Root-knot nematodes (genus Meloidogyne) exhibit a diversity of reproductive modes ranging from obligatory sexual to fully asexual reproduction. Intriguingly, the most widespread and devastating species to global agriculture are those that reproduce asexually, without meiosis. To disentangle this surprising parasitic success despite the absence of sex and genetic exchanges, we have sequenced and assembled the genomes of three obligatory ameiotic and asexual Meloidogyne. We have compared them to those of relatives able to perform meiosis and sexual reproduction. We show that the genomes of ameiotic asexual Meloidogyne are large, polyploid and made of duplicated regions with a high within-species average nucleotide divergence of ~8%. Phylogenomic analysis of the genes present in these duplicated regions suggests that they originated from multiple hybridization events and are thus homoeologs. We found that up to 22% of homoeologous gene pairs were under positive selection and these genes covered a wide spectrum of predicted functional categories. To biologically assess functional divergence, we compared expression patterns of homoeologous gene pairs across developmental life stages using an RNAseq approach in the most economically important asexually-reproducing nematode. We showed that >60% of homoeologous gene pairs display diverged expression patterns. These results suggest a substantial functional impact of the genome structure. Contrasting with high within-species nuclear genome divergence, mitochondrial genome divergence between the three ameiotic asexuals was very low, signifying that these putative hybrids share a recent common maternal ancestor. Transposable elements (TE) cover a ~1.7 times higher proportion of the genomes of the ameiotic asexual Meloidogyne compared to the sexual relative and might also participate in their plasticity. The intriguing parasitic success of asexually-reproducing Meloidogyne species could be partly explained by their TE-rich composite genomes, resulting from allopolyploidization events, and promoting plasticity and functional divergence between gene copies in the absence of sex and meiosis.
Sexual reproduction is evolutionary successful in animals as more than 99% of species reproduce sexually. The few animals that have abandoned sex are usually believed to be short-lived and outcompeted by their sexual relatives. Yet, in the root-knot nematodes, plant pests of global economic importance, an intriguing feature is observed. Species that have abandoned sex cause more damage and have a wider worldwide geographic distribution than their sexual cousins. To understand this puzzling success without sex, we have sequenced and analyzed the genomes of the three most devastating asexual nematodes and compared them to that of a sexual relative. We showed that their genomes are large, duplicated, transposon-rich and have hybrid origins. Due to this polylploid hybrid origin, most of their genes are in several copies with substantial sequence divergence. We detected signs of positive selection between these gene copies and confirmed functional divergence at the expression pattern level. We hypothesize that their peculiar hybrid genome structures provide these nematodes with a potential for adaptation and plasticity and could explain their paradoxical success in the absence of sex.
Citation: Blanc-Mathieu R, Perfus-Barbeoch L, Aury J-M, Da Rocha M, Gouzy J, Sallet E, et al. (2017) Hybridization and polyploidy enable genomic plasticity without sex in the most devastating plant-parasitic nematodes. PLoS Genet 13(6): e1006777. https://doi.org/10.1371/journal.pgen.1006777
Editor: Takashi Gojobori, National Institute of Genetics, JAPAN
Received: January 4, 2017; Accepted: April 24, 2017; Published: June 8, 2017
Copyright: © 2017 Blanc-Mathieu et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All sequence data generated in this study were deposited in the EMBL-EBI European Nucleotide Archive (ENA). The genome assemblies of the 3 Meloidogyne were deposited under project accession number PRJEB8714 with sample accession numbers ERS1696677, ERS671129 and ERS671128 for M. incognita, M. javanica and M. arenaria, respectively. The RNA-seq reads generated for M. incognita, M. javanica and M. arenaria were deposited under accession numbers ERA419974, ERA419971 and ERA419973, respectively. The genome and transcriptome data are also available for download and for exploration at the Meloidogyne portal ([http://meloidogyne.inra.fr).]).
Funding: This work was supported by the Agence Nationale de la Recherche program ANR-13-JSV7-0006—ASEXEVOL, the INRA program AAP SPE 2011, as well as Université de Nice Sophia-Antipolis Postdoc program 2013-2014. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Fully asexual reproduction occurs in only ~0.1% of animal lineages, which generally occupy shallow branches in the tree of life [1,2]. Although there are some exceptions [3–5], the majority of asexual lineages of animals seem to be recently derived from sexual lineages, suggesting they are generally short-lived. Asexual animals lack the possibility to combine advantageous alleles from different individuals via sexual recombination and in association with Hill-Robertson effect and linkage between conflicting alleles, selection is assumed to be less efficient [6,7]. Furthermore, Muller's ratchet  and Kondrashov’s hatchet  models of “clonal decay” predict that they progressively accumulate deleterious mutations. Supporting these models, different studies have demonstrated accelerated accumulation of harmful mutations in asexual lineages [10–14], or (short-term) increased accumulation of transposable elements (TE) in the absence of sex [15,16]. Hence, it is commonly postulated that obligate parthenogenetic animals have evolutionary and adaptive disadvantages compared to their sexual relatives and therefore represent evolutionary dead-ends. Consistent with the geographical parthenogenesis model, parthenogenetic populations of plant and animals are generally present at the edge of the geographical distribution of species, in marginal or anthropologically disturbed environments [17,18]. Their uniparental clonal reproductive mode is supposed to be advantageous for colonizing marginal environments where they escape competition with their sexual relatives. Indeed, parthenogenetic species are frequently found at higher latitudes and altitudes .
Root-knot nematodes (genus Meloidogyne) display a variety of reproductive modes ranging from sexual reproduction (amphimixis) to obligate asexual reproduction (apomixis) with intermediates able to alternate between sexual (amphimixis) and asexual (automixis) reproduction . These notorious plant pests have been ranked number one in terms of economic threat to the agriculture among all nematodes . Challenging the view that fully asexual lineages of animals are outcompeted by their sexual relatives, Meloidogyne species that reproduce without meiosis and without sex have a broader host range, a wider and more southern geographical distribution and are more devastating than their sexual relatives [21,22]. Whether some genomic singularities could account for their higher parasitic success despite the absence of sex remains unclear. In 2008, we coordinated the publication of the draft genome sequence of Meloidogyne incognita , an obligatory asexual nematode and the draft genome of M. hapla, a facultative sexual, was published the same year . One singularity of the M. incognita genome was the presence of genomic regions in two or more copies that spanned several megabases and had an average nucleotide divergence of ~8% . Such a structure was identified neither in the facultative sexual M. hapla nor in any nematode able to reproduce sexually, so far.
The possible origin of the duplicated and diverged genomic regions observed in M. incognita is still debated. Two main hypotheses for the origin of this duplicated genome structure are that (i) duplicated regions represent former paternal and maternal genomes that diverged and became rearranged after their diploid sexual ancestor became asexual or (ii) they result from interspecific hybridization events . As early as 1983, observation of heterozygous patterns of isozymes had led to the hypothesis that M. incognita might have undergone hybridization . Likewise, based on the presence of multiple divergent ITS nuclear markers within apomictic Meloidogyne despite closely related mitochondrial markers between species, it was suggested that these species had undergone hybridization from a set of closely related females with more diverse paternal lineages . Recently, on the basis of the comparative analysis of the initial M. incognita genome and a draft of the meiotic asexual M. floridensis genome, it was also suggested that M. incognita is of hybrid origin .
Regardless of its origin, the potential functional impact conferred by the duplicated genome structure of M. incognita has never been assessed. Furthermore, in the absence of genomes for other apomictic Meloidogyne, it was impossible to state whether such a duplicated genome structure is a specificity of M. incognita or a more general signature of the most economically important root-knot nematodes that, intriguingly, all reproduce asexually.
Here, we aimed at characterizing the genome structures of asexual root-knot nematodes, their most likely origin and the potential consequences at the functional level. We have re-sequenced the genome of M. incognita at much deeper coverage and have sequenced de novo the genomes of M. javanica and M. arenaria, two other apomictic root-knot nematodes of high economic importance. We have assembled these three genomes and validated genome assembly sizes with experimental assays. We confirm that the genomes of M. incognita and of the two other mitotic asexual Meloidogyne are made of duplicated yet diverged and rearranged genome copies. We have annotated the protein-coding genes and TE of the three genomes and performed a comparative genomic analysis, including the genome of the facultative sexual species M. hapla and the meiotic parthenogenetic M. floridensis. We show that the genomes of asexual mitotic Meloidogyne have a higher abundance of TE than M. hapla and any other nematode genome published so far. Using a phylogenomic analysis of the duplicated genomic regions conserved between species, we deciphered the origin and evolutionary history of the peculiar genome architecture of mitotic asexual Meloidogyne. To assess the potential functional outcome of these duplicated regions at a whole-genome scale, we searched and found signs of positive selection between the gene copies defining these genomic blocks. Using RNAseq analysis of different life stages of M. incognita, we show that the majority of gene pairs forming duplicated blocks display diverged expression profiles. Furthermore, gene pairs detected as under positive selection show a significantly higher proportion of diverged expression profiles. Our results show that mitotic asexual Meloidogyne possess duplicated, highly diverged and TE-rich genomes, an ensemble of features unequaled in any other nematode genome so far. We propose that the peculiar genome structures of these nematodes offer potential for adaptive plasticity and might contribute to the paradoxical success of these plant-parasitic animals despite the absence of sex.
Genome assemblies are large, experimentally supported and show high completeness
We sequenced the genomes of three asexually reproducing Meloidogyne species and the assemblies reached 184, 236 and 258 Mb, for M. incognita (Mi), M. javanica (Mj) and M. arenaria (Ma), respectively (Table 1). These genome assemblies are bigger than any Meloidogyne genome assembly reported so far. To confirm these genome sizes, we measured DNA content via flow cytometry experiments and obtained size estimates of 189 ±15, 297 ±27 and 304 ±9 Mb for Mi, Mj and Ma, respectively. The genome assembly of Mi was in the range of estimated size via flow cytometry whereas Mj and Ma assemblies were smaller by 34–60 Mb. To check whether these differences in sizes could be explained by duplicated or repetitive regions collapsed during genome assembly, we plotted the distribution of read coverage along the genome (S1 Fig). We estimated that 17.1 (Mi), 42.9 (Mj) and 21.6 (Ma) Mb of genome assemblies have a coverage twice higher than the rest of the genome sequence and might represent collapsed duplicated regions. Hence, part of the differences in sizes can be explained by these collapsed regions, as previously observed in the genome of the obligate mitotic rotifer Adineta vaga . The genome assemblies contained 97% (Mi), 96% (Mj) and 95% (Ma) of the 248 Core Eukaryotic Genes (CEG) in complete length . These are the highest scores for a Meloidogyne genome so far, suggesting these assemblies are the most complete available to date. We annotated 45,351 (Mi), 98,578 (Mj) and 103,269 (Ma) genes (including, protein-coding genes, ncRNAs, rRNAs and tRNAs). Protein-coding genes spanned up to 43.7 (24%), 75.2 (32%) and 82.2 (32%) Mb of the Mi, Mj and Ma genomes, respectively. Because genome assemblies of the asexual Meloidogyne were ~3–5 times bigger than the haploid genome size of the facultative sexual M. hapla, we suspected these genomes to be polyploid. As a proxy to estimate ploidy levels, we mapped back all the protein-coding sequences (CDS) to their respective genome assemblies and analyzed the proportion of CDS mapping one locus or several loci in the genomes (Fig 1). In the facultative sexual M. hapla, we observed a peak of CDS mapping one single locus in the genome, indicating no sign of whole genome duplication (WGD). In contrast, in the mitotic asexuals, we observed peaks of CDS mapping 3, 3 to 4 and 4 loci in the genomes of Mi, Mj and Ma, respectively. These CDS mapping multiple loci are consistent with the genome sizes of the asexual Meloidogyne (3-5x bigger than the M. hapla genome) and support their polyploid nature.
