, D. E. & Tagirdzhanova, G. Evolutionary biology of lichen symbioses. New Phytol. 234, 1566–1582 (2022).

    Article  PubMed  Google Scholar 

45. Martin, W. F., Garg, S. & Zimorski, V. Endosymbiotic theories for eukaryote origin. Philos. Trans. R. Soc. Lond. B 370, 20140330 (2015).

    Article  Google Scholar 

46. Martijn, J., Vosseberg, J., Guy, L., Offre, P. & Ettema, T. J. G. Deep mitochondrial origin outside the sampled alphaproteobacteria. Nature 557, 101–105 (2018).

    Article  ADS  CAS  PubMed  Google Scholar 

47. Wheeler, W. M. The ant‐colony as an organism. J. Morphol. 22, 307–325 (1911).

    Article  Google Scholar 

48. Boomsma, J. J. & Gawne, R. Superorganismality and caste differentiation as points of no return: how the major evolutionary transitions were lost in translation. Biol. Rev. Camb. Philos. Soc. 93, 28–54 (2018).

    Article  PubMed  Google Scholar 

49. Maynard-Smith, J. & Szathmary, E. The Major Transitions in Evolution, Vol. 49 (Oxford Univ. Press, 1997).

50. West, S. A., Fisher, R. M., Gardner, A. & Kiers, E. T. Major evolutionary transitions in individuality. Proc. Natl Acad. Sci. USA 112, 10112–10119 (2015).

    Article  ADS  CAS  PubMed  PubMed Central  Google Scholar 

51. Rafiqi, A. M., Rajakumar, A. & Abouheif, E. Origin and elaboration of a major evolutionary transition in individuality. Nature 585, 239–244 (2020).

    Article  ADS  CAS  PubMed  Google Scholar 

52. Rhoads, A. & Au, K. F. PacBio sequencing and its applications. Genom. Proteom. Bioinform. 13, 278–289 (2015).

    Article  Google Scholar 

53. Ruan, J. & Li, H. Fast and accurate long-read assembly with wtdbg2. Nat. Methods 17, 155–158 (2019).

    Article  PubMed  PubMed Central  Google Scholar 

54. Zimin, A. V. & Salzberg, S. L. The genome polishing tool POLCA makes fast and accurate corrections in genome assemblies. PLoS Comput. Biol. 16, e1007981 (2020).

    Article  ADS  CAS  PubMed  PubMed Central  Google Scholar 

55. Hu, J., Fan, J., Sun, Z. & Liu, S. NextPolish: a fast and efficient genome polishing tool for long-read assembly. Bioinformatics 36, 2253–2255 (2020).

    Article  CAS  PubMed  Google Scholar 

56. Alonge, M. et al. RaGOO: fast and accurate reference-guided scaffolding of draft genomes. Genome Biol. 20, 224 (2019).

    Article  PubMed  PubMed Central  Google Scholar 

57. Boomsma, J. J. et al. The Global Ant Genomics Alliance (GAGA). Myrmecol. News 25, 61–66 (2017).

    Google Scholar 

58. Xu, M. et al. TGS-GapCloser: a fast and accurate gap closer for large genomes with low coverage of error-prone long reads. Gigascience 9, giaa094 (2020).

    Article  PubMed  PubMed Central  Google Scholar 

59. Vaser, R., Sović, I., Nagarajan, N. & Šikić, M. Fast and accurate de novo genome assembly from long uncorrected reads. Genome Res. 27, 737–746 (2017).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

60. Gurevich, A., Saveliev, V., Vyahhi, N. & Tesler, G. QUAST: quality assessment tool for genome assemblies. Bioinformatics 29, 1072–1075 (2013).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

61. Seppey, M., Manni, M. & Zdobnov, E. M. BUSCO: assessing genome assembly and annotation completeness. Methods Mol. Biol. 1962, 227–245 (2019).

    Article  CAS  PubMed  Google Scholar 

62. Tilak, M.-K. et al. A cost-effective straightforward protocol for shotgun Illumina libraries designed to assemble complete mitogenomes from non-model species. Conserv. Genet. Resour. 7, 37–40 (2015).

    Article  Google Scholar 

63. Meyer, M. & Kircher, M. Illumina sequencing library preparation for highly multiplexed target capture and sequencing. Cold Spring Harb. Protoc. 2010, db.prot5448 (2010).