Occurrences (y axis) of Meloidogyne CDS mapping at minimum 95% identity on minimum 2/3 of their length to one, two, three, four or up to 15 loci (x axis) in their respective genomes. In M. hapla (red), >89% of the CDS map to one single locus while >85% of the CDS map to multiple loci in the apomicts M. incognita (green), M. javanica (blue) and M. arenaria (violet).
Transposable elements (TE) covered 50.0% (Mi), 50.8% (Mj) and 50.8% (Ma) of the genome assemblies. In comparison, only 29.2% of the M. hapla genome was covered by TE, using the same annotation protocol (Table 2). Due to its high fragmentation state, the genome of M. floridensis could not be annotated for TE. On average, 27–30% of the genes of mitotic parthenogenetic species are included within TE, whereas this proportion reaches only 17% in M. hapla (see TE section for more details).
The genomes of asexual Meloidogyne are highly duplicated and gene copies form divergent syntenic blocks
Genome sizes as well as distribution of multi-mapping CDS strongly suggested polyploidy in the asexual Meloidogyne (see above). We used MCScanX  to further investigate the duplication relationships of the protein-coding genes in Meloidogyne genomes. MCScanX classifies protein-coding genes as (i) singleton when no duplicates are found in the assembly, (ii) proximal when duplicates are on the same scaffold and separated by 1 to 10 genes, (iii) tandem when duplicates are consecutive, (iv) WGD or segmental when duplicates form collinear blocks with other pairs of duplicated genes and (v) dispersed when the duplicates cannot be assigned to any of the other categories. In the three mitotic asexual Meloidogyne species, 93.0–94.1% of protein-coding genes were estimated to be duplicated whereas only 46.6% were duplicated in the facultative sexual M. hapla and 52.9% in the meiotic parthenogen M. floridensis (Table 3). We noted that the dispersed category was the most frequent in all Meloidogyne genomes. However, this proportion negatively correlated with N50 values in the mitotic Meloidogyne, suggesting that these duplicates might be re-classified in other categories in the future.
Interestingly, 12,445, 5,806 and 15,632 genes were classified in the WGD / segmental category in Mi, Mj and Ma, respectively (Table 3). They formed 933 (Mi), 581 (Mj) and 1,648 (Ma) pairs of segmentally duplicated genome regions. In contrast, there were only 90 genes forming 11 pairs of duplicated regions in M. hapla and only 12 genes forming one pair of regions in M. floridensis (Table 4). Collinear duplicated regions span up to 58.6, 14.8 and 59.0 Mb of Mi, Mj and Ma genomes, corresponding to 31.8%, 6.3% and 23.0% of their respective sizes (Table 4). Average nucleotide divergence between pairs of duplicated regions was 8.4%, 7.5% and 8.2% for Mi, Mj and Ma, respectively, indicating a similar average divergence of ~8%. The distribution of % divergence between duplicated regions presented one single to two almost totally overlapping peaks (S2 Fig). This observation holds for the three apomictic Meloidogyne and suggests the duplication events have occurred in a same time window. The divergence levels were substantially lower in coding regions (4.7, 6.0 and 5.9%) than in non-coding regions (9.7, 9.0 and 9.7% for intergenic and 11, 10.4 and 11.1% for introns) for Mi, Mj and Ma, respectively (Table 4, S3 Fig).
Median rates of synonymous substitutions (Ks) between gene pairs forming duplicated regions were 0.1 for the three apomictic Meloidogyne. Averaged Ks for duplicated pairs of regions were significantly negatively correlated with collinearity (P<10−8 for Mi, Mj and Ma), which we measured as the fraction of collinear genes within a pair of regions (S4 Fig). This indicates that divergence in terms of number of conserved genes between a pair of regions correlates with the nucleotide divergence of the coding sequences.
In M. incognita and M. arenaria, we found 5 instances of duplicated regions on a same scaffold (Fig 2). In M. incognita, this corresponded to 42 collinear genes present in 4 pairs of tandem regions and 1 palindrome, whereas in M. arenaria, we found 29 collinear genes present in 2 pairs of tandem regions and 3 palindromes. If the duplicated regions represent vestiges of homologous chromosomes, such tandem and palindrome structures appear consistent with absence of chromosome pairing and meiosis, such as in the genome of A. vaga . No similar structure was found in the genomes of M. javanica, M. hapla or M. floridensis. Average Ks value of gene pairs forming tandem or palindromic regions were in the range of Ks values measured for gene pairs in the rest of duplicated regions, suggesting they have the same divergence times.
Five pairs of duplicated homologous collinear regions co-occur on a same scaffold in M. incognita (A) and five in M. arenaria (B). All curves show connections between collinear gene pairs used by MCScanX to define duplicated regions (blue curves show tandem duplications and purple curves show palindromic duplications). Grey lines represent the gene density on the scaffolds.
We found that 29–37% of gene copies forming duplicated regions in mitotic Meloidogyne were TE-derived, a proportion comparable to the proportion observed for the rest of the whole gene sets (27–30%). Furthermore, the distribution of Ks for pairs of TE-derived genes is not significantly different from the distribution of the rest of pairs of genes in duplicated regions, according to a Wilcoxon test. Thus, we can rule out the possibility that the observed duplicated regions are the results of TE multiplications.
In plant genomes, following WGD, fractionation biases can be observed. One genomic copy tends to retain more genes and to accumulate less mutations than the other copy . For each pair of duplicated regions in Meloidogyne, we tested whether a bias of retention of ancestral genes could be observed (Methods). We found 24 (Mi), 1 (Mj) and 36 (Ma) cases where one region had significantly (Chi-square test, 5% level) retained more ancestral genes than its counterpart. By comparing all the genes in duplicated regions and the MCScanX classification of gene copies, we also estimated that only 6%, 4% and 5% of ancestral collinear genes had no more copies anywhere in the Mi, Mj and Ma genomes, respectively and had probably been lost after WGD.
The duplicated regions have different origins and evolutionary histories
To decipher the evolutionary history of the duplicated structure observed in the mitotic Meloidogyne, we conducted a phylogenomic analysis focused on the gene copies forming pairs of duplicated regions. We identified and used a dataset composed of 60 groups of homologous genomic regions defined as follows. The genomic regions must be conserved and contain at least 3 collinear genes in 2 copies in each of the mitotic Meloidogyne vs. one single copy in the amphimictic M. hapla (Fig 3 for an example). These 60 groups of genomic regions encompass 2,202 homologous genes distributed in 222 clusters (Fig 3A for an example, S5 Fig for the 60 conserved and duplicated homologous regions). Within each of the 60 groups of genomic regions, we generated multiple alignments of all the clusters of homologous genes individually. Although duplicated genes within a species are, on average, relatively distant (Ks = 0.1, 5–6% nucleotide divergence), gene copies between species can occasionally be identical. In order to maximize phylogenetic signal, the multiple alignments of each cluster of homologous genes were concatenated in each group of conserved duplicated regions. From the 60 concatenated multiple alignments, we successfully generated 54 maximum-likelihood (ML) phylogenies (6 failed because of short alignments, see Fig 3B for an example tree and S5 Fig for the 54 ML trees of conserved homologous regions).
A. Circos  plot showing the collinear gene pairs (forming homologous regions) that were used for phylogenetic analyses (units = kb). All curves show the connections between the collinear gene pairs used by MCScanX to define segmental duplications. In each Circos plot, color codes are as follows. Collinear orthologs between M. hapla and any of the three asexuals species are in grey. Collinear ‘homoeologs’ within asexual species are in purple. Collinear orthologs between M. arenaria and M. javanica are in green. Collinear orthologs between M. arenaria and M. incognita are in yellow. Collinear orthologs between M. incognita and M. javanica are in red. The outer blue lines represent the gene density on the scaffolds. B. Maximum-likelihood phylogeny of concatenated alignments of collinear protein-coding genes used to form blocks with SH-like branch support. Topologies identical to the mitochondrial phylogeny were considered to represent the maternal contribution to the nuclear genome. The other topologies were considered as representative of paternal contributions.
Only three possible bifurcating monophyletic topologies exist to separate the three mitotic Meloidogyne (Fig 4): (1): (Mi, (Ma; Mj)), (2): (Ma, (Mi; Mj)) or (3): (Mj, (Ma; Mi)). We identified 60 such monophyletic clades and the most frequent corresponded to topology 1, observed 33 times. Topology 2 was observed 15 times and topology 3, 12 times (Fig 4). A total of 20 trees combined two of the three topologies mentioned above and allowed testing whether or not the two duplicated regions present the same evolutionary history across the 3 mitotic Meloidogyne. Among these 20 trees, a majority (13) combined two different topologies for the apomictic Meloidogyne, suggesting that the two regions have different evolutionary histories rather than a common ancestral duplication (Fig 5). The combination of topologies 1 and 2 was observed 7 times (see Fig 3B for an example). The combination of topologies 1 and 3 and the combination of topologies 2 and 3 were each observed 3 times. Only 7 of the 20 trees showed twice the same topology; and in all these cases this was twice topology 1.
The three possible topologies for bifurcating trees separating M. incognita, M. javanica, M. arenaria and their sexual relative M. hapla are represented as (1), (2) and (3) and their relative observed frequencies are indicated in the associated pie chart. The frequencies were calculated from the 40 phylogenetic trees containing at least one monophyletic clade with the 3 apomictic Meloidogyne that were constructed from the concatenated alignments covering a total of 2,202 protein-coding genes.
(A) Schematic representation of two duplicated regions each containing 3 collinear genes (a, b, c and a’, b’, c’). (B) ML trees combining twice the same topology among (1), (2) and (3), for the two duplicated regions (further detailed in Fig 4). These trees suggest the two duplicated regions have the same evolutionary history. (C) ML trees combining two different topologies for the duplicated regions. These trees suggest the two duplicated regions have different evolutionary histories. The relative frequencies of trees combining twice the same (red) or two different (blue) topologies are indicated in the big central pie chart. Relative frequencies within the red and blue categories are indicated by small pie charts next to the corresponding schematic tree. (D) The most frequently observed ML trees consist in a combination of topologies (1) and (2).
Part of the genes forming duplicated regions were present in more than two copies in at least one apomictic species. We identified 387 groups of homologous collinear genes (total of 4,262 genes) forming 3 or 4 duplicated regions in at least one mitotic Meloidogyne and at least 2 duplicated regions in the other mitotic. To decipher the evolutionary history of these additional copies, we counted the number of times the third or fourth copies hold a recent in-paralog (or allele-like position) relative to another copy, vs. the number of times these copies were in a new independent branching position (Fig 6). For genes present in 3 copies in a given genome assembly, the number of allele-like relationship was significantly lower (binomial test, P<10−6) than the number of new phylogenetic position, for all three mitotic species (Table 5). Hence, genes present in 3 copies more frequently formed a new independent branch in the phylogenetic trees than species-specific recent paralogs or allele-like branches. For genes in four copies within a given genome assembly, the number of allelic-like relationship was lower than the number of new positions in all species but the difference was significant (binomial test, P<10−5), for M. arenaria only (Table 5).
(A) An example of homoeologous relationship where each M. arenaria gene copy (arrows) clusters with the copy of another species. (B) An example of allele-like relationship where two of the three M. incognita gene copies (arrows) are more similar to one another than they are to a copy from another species.
Overall, this ensemble of results suggests that the duplicated genome regions have different evolutionary histories and thus probably result from allopolyploidization. These pairs of regions and the corresponding gene pairs can thus be considered as homoeologous . This term refers here to pairs of genes that originated by speciation and were brought back together in the same genome by hybridization.
Asexual Meloidogyne share nearly identical mitochondrial genes
To reveal the maternal evolutionary history of Meloidogyne species included in our analysis, we performed a phylogeny based on mitochondrial protein-coding genes as well as the 12S and 16S rRNAs (S1 Text). The phylogenetic tree (Fig 7, S6 Fig, S1 Table) returned the following highly-supported topology: (Ma,((Mi,Mf),Mj))). This topology corresponds to topology 2, the second most frequently observed in the analysis of the 60 groups of homoeologous duplicated regions (Fig 4). This suggests that genomic regions displaying topology 2 correspond to the maternal contribution to the nuclear genome. We also measured the average nucleotide divergence of mitochondrial genes between the 3 apomictic Meloidogyne (Mi, Mj and Ma). On average, the inter-species nucleotide divergence was very low (0.17%) and ranged from 0.00 to 0.33%. In contrast, the average nucleotide divergence between the meiotic M. hapla and the three mitotic was 24.50% and ranged from 24.42 to 24.58%. Hence, mitochondrial phylogenetic analysis reveals a high similarity between mitochondrial genomes of the three mitotic species and a substantial distance to their sexual relative M. hapla.
Consensus phylogeny obtained using ML and Bayesian analyses on 14 concatenated mitochondrial genes (12 protein-coding and 2 rRNA). Posterior probability (above) and bootstrap (below) support values are given at each corresponding branches. Phylogenies were rooted with Pratylenchus vulnus as an outgroup. The tree is represented without taking branch lengths into an account for better visibility of phylogenetic relations; the same tree with actual branch lengths is available in S6 Fig. Species with a (L) suffix indicate sequences coming from the genome sequencing effort undertaken as part of this paper. Species with a (D) suffix indicate sequences coming from other databases. Full species names and accession numbers for the sequences coming from external databases are given in S1 Table.
The duplicated genomes of asexual Meloidogyne enable positive selection between gene copies
We tested whether gene redundancy due to the duplicated genomic regions might result in a relaxation of selective pressure on the gene copies. We employed two different strategies to detect positive and episodic diversifying selection. One raw approach based on pairwise computation of the ratios of rates of non-synonymous (Ka) vs. synonymous mutations (Ks); and a phylogeny-based statistical approach. We found that 612 (8.8%) (Mi), 698 (22.4%) (Mj) and 2,061 (20.9%) (Ma) homoeologous gene pairs had a Ka / Ks ratio greater than 1, indicating possible positive selection (Fig 8). In a second, phylogeny-based approach, we looked for signs of episodic diversifying selection (EDS) in homoeologous genes shared by the 3 apomictic Meloidogyne genomes. We retrieved all homoeologous genes present in the three apomictic species and M. hapla and in 2 or more copies in at least one apomictic Meloidogyne. We found 1,735 such groups and used them to generate multi-gene alignments and their respective ML midpoint-rooted phylogenies. Using the random effects branch-sites model , we found 172 (Mi), 109 (Mj) and 208 (Ma) gene copies showing evidence of EDS (S7 Fig for an example) at the 0.05 confidence level (P_Holm: corrected for 9 tests using the Holm-Bonferroni procedure). Among these genes, 20 (Mi), 21 (Mj) and 47 (Ma) were also found to have Ka / Ks ratios >1.
Histogram displaying the distribution of the ratio of rates of non-synonymous (Ka) / rates of synonymous (Ks) nucleotide substitutions for pairs of duplicated collinear genes in M. incognita, M. javanica and M. arenaria. Pairs with a Ka / Ks > 1 indicate positive selection between genes.
To assess which functional categories were affected by positive selection or EDS, we examined Pfam domains and gene ontology (GO) terms associated to these genes (S1 Text). Overall, a large variety of Pfam domains and associated GO terms were identified among proteins encoded by genes under positive selection (Ka / Ks >1) or subject to EDS, in the three apomictic species. As many as 363, 304 and 674 distinct Pfam domains, corresponding to 177, 167, and 310 distinct GO terms were found in proteins encoded by genes with Ka / Ks >1 in Mi, Mj and Ma, respectively (S2 Table). Similarly, we identified 123, 78 and 174 distinct Pfam domains, corresponding to 93, 56 and 112 distinct GO terms, in proteins encoded by genes under EDS in Mi, Mj and Ma, respectively (S3 Table). Regardless of the dataset (Ka / Ks or EDS), few Pfam domains and GO terms were common to the 3 apomictic species and the majority of them were related to enzymatic, binding and metabolic activities as well as cell cycle-related and transport functions (S2 and S3 Tables). Mapping raw GO terms to the more generic GO-slim terms, revealed more overlap between the three mitotic species. The vast majority of GO-slim terms were shared by the three species in the EDS as well as in the Ka / Ks datasets (S2 and S3 Tables). Overlap between the EDS and Ka / Ks datasets was also high as 23 of the 28 GO terms shared by the three mitotic in the EDS dataset were also shared by them in the Ka / Ks dataset. These terms were mainly related to diverse enzymatic, catabolic, metabolic and biosynthetic functions. We identified significantly enriched GO and GO-slim terms in the Ka / Ks datasets as compared to the rest of homoeologous gene pairs (S2 Table). However, no significantly enriched GO or GO-slim term was common to the three species. No GO or GO-slim term was found to be enriched at the significance threshold (FDR<0.05) in the proteins under EDS as compared to the rest of homoeologous proteins.
Homoeologous gene copies show functional divergence at the expression pattern level
Functional divergence between gene copies can be viewed at different levels, including the biochemical function, the biological process or the expression pattern. Gene copies featuring the same biochemical function but expressed in different tissues or time points can be involved in different biological processes (e.g. development of different organs). To biologically assess whether functional divergence actually occurs between homoeologous gene copies, we analyzed their expression patterns across four developmental life stages (eggs, J2 infective juveniles, J3-J4 larval stages and adult females) of M. incognita using RNAseq (methods). We generated between 60.5 (J3-J4 replicate 2) and 94.2 million (egg replicate 2) 2x75bp paired-end reads across the 12 libraries (4 stages x 3 replicates). After all the cleaning steps, between 11.8 (J3-J4 replicate 2) and 48.5 (egg replicate 2) million clean paired-end reads were mapped to the M. incognita reference genome. The proportion of paired-end reads aligned on the genome varied between 75.0% (female replicate 3) and 96.2% (egg replicate 1). The majority of the read pairs (57.2–76.1%) mapped to a unique position on the genome (Table 6). Overall, a total of 42,705 M. incognita protein-coding genes (or 97.7% of 43,718 in total) had a log10(FPKM+1) >1 in at least one sample and were considered as expressed. After filtering out low-signal values as well as low-complexity genes, a total of 38,870 expressed genes remained, including 6,767 homoeologous gene pairs (7,299 initially). This ensemble of expressed genes was classified into 24 distinct expression clusters (methods). We assessed whether homoeologous gene pairs tended to fall in the same expression cluster or in different expression clusters, indicating functional divergence. We found that 4,326 out of 6,767 (63.9%) expressed homoeologous gene pairs showed signs of diverged expression by being assigned to two different expression clusters (Fig 9). Interestingly, pairs of homoeologous genes showing evidence of positive selection in M. incognita tend to be more often in different expression clusters than those showing no sign of positive selection (74.1% vs 63.9%, p-value <10−7). This ensemble of results biologically confirms that the peculiar allopolyploid genome structure of asexual root-knot nematodes is associated to functional divergence between gene copies.
For each of the 24 expression clusters, the information ‘n = a/b/c’ indicates. a: the total number of M. incognita genes having this expression pattern. b: the number of times the two genes from a same pair have this same expression pattern. c: the number of genes part of a pair having this pattern, while the other cognate gene has a different expression pattern (i.e. belongs to another expression cluster). Hence, c refers to gene pairs with diverged expression patterns. The expression ranks across the four developmental life stages (eggs, J2 infective juveniles, J3-J4 larval stages and adult females) are represented in colors as follows rank1: lowest expression (blue), rank2: second lowest expression (green), rank3: second highest expression (orange) and rank4: highest expression (red).
Asexual Meloidogyne genomes are rich in genes and in transposable elements
Although recombination can prevent the accumulation of TE, sexual reproduction can favor transmission of TE between individuals. In parallel, hybridization can initially favor TE multiplication by exposing naïve host genomes to TE uncontrolled by their inactivation machinery (e.g. chromatin modification or small RNAs). Thus, we investigated whether differences in TE abundance could be revealed between sexual and asexual Meloidogyne. With 29.2% of its genome occupied by TE, M. hapla has a relatively high TE abundance compared to other nematodes. Indeed, TE span 16.5% and 22.4% of the genomes of C. elegans and C. briggsae, respectively , 18% in Trichinella spiralis , 14–15% in Brugia malayi  and 22% in Bursaphelenchus xylophilus . We found that TE span 50.0, 50.8 and 50.8% of the genome assemblies of the asexual Mi, Mj and Ma, respectively (Table 2). The genomes of the asexually reproducing Meloidogyne thus appear to be particularly rich in TE and 1.7 times richer than the only sexual Meloidogyne genome available to date. Consistent with this observation, Class I retro-elements are on average 1.5 times more abundant in the asexual species. Within Class I elements, DIRS-like (Dictyostelium intermediate repeat sequence) appear to have undergone a particular expansion in the asexuals as they are on average 5.5 times more abundant than in the sexual species. Class II DNA transposons are 1.9 times more abundant in the three apomictic species than in the M. hapla genome. Although Helitron occupy a comparable proportion in asexuals and in the sexual, all the other categories are more than 2 times more abundant in the asexuals. This includes Maverick-like and TIR (terminal inverted repeats) elements as well as “unclassified” TE that possessed characteristics of Class II elements but could not be further assigned to one family. The rest of the potential TE is in the “other” category, which gathers DNA fragments displaying contradictory features of both Class I and II elements. This category was also more abundant in the asexuals than in the sexual species (~1.8 times). This overall abundance of TE in asexual Meloidogyne has implications at the protein-coding level. While 27–30% of the protein-coding genes of asexual Meloidogyne are totally included within TE, only 17% of M. hapla genes are within TE. Hence, TE abundance partly explains the higher number of genes observed in the asexual Meloidogyne (43,718–102,269 compared to 14,207 in M. hapla).