    Article  Google Scholar 

64. Chen, S., Zhou, Y., Chen, Y. & Gu, J. fastp: an ultra-fast all-in-one FASTQ preprocessor. Bioinformatics 34, i884–i890 (2018).

    Article  PubMed  PubMed Central  Google Scholar 

65. Vasimuddin, M., Misra, S., Li, H. & Aluru, S. Efficient architecture-aware acceleration of BWA-MEM for multicore systems. In 2019 IEEE International Parallel and Distributed Processing Symposium (IPDPS) 314–324 (IEEE, 2019).

66. Li, H. et al. The Sequence Alignment/Map format and SAMtools. Bioinformatics 25, 2078–2079 (2009).

    Article  PubMed  PubMed Central  Google Scholar 

67. Chevreux, B. et al. Using the miraEST assembler for reliable and automated mRNA transcript assembly and SNP detection in sequenced ESTs. Genome Res. 14, 1147–1159 (2004).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

68. Li, D. et al. MEGAHIT v1.0: a fast and scalable metagenome assembler driven by advanced methodologies and community practices. Methods 102, 3–11 (2016).

    Article  CAS  PubMed  Google Scholar 

69. Allio, R. et al. MitoFinder: efficient automated large-scale extraction of mitogenomic data in target enrichment phylogenomics. Mol. Ecol. Resour. 20, 892–905 (2020).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

70. DePristo, M. A. et al. A framework for variation discovery and genotyping using next-generation DNA sequencing data. Nat. Genet. 43, 491–498 (2011).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

71. Danecek, P. et al. The variant call format and VCFtools. Bioinformatics 27, 2156–2158 (2011).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

72. Juvé, Y. et al. One mother for two species: obligate cross-species cloning in ants. Zenodo https://doi.org/10.5281/zenodo.11506545 (2025).

73. Kuznetsov, D. et al. OrthoDB v11: annotation of orthologs in the widest sampling of organismal diversity. Nucleic Acids Res. 51, D445–D451 (2023).

    Article  CAS  PubMed  Google Scholar 

74. Danecek, P. et al. Twelve years of SAMtools and BCFtools. Gigascience 10, giab008 (2021).

    Article  PubMed  PubMed Central  Google Scholar 

75. Katoh, K. & Toh, H. Recent developments in the MAFFT multiple sequence alignment program. Brief. Bioinform. 9, 286–298 (2008).

    Article  CAS  PubMed  Google Scholar 

76. Ranwez, V., Douzery, E. J. P., Cambon, C., Chantret, N. & Delsuc, F. MACSE v2: toolkit for the alignment of coding sequences accounting for frameshifts and stop codons. Mol. Biol. Evol. 35, 2582–2584 (2018).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

77. Romiguier, J. et al. Ant phylogenomics reveals a natural selection hotspot preceding the origin of complex eusociality. Curr. Biol. 32, 2942–2947 (2022).

    Article  CAS  PubMed  Google Scholar 

78. Lau, M. K. et al. Draft Aphaenogaster genomes expand our view of ant genome size variation across climate gradients. PeerJ 7, e6447 (2019).

    Article  PubMed  PubMed Central  Google Scholar 

79. Nygaard, S. et al. The genome of the leaf-cutting ant Acromyrmex echinatior suggests key adaptations to advanced social life and fungus farming. Genome Res. 21, 1339–1348 (2011).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

80. Minh, B. Q. et al. IQ-TREE 2: new models and efficient methods for phylogenetic inference in the genomic era. Mol. Biol. Evol. 37, 1530–1534 (2020).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

81. Capella-Gutiérrez, S., Silla-Martínez, J. M. & Gabaldón, T. trimAl: a tool for automated alignment trimming in large-scale phylogenetic analyses. Bioinformatics 25, 1972–1973 (2009).

    Article  PubMed  PubMed Central  Google Scholar 

82. Yang, Z. PAML 4: phylogenetic analysis by maximum likelihood. Mol. Biol. Evol. 24, 1586–1591 (2007).

    Article  CAS  PubMed  Google Scholar 

83. dos Reis, M. & Yang, Z. Approximate likelihood calculation on a phylogeny for Bayesian estimation of divergence times. Mol. Biol. Evol. 28, 2161–2172 (2011).