We tested whether the higher gene numbers observed in asexual Meloidogyne were homogeneously distributed along all protein domain families. We plotted the abundance of protein domains in Mi, Mj and Ma as a function of their abundance in Mh (Fig 10). The abundances of protein domains in Mi, Mj and Ma were all positively correlated to the abundance in Mh (R2 = 0.92, R2 = 0.89 and R2 = 0.87 for Mi, Mj and Ma, respectively). The slopes of the linear regressions were 3.06, 4.49 and 4.80 for Mi, Mj and Ma, respectively, suggesting that most of the protein domains are between 3 and 5 times more abundant in the three asexuals as compared to M. hapla. We compared the abundance of Pfam domains known to be found in TE-related genes and important for their own transposition activity (e.g. reverse transcriptase, integrase, transposase) in the four Meloidogyne species. We found that, on average, these domains were 3.4 to 9.8 times more abundant in asexual Meloidogyne than in M. hapla (S4 Table). For instance, rve (integrase core domain) is present in 205–689 copies in the three asexuals while it is found in only 59 copies in M. hapla. Similarly, the DDE_ 3 (DDE superfamily endonuclease) domain is absent in the M. hapla protein set while it is found in 13–61 copies in the three asexuals. This suggests that expansion of at least some families of TE might be in part responsible for the higher number of protein-coding genes in the asexuals.
The x-axis represents abundance of Pfam domains in M. hapla and the y-axis represents abundance of the same domains in M. incognita (red diamonds) M. javanica (green squares) and M. arenaria (yellow triangles). Linear regressions of M. incognita, M. javanica and M. arenaria Pfam domain abundances as a function of the abundance in M. hapla are plotted alongside their respective equations and correlation coefficients. Pfam domains associated to transposable elements (TE) are represented as crossed red, green and yellow squares in M. incognita, M. javanica and M. arenaria, respectively.
Asexual root-knot nematodes have large and diverged polyploid genomes
Meiotic pairing and segregation require high sequence identity and collinearity between homologous chromosomes. Usually, sequencing the genome of a diploid sexual animal involves performing repeated cycles of inbreeding to obtain lineages virtually homozygous at all loci. Genome assembly then results in collapsing all these virtually identical paternal and maternal variants into one single haploid reference sequence. We actually observed this in M. hapla which was assembled into a ~54 Mb genome , similar to previously reported measures of haploid genome sizes (~50 Mb) [38,39], and confirmed by our flow cytometry measures (~60 ± 1.5 Mb, Table 1). Concordance of genome assembly size with experimental measures, associated to the absence of extensive duplications of genomic regions, indicate a canonical sexual diploid genome. The haploid chromosome number of M. hapla is n = 16, similar to the putative ancestral haploid number of chromosomes (n = 18) in Meloidogyne species [19,22,40]. Hence, we can hypothesize that the ancestral haploid genome size for a Meloidogyne is ~55–60 Mb with n = 16–18 chromosomes. The genome assembly sizes of the three mitotic Meloidogyne species we describe here reach ~180 Mb, ~235 Mb and ~260 Mb for Mi, Mj and Ma, respectively. This represents ~3x, ~4x and >4x the expected haploid genome size for a Meloidogyne species. Flow cytometry measures of nuclear DNA content suggest an even larger genome size of up to ~300 Mb for Mj and Ma (Table 1). Hence, the genomes of apomictic Meloidogyne (~180–300 Mb) are 3 (Mi) to 5 times (Ma) bigger than the ancestral haploid genome size for a Meloidogyne species. Furthermore, Pfam domains are on average 3, 4.5 and ~5 times more abundant in the mitotic Mi, Mj and Ma genomes, respectively than in the M. hapla genome (Fig 10). Moreover, alignment of the CDS to the respective Meloidogyne genomes showed that while in the sexual M. hapla, most CDS map one single locus; we observe a peak at 3 matching loci for Mi, a peak between 3 and 4 for Mj and a peak at 4 matching loci for Ma (Fig 1). Finally, we showed that a substantial proportion of gene copies form collinear blocks of duplicated genome regions. Taken together, these results strongly suggest that the genomes of mitotic Meloidogyne are polyploid, with M. incognita being most likely triploid, M. javanica tetraploid and M. arenaria tetra- to pentaploid. Observation of chromosome numbers (supplementary discussion S1) further supports polyploidy of the three asexual Meloidogyne. Similarity between genome assembly sizes and measures of total nuclear DNA content via flow cytometry suggests that most of the former paternal and maternal donor genomes have been separately assembled, probably due to their high within-species divergence. Indeed, the duplicated collinear genome regions span several Mb and thousands of genes in each mitotic Meloidogyne genome and show a similarly high within-species average nucleotide divergence of ~8%, confirming initial analysis of the Mi draft genome. Likewise, the per-site synonymous substitution rate (Ks) of collinear gene pairs that define these duplicated genome regions had a very similar median of 0.1 for all three species. This homogeneity of nucleotide divergence and Ks between pairs of collinear regions for the three mitotic Meloidogyne species suggests that they have duplicated in a same time window and thus separated for a similar amount of time.
The genome structures are consistent with absence of meiosis
Due to multiple synteny breakpoints, no long scaffold could be aligned on its whole length to another long scaffold in any of the three mitotic Meloidogyne. This rearranged chromosomal structure combined with the high average divergence between homoeologous blocks suggest chromosome pairing must be complicated if not impossible. Furthermore, in Mi and Ma, we observed collinear regions present in palindromic or tandem arrangement on a same scaffold. Such structures, similar to the ones observed in the ancient asexual bdelloid rotifer A. vaga , appear incompatible with segregation of homologous chromosomes in conventional meiosis. Both the difficulty of pairing homologous chromosomes and the impossibility to separate the genome in two equivalent chromosome sets are consistent with the absence of meiosis and the strict asexual reproduction of these organisms. Palindromes or tandem blocks were not observed in the genome of the meiotic M. hapla, and because this genome presents high contiguity (the highest for a Meloidogyne); this most probably represents true absence. Similar analysis could not be performed for the M. floridensis genome; due to its low contiguity and fragmented nature, (only 12 genes were found in one pair of duplicated regions).
The asexual Meloidogyne species most likely have recent hybrid origins
The duplicated genome regions in mitotic Meloidogyne tend to be more similar across different species than they are to their other copies within the same species. Furthermore, when duplicated regions form two or more clades in phylogenomic analysis, these clades more frequently present distinct topologies. Thus, collinear duplicated regions within a species have different origins and evolutionary histories and probably do not originate from common ancestral allelic regions that accumulated mutations separately (i.e. no Meselson-White effect). Contrasting with the high within-species divergence of duplicated blocks in the nuclear genome (avg. divergence ~8%), mitochondrial genes are almost identical in Mi, Ma and Mj (avg. divergence ~0.17%). This confirms previous observations that these three species share virtually identical mtDNA markers [26,41,42] and suggests that Mi, Mj, and Ma share closely related or common maternal ancestors. The mitochondrial genome is expected to accumulate mutations faster than the nuclear genome (e.g. 100–1,000 times more rapidly than the nuclear genome in C. elegans [43–45]). Hence, the divergence time between the nuclear genome copies within a same species is assumed much higher than the divergence time between the different species themselves, based on mitochondrial data. Inter-specific hybridization is the most likely hypothesis that could resolve at the same time the discrepancy between low between-species mitochondrial divergence and high within-species nuclear divergence levels and the observed topologies in the phylogenomic analysis (alternative hypotheses are discussed in S1 Text). We propose that not only Mi but also Mj and Ma most likely originated from multiple hybridization events with a same or closely related maternal donor lineage and different paternal donors. This confirms, at a whole genome scale, a previous formulation of this hypothesis based exactly on the same kind of observed discrepancy between low divergence in mitochondrial markers between species and high divergence in ITS nuclear markers within species of apomictic Meloidogyne . In this regard, the case of salamander in the Ambystoma genus constitute an interesting parallel. Indeed, some unisexual species in this genus are characterized by their absence of meiosis, their various ploidy levels (from 2n to 5n), their hybrid origin and their closely related mitochondrial genomes .
In apomictic Meloidogyne, because the mitochondrial divergence is very low, the speciation between Mi, Mj and Ma as well as the hybridization and associated loss of sex must be very recent. The very low proportion of collinear genes lost in duplicated genomic blocks further support recent whole genome duplication events (via hybridization). Indeed, after a whole genome duplication event, regardless whether it involves hybridization or not, most of the redundant gene copies are expected to be lost relatively rapidly. For instance, it has been shown in teleost fish that 70–80% of the genes have been rapidly lost after the latest WGD event .
Multiple hybridization steps probably led to the 3-5n polyploid genomes
Based on genome sizes, Pfam domain abundance and peaks of genes in 3, 4 or more copies, we estimated that Mi was most likely triploid while Mj and Ma were respectively most likely tetra to penta-ploids (see above). Some collinear regions are conserved in more than 2 copies within all the genomes, which allowed assessing the evolutionary histories of the third and fourth genome copies. For the three asexual Meloidogyne, the third copies of collinear genes were significantly more frequently similar to a cognate gene in another Meloidogyne species than to any of the other two copies found in the same species (Table 5). Thus, the third copies probably derive from a distinct hybridization event. This suggests a two-step hybridization process. First, homoploid hybridization (hybridization between two diploid AA and BB progenitors without associated genome doubling [48,49]) took place and led to a diploid AB hybrid. Then, a second hybridization between an unreduced AB gamete of the homoploid hybrid with a reduced C gamete of another sexual species led to the presence of three distinct copies (ABC) of nuclear genomes within a same species. It should be noted that unreduced gametes are frequently produced by inter-specific hybrids . Although this hypothesis could explain the triploid genome of M. incognita, additional steps are needed to explain the tetra to penta-ploid genomes of M. javanica and M. arenaria. We hypothesize that the triploid bridge pathway described in several polyploid plants [51,52] could explain the transition between triploids (similar to M. incognita) and tetraploids (similar to M. javanica). Indeed, under this hypothesis, triploids can constitute a bridge towards tetraploidy by producing unreduced triploid gametes, that, by fusing with a haploid gamete, would lead to a tetraploid progeny. Those tetraploids would in turn produce diploid gametes that would combine with haploid gametes of other species, creating a new triploid. And this triploid, could in turn serve as an intermediate towards new tetraploid by fusing unreduced triploid gametes with haploid gametes, constituting a “triploid-tetraploid-triploid” circle as suggested in . Finally, fusion of an unreduced triploid gamete with either a reduced gamete from a tetraploid or an unreduced gamete from a diploid, could lead to a pentaploid hybrid similar to M. arenaria.
In this perspective, the plant genus Boechera constitutes a good exemplary system for cases of hybridization at different ploidy levels and ecological success . Indeed, it constitutes a genus in which both homoploid diploid and triploid hybrids are present with also less frequent tetraploids or species of higher ploidy levels. Similarly to the Meloidogyne, the hybrids are apomicts and highly heterozygous.
The whiptail lizards constitute an interesting similar example of animal with fully asexual reproduction. Like the asexual Meloidogyne, these lizards are of hybrid origin. Interestingly, several polylploid lineages were identified and they also present a fixed heterozygosity [55,56].
High abundance of TE in asexual Meloidogyne possibly provides genome plasticity
Following loss of sexuality, it has been hypothesized that TE could invade the genomes because recombination would tend to favor their elimination . Alternatively, it has been suggested that the only asexual animals that survive are those that control TE multiplication in their genomes. Examples supporting these two different hypotheses exist in the literature. In Daphnia arthropods, sexual reproduction seems to be correlated with an initial slower accumulation of TE in genomes whereas, at the long-term sex is associated with higher TE loads . In parasitoid wasps, it has been shown that TE are more abundant in Wolbachia-induced asexual lineages than in sexual lineages . However, whether this is a consequence of sex loss or of Wolbachia infection remains to be clarified. In contrast, in the ancient asexual bdelloid rotifer A. vaga, TE occupy only 3% of the genome and while a high diversity of TE was found, they are generally present at very low copy numbers . This suggests that TE proliferation might be under control in this species. Recently, a comparison of the TE load in five sexual vs. asexual lineages of arthropods showed no evidence for TE accumulation in the asexuals . In Meloidogyne, we found that TE occupy ~50% of the genomes of the three mitotic species while they occupy only 29% of the genome of M. hapla. Although it appears that TE have proliferated in the genomes of the asexual Meloidogyne, this might be a consequence of their hybrid origin rather than of their mode of reproduction. Regardless their origin, this abundance of TE might constitute a potential for genomic plasticity in the absence of sexual recombination. Supporting this hypothesis, some canonical full length TE were previously experimentally identified in these Meloidogyne species . Furthermore, a Tm1 transposon has been identified in apomictic Meloidogyne but no homolog with an intact transposase could be found in the sexually-reproducing relative M. hapla . Interestingly, the Cg-1 gene, whose deletion is associated to resistance-breaking strains of M. javanica, has been identified within one of these Tm1 transposons. Thus, TE possibly have an adaptive impact on these nematodes, including on their host plant range.
The allopolyploid genomes are associated with positive selection between gene copies
Mi, Mj and Ma are exceptionally successful, globally distributed parasites of diverse agricultural crops [62,63]. Intriguingly, their geographical distributions and host ranges are wider than those of their sexual relatives. Furthermore, in controlled condition, they are able to overcome plant resistance within a few generations . In the absence of sex and meiotic recombination to provide genomic plasticity and adaptability, their allopolyploid nature may provide benefits contributing to their parasitic success. First, polyploidy can provide the raw material for neo- and sub-functionalization of duplicated gene copies, resulting in novel genetic variation [64,65]. It has been shown in yeast that ploidy level is correlated to faster adaptation . Also, it has been suggested that polyploidy could mask deleterious recessive alleles  and limit their accumulation via gene conversion between homologous regions . Furthermore, allopolyploidy, by combining several genomes in one species, may lead to transgressive phenotypes that surpass those of the parent species via novel genetic combination and heterosis [50,67,68]. Ambystoma salamanders constitute one clear case of transgressive phenotype in animals. Indeed the hybrids between one native and one introduced species are ecologically fitter and more successful than the parental native species as well as other related species in the native environment [69,70].
Here, we tested whether the presence of several divergent genomic copies in a same species, could have functional consequences at the coding level. Hybridization brings together homoeologs chromosomes and therefore orthologous gene copies within an individual. Because the three mitotic Meloidogyne have very close mitochondrial genomes, their speciation was certainly recent. Hence, we can hypothesize that most of the high within-species nucleotide divergence between duplicated genomic regions is due to hybridization rather than long-term divergence. Most likely, the hybrid inherits orthologs that had retained similar function and following functional redundancy, selective pressure on these genes may relax and drive them to different evolutionary trajectories [71,72]. In some cases, the relaxation of selective pressure can allow emergence of new adaptive mutations. We have shown that ~8 to 20% of gene copies coming from the duplicated genomic regions harbor signs of positive selection. A diversity of Pfam domains and associated gene ontology terms were predicted in proteins encoded by positively selected genes. Although many terms and domains were related to enzymatic and other catalytic functions, there was a poor overlap between the three apomictic species, and different domains and functions were specifically enriched in positively selected genes in each species. These observations suggest that the functional consequences of the hybrid genome structure were different in each species.
Functional divergence at the expression pattern level
In the model root-knot nematode M. incognita, we showed that more than 60% of homoeologous gene copies display diverged expression patterns. These gene copies resulting from hybridization have only single-copy equivalents in the sexual relative M. hapla. This biological confirmation of functional divergence suggests that additional genes in asexual root-knot nematodes are not just merely functionally redundant with their single-copy orthologs in the sexual relatives but actually support plasticity and variability. Thus, we can assume that the allopolyploid genome structures of asexual root-knot nematodes provide them with a reservoir of variability and adaptability that could partly compensate the absence of sexual reproduction. Noteworthy, these results are consistent with two other recent studies of hybrid animal genomes (the Atlantic salmon and the frog Xenopus laevis) that both also showed extensive functional divergence at the expression level between homoeologous gene copies [73,74]. Interestingly, we noted that the proportion of expression divergence is significantly higher (>70% vs. >60%, p-value<1.10−7) in homoeologous gene pairs that are under positive selection. These gene pairs combine both divergence at the expression level and accumulation of non-synonymous mutations that could lead to functional divergence at the biochemical level. They are thus the most obvious candidates for neo or sub functionalization.
How an animal can survive without sexual reproduction and compete with its sexual relatives remains an evolutionary puzzle. Intriguingly, asexually reproducing (apomictic) root-knot nematodes outcompete their sexual relatives as plant parasites of global economic impact. We have shown here that the genomes of the apomictic Meloidogyne are duplicated most likely because of a complex series of hybridization events. Although the parental lineages are unknown, they probably belong to the sexual relative clades. Hence, this parasitic success could be viewed as a case of transgressive phenotype, where the ecological success of the hybrid progeny surpasses those of the parents [27,68]. Furthermore, hybridization has been proposed as an important evolutionary phenomenon that could give rise to new parasites and pathogens. For instance, hybridization of two host-specific plant parasitic tephritid fruit flies gave rise to a new species able to parasitize a new invasive host plant . Similarly, hybridization of two Blumeria fungal pathogens gave rise to a new species that is able to infest a host plant of economic interest resistant to both progenitor species . In the asexual root-knot nematodes, the presence of duplicated and diverged genomic regions probably promotes functional novelty between resulting gene copies, following positive selection. We confirmed this functional divergence at the expression level at the whole genome scale. Furthermore, the TE-rich nature of their genomes might also foster genomic plasticity not only actively by TE movements across the genomes but also passively by promoting chromosomal shuffling between these repeated genomic regions. Such a TE-promoted chromosomal shuffling associated to adaptation to different host plants has already been shown in a plant-pathogenic fungus . Part of the intriguing success of mitotic asexual Meloidogyne could thus reside in their duplicated, diverged and TE-rich genomes resulting from hybridization. It would be interesting to explore the potential for plasticity and adaptation in the genomes of other asexual animals, particularly parasites and pathogens, to assess whether convergent or independent genomic characteristics support this potential.
Materials and methods
DNA samples preparation protocols for genome sequencing are detailed in the S1 Text. Genome assemblies were performed in four steps, following the same procedure as developed for resolving the degenerate tetraploid genome structure of the bdelloid rotifer A. vaga : (i) assembly of 454 data into contigs, (ii) correction of the 454 contigs using Illumina data, (iii) scaffolding of 454 contigs and (iv) gap closing using Illumina data. For the first step, we used the multi-pass assembler MIRA  version 3.9.4 (normal mode, default options except the number of cycles) to generate contigs from the 454 genomic libraries (S5 and S6 Tables). Although computationally demanding, running MIRA with multiple cycles is particularly appropriate to separate heterozygous regions in genomes, as anticipated in polyploid species. Moreover, Sanger reads of the M. incognita first draft genome sequence  were used to generate the current assembly. Twelve (M. arenaria and M. javanica) or sixteen (M. incognita) cycles were performed to separate a maximum of repeats and heterozygous regions. We subsequently used Illumina data to correct the homopolymer errors of the 454 contigs following a standard procedure . The corrected contigs were linked into scaffolds using the program SSPACE  with 454, Sanger and Illumina data. Finally, assemblies were gap-closed using GapCloser from the SOAPdenovo 2 package  with Illumina data. The statistics of the three genome assemblies are summarized in S5 Table. We assessed the completeness of the three genome assemblies by counting the number of Core Eukaryotic Gene (CEG) using CEGMA .
Experimental determination of nuclear DNA content
Flow cytometry was used to perform accurate measurement of nuclear DNA content in the three apomictic Meloidogyne (M. incognita, M. javanica and M. arenaria) as well as in the facultative sexual M. hapla, compared to internal standards with known genome sizes. Caenorhabditis elegans strain Bristol N2 (approximately 200 Mb at diploid state [83,84]) and Drosophila melanogaster strain Cantonese S. (approximately 350 Mb at diploid state [85,86]) were used as internal standards. Extraction of nuclei was performed as previously described . Briefly, for each Meloidogyne species about two hundred thousand stage 2 juveniles (J2s) were suspended in 2 mL lysis buffer (1mM KCl, 30 mM NaCl, 10 mM MgCl2, 0.2 mM EDTA, 30 mM Tris, 300 mM sucrose, 5 mM sodium butyrate, 0.1 mM PMSF, 0.5 mM DTT, 40 μl Igepal), grinded for 10 min with a Dounce homogenizer and filtered through a 0.20 μm nylon mesh. Subsequently, this 2 mL suspension was overlaid on top of 8 mL suspension buffer (same as lysis buffer except for sucrose, 1.2 M, and without Igepal) so that the tubes were ready for centrifugation (10,000 rpm, 30 min, 4°C) to reduce the level of debris and to pellet nuclei. Supernatant was completely discarded and pelleted nuclei were re-suspended in suspension buffer. Then nuclei suspension was stained, at 37°C for 30 min, with 75 μg/mL propidium iodide and 50 μg/mL DNAse-free RNAse. The same nuclei extraction protocol was performed at the same time on the samples and on the two internal standards. Flow cytometry analyses were carried out using a LSRII / Fortessa (BD Biosciences) flow cytometer operated with FACSDiva v6.1.3 (BD Biosciences) software. Data were analyzed with Kaluza v1.2 software (Beckman Coulter) and cytograms exhibiting peaks for each phase of the cell cycle (G0/G1, S and G2/M) were obtained. Standards and samples were processed both alone and together. Only mean fluorescence intensity of the first peak (arbitrary units), corresponding to G0/G1 phase of the cell cycle of the cytograms, was taken into account to estimate DNA content. In this method [88,89], the amounts of DNA in the Meloidogyne samples were determined by interpolating the fluorescence signals generated from the standards using the following equation:
The estimated DNA contents of the Meloidogyne samples were calculated by averaging the values obtained from three biological replicates (S8 Fig).
Search for collapsed duplicated regions
To check whether some nearly identical duplicated genomic regions had been collapsed during the assembly (as previously observed in the A. vaga genome ), we aligned the Illumina PE-reads of each species against their respective genome assembly sequence, using BWA  with default parameters. We computed the per base read coverage using BEDtools genomeCoverageBed  and plotted the distribution of the per-base coverage depth. This clearly showed 2 peaks for the three species, one systematically at twice the coverage of the first peak (S1 Fig). We calculated the number of bases with per base coverage comprised in the range of the second peak and summed it up to obtain the total size of the duplicated regions that had been collapsed during the assembly.