    Article  PubMed  Google Scholar 

84. Branstetter, M. G., Longino, J. T., Reyes-López, J. L., Brady, S. G. & Schultz, T. R. Out of the temperate zone: a phylogenomic test of the biogeographical conservatism hypothesis in a contrarian clade of ants. J. Biogeogr. 49, 1640–1653 (2022).

    Article  Google Scholar 

85. Rambaut, A., Drummond, A. J., Xie, D., Baele, G. & Suchard, M. A. Posterior summarization in Bayesian phylogenetics using Tracer 1.7. Syst. Biol. 67, 901–904 (2018).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

86. Purcell, S. et al. PLINK: a tool set for whole-genome association and population-based linkage analyses. Am. J. Hum. Genet. 81, 559–575 (2007).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

87. Romiguier, J. et al. Comparative population genomics in animals uncovers the determinants of genetic diversity. Nature 515, 261–263 (2014).

    Article  ADS  CAS  PubMed  Google Scholar 

88. Schlupp, I. The evolutionary ecology of gynogenesis. Annu. Rev. Ecol. Evol. Syst. 36, 399–417 (2005).

    Article  Google Scholar 

89. Lavanchy, G. & Schwander, T. Hybridogenesis. Curr. Biol. 29, 539 (2019).

    Article  CAS  PubMed  Google Scholar 

Download references

Author notes

1. These authors contributed equally: Y. Juvé, C. Lutrat, A. Ha

2. Unaffiliated: C. Lebas

Authors and Affiliations

1. ISEM, University of Montpellier, CNRS, IRD, Montpellier, France

   Y. Juvé, C. Lutrat, A. Ha, A. Weyna, E. Lauroua, N. Galtier & J. Romiguier

2. Department of Ecology and Evolution, University of Lausanne, Lausanne, Switzerland

   A. Weyna

3. Evo-Eco-Paleo UMR 8198, University of Lille, CNRS, Lille, France

   A. C. Afonso Silva & C. Roux

4. Department of Chemistry, Life Sciences, and Environmental Sustainability, University of Parma, Parma, Italy

   E. Schifani

5. Institut de Biologia Evolutiva, CSIC, University of Pompeu Fabra, Barcelona, Spain

   E. Schifani

6. Antarea, Saint-Aubin-de-Médoc, France

   C. Galkowski

7. Centre de Biologie pour la Gestion des Populations, INRAE, CIRAD, IRD, Montpellier SupAgro, Université de Montpellier, Montpellier, France

   R. Allio

8. Department of Developmental Biology, University of Plovdiv ‘Paisii Hilendarski’, Plovdiv, Bulgaria

   I. Stoyanov

9. Department of Ecology, Universität Innsbruck, Innsbruck, Austria

   B. C. Schlick-Steiner & F. M. Steiner

10. Institut NeuroMyoGène, CNRS UMR-5284, INSERM U-1314, MeLis, Université Lyon, Université Claude Bernard Lyon 1, Lyon, France

    D. Baas

11. LEHNA UMR 5023, Université Claude Bernard Lyon 1, CNRS, ENTPE, Villeurbanne, France

    B. Kaufmann

Contributions

J.R., Y.J., B.K., E.S., I.S., C.G., B.C.S.-S., F.M.S., C. Lebas, A.W., C. Lutrat, A.H. and E.L. sampled individuals. E.L., C. Lutrat, B.K., D.B., R.A. and J.R. performed lab work. A.C.A.S. supervised by C.R. and J.R. assembled the reference genome. A.W. supervised by J.R. and N.G. performed preliminary analyses of genetic data. C. Lutrat and J.R. monitored ant colonies in artificial nests and analysed egg genotyping. A.H. and J.R. performed analyses of species delimitation. Y.J., J.R., B.C.S.-S. and F.M.S. carried out the morphological descriptions and analyses. J.R. performed the final analyses of population genetics, phylogeny and molecular dating, conducted the statistical analyses and produced all the figures. J.R. wrote the manuscript with comments from all co-authors. J.R. conceived the project, acquired the funding and supervised all experiments.

Corresponding author

Correspondence to J. Romiguier.

Competing interests

The authors declare no competing interests.

Peer review information

Nature thanks Ehab Abouheif and the other, anonymous, reviewers for their contribution to the peer review of this work. Peer reviewer reports are available.

Reporting Summary

Peer Review File

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and permissions