Predictions of protein-coding genes were performed using EuGene 4.1c , optimized and tested for M. incognita on a dataset of 301 non-redundant full-length cDNAs. Translation starts and splice sites were predicted using SpliceMachine . Three datasets of M. incognita transcribed sequences were provided to EuGene to contribute to the prediction of gene models: i) Sanger ESTs (Genbank 20110419), ii) a dataset of seven Illumina transcriptomes obtained in our lab in a previous study , and iii) a dataset of nine Trinity  assemblies of RNAseq data, generated in this study (S7 Table, S1 Text). Transcribed sequences were aligned on the genome using GMAP ; spliced alignments spanning 80% of the transcript sequence length at a 90% identity cut-off were retained. Similarities to i) C. elegans release Wormpep221, ii) G. pallida, release 1.0 , and iii) Swiss-Prot release December 2013 (excluding proteins similar to REPBASE ) were searched using BLAST  and provided to EuGene to contribute to gene modelling. The gene modelling algorithm used the standard parameters for the 4.1c version, except for the fact that i) the gene finding algorithm was applied on both strands independently allowing overlapping gene models, ii) non-canonical GC/donor and AC/acceptor sites were allowed on the basis of transcriptional evidences, iii) a gene model was not allowed to span a gap (‘N’) longer than 1,000 nucleotides, iv) the minimum length of introns was set to 35 nucleotides, and v) the minimum CDS length cut-off was set to 150 nucleotides. For M. arenaria and M. javanica, the EuGene pipeline, with models and parameters tuned on M. incognita, was used to annotate both genomes. Two modifications were applied on the selection of reference datasets i) Swiss-Prot (excluding proteins similar to REPBASE) and the proteome of M. incognita were used as reference proteomes ii) assemblies of M. arenaria and M. javanica RNAseq data were used as sources of transcription evidences.
The predicted protein sequences of M. incognita, M. javanica, M. arenaria and M. hapla were scanned for the presence of Pfam protein domains using the program PfamScan  against the Pfam-A HMM domain library (release 27.0), using default thresholds and parameters. A gene ontology annotation was inferred from the Pfam protein domain annotation using the pfam2go mapping file maintained at the gene ontology portal and generated from the InterPro2GO mapping . Gene ontology terms were also mapped on the generic GO-slim ontology using the GOSlimViewer utility developed as part of AgBase .
Analysis of whole-genome duplication
The duplicated structures of Meloidogyne species were estimated by detecting conserved blocks of duplicated genes. The protein sequences of each genome were initially self-blasted to determine a homologous relationship with an e-value threshold of 1e-10. Conserved blocks of duplicated genes were detected based on the gene locations in the genome using MCScanX  with default parameters. We required at least 3 collinear genes pairs for MCScanX to form a block. Using the perl script “add_ka_and_ks_to_colinearity.pl” included in the MCScanX package, we calculated Ks values for each homologous gene pairs between duplicated blocks. The median Ks value was considered a representative of the divergence between duplicated regions.
We used a custom python script (S2 Text) to compute the pairwise nucleotide identity between collinear blocks for each species. Briefly, pairs of duplicated genomic regions were extracted according to the GFF3 positions of their first and last collinear genes. They were then aligned using NUCmer from MUMmer v3.23 with default parameters. We then filtered out sub-alignment shorter than 50 nt (delta-filter -l 50) and summarized alignment using the dnadiff program from the MUMmer package. The average identity at the nucleotide level between duplicated regions was obtained from the output of dnadiff. Identity within coding and non-coding sequences was obtained by masking coding or non-coding sequences in each duplicated region before NUCmer alignment using BEDtools maskFastaFromBed v2.17.0 .
To analyze synteny conservation between genomes, we concatenated all the inter-/intra-species BLAST hits (e-value threshold of 1e-10) of M. incognita, M. javanica and M. arenaria and M. hapla protein sequences and fed MCScanX with this pooled BLAST result as well as with information on the location of the corresponding genes in the respective genomes, as recommended in the MCScanX manual for multi-species comparisons. The M. floridensis genome had to be discarded from this comparative analysis because only one pair of regions composed of 12 genes block was detected for this genome preventing any large-scale analysis of conserved synteny. We required at least 3 collinear genes pairs for MCScanX to detect a block. We parsed the results of the collinearity analysis between genomes of Meloidogyne species (HTML files output by MCScanX) to extract collinear genes forming duplicated regions conserved between Meloidogyne species. We used those homologous collinear genes to perform phylogenomics analyses (see below).
Determination of fragmentation bias
For each pair of duplicated regions, the genes present on each region were counted and the number of genes that were present in the ancestor of these two regions was calculated as the total number of genes on the two collinear regions minus the number of gene pairs. We then compared the number of genes in each region to the number of estimated genes in the ancestral region to statistically determine whether one region had retained significantly more genes than the other within a pair.
Alignments, phylogenies and topologies searching
First, protein sequences were aligned using MUSCLE v3.8.31 [107,108]. Second, protein alignments were back translated into codon alignment using PAL2NAL v.14  with the ‘nogap’ option. Third, codon alignments were trimmed using GBLOCKS  with default options. The fittest model of nucleotide evolution was searched using the function ModelTest as implemented in the R package phangorn . We then used PhyML  (-d nt -b -4 -m GTR -f e -t e -v e -a e -s BEST) to build maximum likelihood phylogenies with SH-like branches support on these pruned alignments. We rooted the phylogenies using the midpoint function of R package phangorn. For the rest of the analyses, we only retained the trees in which M. hapla displayed an outgroup position relative to the other Meloidogyne species in the midpoint-rooted topologies. Tree topologies were classified and counted using a custom R script. Phylogenetic tree figures were formatted and edited using EvolView .
Identifying gene copies subject to positive selection
To compute the Ka / Ks ratios per pair of collinear homologous genes, we used the Nei-Gojobori method  implemented in MCScanX. We eliminated all cases where 0.01<Ks<1 to discard genes that were either evolving extremely slowly or extremely rapidly and could potentially yield erroneous Ka/Ks estimates. We performed tests of episodic diversifying selection (EDS, a form of positive selection) using the random effects branch-sites model  implemented in the HYPHY package . We looped the branchSiteREL.bf script over the 1,735 multi-sequence alignments and their respective ML midpoint rooted trees. Each alignment contained at least one collinear protein-coding gene for all three apomictic species and M. hapla and a duplicate in at least one asexual species. We chose the adaptive version of BSRE and allowed branch-site variation in synonymous rates. Branch with length less than 0.01 were not considered because ω rate classes cannot be inferred reliably for very small branches (< 0.01).
RNAseq sample preparation and sequencing
Total RNAs were extracted from 4 M. incognita developmental stages (pre-parasitic J2s, parasitic J3-J4, adult females and eggs) using TRIzol Reagent (Invitrogen); three independent biological replicates were performed for each stage. Total RNA quality and quantity were assessed by a 2100 Bioanalyser (Agilent technologies). Samples with RNA integrity number (RIN) over 8.5 were kept for cDNA library construction, except for eggs samples for which RIN ranged between 6 and 7. An input of 100 ng total RNA was provided to construct cDNA libraries via the Ovation Universal RNAseq system (Nugen technologies). To eliminate unwanted rRNA transcripts, we designed 101 InDA-C primers to target M. incognita 28S and 18S transcripts for depletion. The 12 cDNA libraries (4 stages x 3 replicates) were quantified and equilibrated to 4 nM using Kapa QPCR (Kapa Biosystems). Finally, multiplexed libraries were sequenced on an Illumina NextSeq 500 sequencer on two High 150 flow cells (400M PE75 reads), on the UCA Genomix sequencing platform of Nice Sophia-Antipolis.
RNAseq reads cleaning, mapping and counts
The quality of the raw read fastq files were manually checked using FastQC and the following series of filters were applied to all the files. We first eliminated possible remaining ribosomal RNA contamination using SortMeRNA . We then used PRINSEQ  to trim sequence ends with quality score lower than 28, and only kept reads with an overall score >28 and a length >60 nucleotides. We aligned the cleaned reads to the M. incognita indexed genome assembly using the STAR 2-pass procedure .
We used RSEM  to estimate read counts, transcripts per million (TPM) as well as fragments per kilobase per million mapped reads (FPKM) for the M. incognita predicted protein-coding genes. RSEM takes into account multi-mapped reads and assigns them proportionally to the different loci according to probabilities estimated based on uniquely mapped reads.
Identification of homoeologous gene copies with different gene expression patterns
We transformed the raw FPKM values in log10 (FPKM+1) values. To avoid the risk of spurious read counts due to low complexity regions in transcripts, we eliminated all the transcripts that had more than 1/3 of their length covered by low-complexity regions, as measured by RepeatMasker . We also filtered out genes showing too much variability between replicates and showing an inside-replicate coefficient of variation of log10(FPKM+1) higher than 0.8. We finally averaged expression over each triplicate, and filtered out genes with log10(FPKM+1) mean expression values lower than 0.3 in all conditions, as this corresponded to low signal.
We then clustered genes according to their expression pattern in the four conditions. To be as robust and conservative as possible, we clustered together genes showing the same relative expression patterns in the four conditions, i.e. genes whose expression values are ranked in the same order between the four conditions, resulting in 24 (= 4!) groups. Homoeologous genes from a same pair that were located in two different gene expression clusters were considered as having diverged expression patterns.
Transposable elements annotation and analysis
TE detection and annotation.
Repeat annotation was performed using the REPET pipelines TEdenovo and TEannot . The TEdenovo pipeline was used to search for repeats in the contigs of the four Meloidogyne genomes pooled together (M. incognita, M. javanica, M. arenaria, M. hapla) to ensure that the annotation results were based on the same de novo predictions and were comparable between species. The high scoring segments pairs detection was performed by aligning all genomes against themselves using BLASTER  with the following parameters: identity >90%, High Scoring segments Pairs (HSP) length >100b & <20kb, e-value <1e-300. The LTRs retrotransposons were searched using LTRharvest  with the following parameters: LTR similarity >90% min LTR size >100bp & <1000bp. The repetitive HSPs identified from the BLASTER output were clustered using Recon , Grouper  and Piler . The predictions from LTRharvest were clustered using the MCL algorithm. The consensus for each cluster are obtained using MAP  and classified with PASTEC Classifier . The unclassified consensus sequences (noCAT) were filtered out to keep only consensus sequences built from at least 10 sequences in the cluster. The library of 10,535 consensus sequences obtained with TEdenovo was used for separate annotation of each genome using TEannot with the same parameters. The alignment of the reference consensus TE sequences on the genome was made using Blaster, CENSOR  and RepeatMasker (http://www.repeatmasker.org/). The results of the three methods were concatenated and MATCHER  was used to remove overlapping HSPs and make connections with the "join" procedure. Using in-house perl script, we retrieved TE annotation with >80% sequence identity to consensus sequences & length >150 nucleotides.
TE-related Pfam domains.
A published list of 124 Pfam domains associated with TE  was combined with the list of TE-related Pfam domains included in the LTR Digest software , resulting in a non-redundant list of 129 TE-related Pfam domains. The Meloidogyne species analyzed in the present article possessed 28 of these 129 TE-related domains. The compared abundance of these 28 domains in the three apomictic Meloidogyne is presented in S4 Table.
S1 Fig. Per base coverage distribution of Illumina PE-reads mapped on asexual Meloidogyne genomes.
Two peaks are observed. One corresponding to the average coverage of the genome and a second at ~ twice the coverage of the first peak. This second peak shows that a substantial portion of the genomes has a twice higher coverage. Sequences with this twice higher coverage may represent nearly identical duplicated sequences merged into single sequences in our assembly because they are almost identical.
S2 Fig. Distribution of % identity between pairs of duplicated regions in the asexual Meloidogyne.
Counts of pairs of duplicated genomic blocks (y-axis) as a function of their % identity (x-axis), in M. incognita (M. inc), M. javanica (M. jav) and M. arenaria (M. are).
S3 Fig. Vista plot of alignment of a pair of duplicated blocks taken as an example for M. incognita.
To illustrate the divergence between pair of duplicated collinear blocks formed by MCScanX, we aligned (using LAGAN: at http://genome.lbl.gov/vista/index.shtml), 2 collinear blocks of M. incognita. Block located on scaffold “Minc3s00001” from position 286,600 to 315,888 is aligned against its homoeolog block located on scaffold “Minc3s00990” from position 3,627 to 27,194. Percentage identity (between 50 and 100%) is plotted against the scaffold sequence (light grey line) for local alignment with 70% coverage and a minimum of 100 bp. Location of homologous genes is indicated by dark grey arrows. Identity is in salmon for conserved non-coding sequences (CNS) and in light slate blue for protein-coding sequences.
S4 Fig. Collinearity plotted against average Ks per pair of duplicated blocks.
For each pair of collinear blocks collinearity is computed as the fraction of collinear genes within a pair of blocks and Ks is obtained using the add_ka_and_ks_to_collinearity.pl script of the MCScanX package.
S5 Fig. Circos and corresponding phylogenies for genomic regions in one copy in M. hapla and two or more copies in asexual Meloidogyne.
In each Circo (unit = kb), color codes between collinear blocks are as follows. Collinear orthologs between M. hapla and any of the three asexuals species in grey. Collinear ‘homoeologs’ within asexual species in purple. Collinear orthologs between M. arenaria and M. javanica in green. Collinear orthologs between M. arenaria and M. incognita in yellow. Collinear orthologs between M. incognita and M. javanica in red. The outer scaled blue line represents gene density on scaffolds. The corresponding ML phylogenies performed on the concatenated alignments per blocks are given below each Circo.
S6 Fig. Mitochondrial phylogeny with branch lengths.
Mitochondrial phylogenetic tree of Meloidogyne with actual branch lengths showing high identity level between apomictic root-knot nematodes.
S7 Fig. Example of episodic diversifying selection detected in a phylogenetic tree.
Example of a duplicated gene copy harboring signature of episodic diversifying selection as detected by the Branch-Site REL model. The branch leading to the gene Mjav1s00402g005840 is inferred to be under (or to have underwent) episodic diversifying selection (corrected p = 0.013). The hue of each color indicates strength of selection, with primary red corresponding to ω > 5, primary blue to ω = 0 and grey to ω = 1. The width of each color component represents the proportion of sites in the corresponding class. Thicker branches have been classified as undergoing episodic diversifying selection by the sequential likelihood ratio test at corrected p ≤ 0.05. ω: the inferred rates of non-synonymous mutations / rate of synonymous mutations ratio.
S8 Fig. Relative DNA staining in nuclei of Meloidogyne spp.
Example of an unfiltered cytogram (arbitrary units) where nuclei were processed alone in a single acquisition: (A) M. incognita sample and (B) C. elegans standard. (C) Cytogram example obtained after gating on G0/G1 nuclei (arbitrary units) from each Meloidogyne species (M. hapla, M. incognita, M. javanica and M. arenaria) when processed mixed altogether with an internal standard (D. melanogaster: approximately 350 Mb).
S1 Text. Supplementary methods, results and discussion.
S1 Table. Species list and sequence accession number for the mitochondrial phylogeny.
S2 Table. Functional annotation of genes under positive selection (Ka/Ks >1).
S3 Table. Functional annotation of genes under episodic diversifying selection (EDS).
S4 Table. Abundance of Pfam domains associated to transposable elements.
S6 Table. Statistics of genome reads used for genome assemblies.
This work was developed in close collaboration with the Université Côte d'Azur functional genomics platform of Nice Sophia Antipolis (UCAGenomix), a partner of the National Infrastructure France Génomique. We thank A. Robichon, JJ. Remy and DA. Fernandes De Abreu for providing D. melanogaster and C. elegans used as standards for flow cytometry experiments, C. Kim-Jo for help in counting tree topologies and formatting supplementary figures, D. Pratella for initial analysis of RNAseq data. We thank V. Jamilloux and the REPET team at INRA Versailles for help in installing and configuring REPET. The authors are grateful to the GenoToul bioinformatics platform Toulouse Midi-Pyrenees for providing computing resources.
- Conceptualization: EGJD RBM LPB JMA.
- Data curation: EGJD RBM LPB MDR CR JFF JGo ES MBB JC.
- Formal analysis: RBM MBB MDR CR JGo ES TS DKK YJKJ.
- Funding acquisition: EGJD LPB PA PCS PW.
- Investigation: RBM CMJ MDR EGJD LPB JMA JGo ES MBB JFF DKK AC CDS JGu YJKJ CR TS.
- Methodology: RBM JFF JMA LPB JGo ES TS JC.
- Project administration: EGJD LPB PW.
- Resources: PW PCS JGo ES TS EGJD JC.
- Software: RBM MDR CR JGo ES JFF.
- Supervision: EGJD LPB JGo JMA.
- Validation: RBM LPB MDR CR MBB JFF JMA JG ES.
- Visualization: RBM EGJD LPB MBB JMA MDR.
- Writing – original draft: EGJD RBM.
- Writing – review & editing: EGJD RBM JFF PCS PA LPB.
- 1. Vrijenhoek RC. Animal Clones and Diversity Are natural clones generalists or specialists? BioScience. 1998;48: 617–628.
- 2. Rice WR. Experimental tests of the adaptive significance of sexual recombination. Nat Rev Genet. 2002;3: 241–251. pmid:11967549
- 3. Neiman M, Meirmans S, Meirmans PG. What Can Asexual Lineage Age Tell Us about the Maintenance of Sex? Ann N Y Acad Sci. 2009;1168: 185–200. pmid:19566708
- 4. Danchin E, Flot J-F, Perfus-Barbeoch L, Van Doninck K. Genomic perspectives on the long-term absence of sexual reproduction in animals. In: Pontarotti P, editor. Evolutionary Biology—Concepts, Biodiversity, Macroevolution and Genome Evolution. Berlin Heidelberg: Springer; 2011.
- 5. Flot J-F, Hespeels B, Li X, Noel B, Arkhipova I, Danchin EGJ, et al. Genomic evidence for ameiotic evolution in the bdelloid rotifer Adineta vaga. Nature. 2013;500: 453–457. pmid:23873043
- 6. Hill WG, Robertson A. The effect of linkage on limits to artificial selection. Genet Res. 1966;8: 269–294. pmid:5980116
- 7. Glémin S, Galtier N. Genome Evolution in Outcrossing Versus Selfing Versus Asexual Species. In: Anisimova M, editor. Evolutionary Genomics. Totowa, NJ: Humana Press; 2012. pp. 311–335. http://link.springer.com/10.1007/978-1-61779-582-4_11
- 8. Muller HJ. The Relation of Recombination to Mutational Advance. Mutat Res. 1964;106: 2–9. pmid:14195748
- 9. Kondrashov AS. Deleterious mutations and the evolution of sexual reproduction. Nature. 1988;336: 435–40. pmid:3057385
- 10. Paland S, Lynch M. Transitions to asexuality result in excess amino acid substitutions. Science. 2006;311: 990–2. pmid:16484491
- 11. Barraclough TG, Fontaneto D, Ricci C, Herniou EA. Evidence for inefficient selection against deleterious mutations in cytochrome oxidase I of asexual bdelloid rotifers. Mol Biol Evol. 2007;24: 1952–62. pmid:17573376
- 12. Neiman M, Hehman G, Miller JT, Logsdon JM, Taylor DR. Accelerated mutation accumulation in asexual lineages of a freshwater snail. Mol Biol Evol. 2010;27: 954–63. pmid:19995828
- 13. Henry L, Schwander T, Crespi BJ. Deleterious mutation accumulation in asexual Timema stick insects. Mol Biol Evol. 2012;29: 401–8. pmid:21940645
- 14. Hollister JD, Greiner S, Wang W, Wang J, Zhang Y, Wong GK-S, et al. Recurrent Loss of Sex Is Associated with Accumulation of Deleterious Mutations in Oenothera. Mol Biol Evol. 2015;32: 896–905. pmid:25534028
- 15. Schaack S, Choi E, Lynch M, Pritham EJ. DNA transposons and the role of recombination in mutation accumulation in Daphnia pulex. Genome Biol. 2010;11: R46. pmid:20433697
- 16. Ågren JA, Greiner S, Johnson MTJ, Wright SI. No evidence that sex and transposable elements drive genome size variation in evening primroses. Evolution. 2015;69: 1053–1062. pmid:25690700
- 17. Vrijenhoek RC, Parker ED. Geographical Parthenogenesis: General Purpose Genotypes and Frozen Niche Variation. In: Schön I, Martens K, Dijk P, editors. Lost Sex. Dordrecht: Springer Netherlands; 2009. pp. 99–131. Available: http://www.springerlink.com/index/10.1007/978-90-481-2770-2_6
- 18. Hörandl E. Geographical Parthenogenesis: Opportunities for Asexuality. In: Schön I, Martens K, Dijk P, editors. Lost Sex. Springer Netherlands; 2009. pp. 161–186. http://link.springer.com/chapter/10.1007/978-90-481-2770-2_8
- 19. Castagnone-Sereno P, Danchin EG, Perfus-Barbeoch L, Abad P. Diversity and Evolution of Root-Knot Nematodes, Genus Meloidogyne: New Insights from the Genomic Era. Annu Rev Phytopathol. 2013;51: 130517121443006. pmid:23682915
- 20. Jones JT, Haegeman A, Danchin EGJ, Gaur HS, Helder J, Jones MGK, et al. Top 10 plant-parasitic nematodes in molecular plant pathology. Mol Plant Pathol. 2013;14: 946–961. pmid:23809086
- 21. Castagnone-Sereno P, Danchin EGJ. Parasitic success without sex—the nematode experience. J Evol Biol. 2014;27: 1323–1333. pmid:25105196
- 22. Castagnone-Sereno P. Genetic variability and adaptive evolution in parthenogenetic root-knot nematodes. Heredity. 2006;96: 282–9. pmid:16404412
- 23. Abad P, Gouzy J, Aury J-M, Castagnone-Sereno P, Danchin EGJ, Deleury E, et al. Genome sequence of the metazoan plant-parasitic nematode Meloidogyne incognita. Nat Biotechnol. 2008;26: 909–915. pmid:18660804
- 24. Opperman CH, Bird DM, Williamson VM, Rokhsar DS, Burke M, Cohn J, et al. Sequence and genetic map of Meloidogyne hapla: A compact nematode genome for plant parasitism. Proc Natl Acad Sci U A. 2008;105: 14802–7.
- 25. Dalmasso A, Berge JB. Enzyme polymorphism and the concept of parthenogenetic species, exemplified by Meloidogyne. Stone AR, Platt HM, Khalil LF, editors. Concepts Nematode Syst Proc Intern Symp Camb Univ 2–4 Sept 1981 Syst Assoc Spec Vol 22. 1983; 187–196.
- 26. Hugall A, Stanton J, Moritz C. Reticulate evolution and the origins of ribosomal internal transcribed spacer diversity in apomictic Meloidogyne. Mol Biol Evol. 1999;16: 157–164. pmid:10028283
- 27. Lunt DH, Kumar S, Koutsovoulos G, Blaxter ML. The complex hybrid origins of the root knot nematodes revealed through comparative genomics. PeerJ. 2014;2: e356. pmid:24860695
- 28. Parra G, Bradnam K, Ning Z, Keane T, Korf I. Assessing the gene space in draft genomes. Nucleic Acids Res. 2009;37: 289–297. pmid:19042974
- 29. Wang Y, Tang H, Debarry JD, Tan X, Li J, Wang X, et al. MCScanX: a toolkit for detection and evolutionary analysis of gene synteny and collinearity. Nucleic Acids Res. 2012;40: e49. pmid:22217600
- 30. Murat F, Zhang R, Guizard S, Flores R, Armero A, Pont C, et al. Shared Subgenome Dominance Following Polyploidization Explains Grass Genome Evolutionary Plasticity from a Seven Protochromosome Ancestor with 16K Protogenes. Genome Biol Evol. 2014;6: 12–33. pmid:24317974
- 31. Krzywinski MI, Schein JE, Birol I, Connors J, Gascoyne R, Horsman D, et al. Circos: An information aesthetic for comparative genomics. Genome Res. 2009; pmid:19541911
- 32. Huskins CL. A cytological study of Vilmorin’s unfixable dwarf wheat. J Genet. 1931;25: 113–124.
- 33. Pond SLK, Murrell B, Fourment M, Frost SDW, Delport W, Scheffler K. A random effects branch-site model for detecting episodic diversifying selection. Mol Biol Evol. 2011; msr125. pmid:21670087
- 34. Stein LD, Bao Z, Blasiar D, Blumenthal T, Brent MR, Chen N, et al. The Genome Sequence of Caenorhabditis briggsae: A Platform for Comparative Genomics. PLoS Biol. 2003;1: e45. pmid:14624247
- 35. Mitreva M, Jasmer DP, Zarlenga DS, Wang Z, Abubucker S, Martin J, et al. The draft genome of the parasitic nematode Trichinella spiralis. Nat Genet. 2011;43: 228–235. pmid:21336279
- 36. Scott AL, Ghedin E. The genome of Brugia malayi—All worms are not created equal. Parasitol Int. 2009;58: 6–11. pmid:18952001
- 37. Kikuchi T, Cotton JA, Dalzell JJ, Hasegawa K, Kanzaki N, McVeigh P, et al. Genomic Insights into the Origin of Parasitism in the Emerging Plant Pathogen Bursaphelenchus xylophilus. PLoS Pathog. 2011;7: e1002219. pmid:21909270
- 38. Lapp NA, Triantaphyllou AC. Relative DNA Content and Chromosomal Relationships of some Meloidogyne, Heterodera, and Meloidodera spp. (Nematoda: Heteroderidae). J Nematol. 1972;4: 287–291. pmid:19319282
- 39. Pableo EC, Triantaphyllou AC. DNA Complexity of the Root-knot Nematode (Meloidogyne spp.) Genome. J Nematol. 1989;21: 260–3. pmid:19287606
- 40. Triantaphyllou AC. Cytogenetics, cytotaxonomy and phylogeny of root-knot nematodes. In: Sasser JN, Carter CC, editors. An advance treatise on Meloidogyne. Raleigh, USA: North Carolina State University Graphics; 1985. pp. 113–26.
- 41. Fargette M, Berthier K, Richaud M, Lollier V, Franck P, Hernandez A, et al. Crosses prior to parthenogenesis explain the current genetic diversity of tropical plant-parasitic Meloidogyne species (Nematoda: Tylenchida). Infect Genet Evol. 2010;10: 806–813. pmid:19393769
- 42. Lunt DH. Genetic tests of ancient asexuality in Root Knot Nematodes reveal recent hybrid origins. BMC Evol Biol. 2008;8: 194. pmid:18606000
- 43. Denver DR, Morris K, Lynch M, Vassilieva LL, Thomas WK. High direct estimate of the mutation rate in the mitochondrial genome of Caenorhabditis elegans. Science. 2000;289: 2342–4. pmid:11009418
- 44. Denver DR, Morris K, Lynch M, Thomas WK. High mutation rate and predominance of insertions in the Caenorhabditis elegans nuclear genome. Nature. 2004;430: 679–682. pmid:15295601
- 45. Denver DR, Dolan PC, Wilhelm LJ, Sung W, Lucas-Lledó JI, Howe DK, et al. A genome-wide view of Caenorhabditis elegans base-substitution mutation processes. Proc Natl Acad Sci. 2009;106: 16310–16314. pmid:19805298
- 46. Lampert KP, Schartl M. A little bit is better than nothing: the incomplete parthenogenesis of salamanders, frogs and fish. BMC Biol. 2010;8: 78. pmid:20687905
- 47. Inoue J, Sato Y, Sinclair R, Tsukamoto K, Nishida M. Rapid genome reshaping by multiple-gene loss after whole-genome duplication in teleost fish suggested by mathematical modeling. Proc Natl Acad Sci. 2015;112: 14918–14923. pmid:26578810
- 48. Abbott RJ, Hegarty MJ, Hiscock SJ, Brennan AC. Homoploid hybrid speciation in action. Taxon. 2010;59: 1375–1386.
- 49. Mavárez J, Linares M. Homoploid hybrid speciation in animals. Mol Ecol. 2008;17: 4181–4185. pmid:19378399
- 50. Mason AS, Pires JC. Unreduced gametes: meiotic mishap or evolutionary mechanism? Trends Genet. 2015;31: 5–10. pmid:25445549
- 51. Ramsey J, Schemske DW. Pathways, mechanisms, and rates of polyploid formation in flowering plants. Annu Rev Ecol Syst. 1998;29: 467–501.
- 52. Yamauchi A, Hosokawa A, Nagata H, Shimoda M. Triploid Bridge and Role of Parthenogenesis in the Evolution of Autopolyploidy. Am Nat. 2004;164: 101–112. pmid:15266374
- 53. Yahara T. Evolution of Agamospermous Races in Boehmeria and Eupatorium. Plant Species Biol. 1990;5: 183–196.
- 54. Rushworth CA, Song B-H, Lee C-R, Mitchell-Olds T. Boechera, a model system for ecological genomics. Mol Ecol. 2011;20: 4843–4857. pmid:22059452
- 55. Lutes AA, Neaves WB, Baumann DP, Wiegraebe W, Baumann P. Sister chromosome pairing maintains heterozygosity in parthenogenetic lizards. Nature. 2010;464: 283–286. pmid:20173738
- 56. Cole CJ, Taylor HL, Baumann DP, Baumann P. Neaves’ whiptail lizard: the first known tetraploid parthenogenetic tetrapod (Reptilia: Squamata: Teiidae). 2014; https://www.bioone.org/doi/full/10.3099/MCZ17.1
- 57. Dolgin ES, Charlesworth B. The Fate of Transposable Elements in Asexual Populations. Genetics. 2006;174: 817–827. pmid:16888330
- 58. Kraaijeveld K, Zwanenburg B, Hubert B, Vieira C, De Pater S, Van Alphen JJM, et al. Transposon proliferation in an asexual parasitoid. Mol Ecol. 2012;21: 3898–3906. pmid:22548357
- 59. Bast J, Schaefer I, Schwander T, Maraun M, Scheu S, Kraaijeveld K. No accumulation of transposable elements in asexual arthropods. Mol Biol Evol. 2015; msv261. pmid:26560353
- 60. Leroy H, Leroy F, Augé-Gouillou C, Castagnone-Sereno P, Vanlerberghe-Masutti F, Bigot Y, et al. Identification of mariner-like elements from the root-knot nematode Meloidogyne spp. Mol Biochem Parasitol. 2000;107: 181–190. pmid:10779595
- 61. Gross SM, Williamson VM. Tm1: A Mutator/Foldback Transposable Element Family in Root-Knot Nematodes. PLoS ONE. 2011;6: e24534. pmid:21931741
- 62. Trudgill DL, Blok VC. Apomictic, polyphagous root-knot nematodes: exceptionally successful and damaging biotrophic root pathogens. Annu Rev Phytopathol. 2001;39: 53–77. pmid:11701859
- 63. Perry RN, Moens M, Starr JL, editors. Root-knot nematodes. Cambridge, MA: CABI North American Office; 2009.
- 64. Cuypers TD, Hogeweg P. A Synergism between Adaptive Effects and Evolvability Drives Whole Genome Duplication to Fixation. PLoS Comput Biol. 2014;10: e1003547. pmid:24743268
- 65. Soltis DE, Visger CJ, Soltis PS. The polyploidy revolution then…and now: Stebbins revisited. Am J Bot. 2014;101: 1057–1078. pmid:25049267
- 66. Selmecki AM, Maruvka YE, Richmond PA, Guillet M, Shoresh N, Sorenson AL, et al. Polyploidy can drive rapid adaptation in yeast. Nature. 2015;519: 349–352. pmid:25731168
- 67. Madlung A. Polyploidy and its effect on evolutionary success: old questions revisited with new tools. Heredity. 2013;110: 99–104. pmid:23149459
- 68. Dittrich-Reed DR, Fitzpatrick BM. Transgressive Hybrids as Hopeful Monsters. Evol Biol. 2012;40: 310–315. pmid:23687396
- 69. Fitzpatrick BM, Shaffer HB. Hybrid vigor between native and introduced salamanders raises new challenges for conservation. Proc Natl Acad Sci. 2007;104: 15793–15798. pmid:17884982
- 70. Ryan ME, Johnson JR, Fitzpatrick BM. Invasive hybrid tiger salamander genotypes impact native amphibians. Proc Natl Acad Sci. 2009;106: 11166–11171. pmid:19564601
- 71. Lynch M, Conery JS. The evolutionary fate and consequences of duplicate genes. Science. 2000;290: 1151–5. pmid:11073452
- 72. Ohno S. Evolution by gene duplication. 1St Edition. Springer-Verlag; 1970.
- 73. Lien S, Koop BF, Sandve SR, Miller JR, Kent MP, Nome T, et al. The Atlantic salmon genome provides insights into rediploidization. Nature. 2016;533: 200–205. pmid:27088604
- 74. Session AM, Uno Y, Kwon T, Chapman JA, Toyoda A, Takahashi S, et al. Genome evolution in the allotetraploid frog Xenopus laevis. Nature. 2016;538: 336–343. pmid:27762356
- 75. Schwarz D, Matta BM, Shakir-Botteri NL, McPheron BA. Host shift to an invasive plant triggers rapid animal hybrid speciation. Nature. 2005;436: 546–549. pmid:16049486
- 76. Menardo F, Praz CR, Wyder S, Ben-David R, Bourras S, Matsumae H, et al. Hybridization of powdery mildew strains gives rise to pathogens on novel agricultural crop species. Nat Genet. 2016;48: 201–205. pmid:26752267
- 77. de Jonge R, Bolton MD, Kombrink A, van den Berg GCM, Yadeta KA, Thomma BPHJ. Extensive chromosomal reshuffling drives evolution of virulence in an asexual pathogen. Genome Res. 2013;23: 1271–1282. pmid:23685541
- 78. Chevreux B, Wetter T, Suhai S. Genome Sequence Assembly Using Trace Signals and Additional Sequence Information. Computer Science and Biology: Proceedings of the German Conference on Bioinformatics. 1999. pp. 45–56.
- 79. Aury J-M, Cruaud C, Barbe V, Rogier O, Mangenot S, Samson G, et al. High quality draft sequences for prokaryotic genomes using a mix of new sequencing technologies. BMC Genomics. 2008;9: 603. pmid:19087275
- 80. Boetzer M, Henkel CV, Jansen HJ, Butler D, Pirovano W. Scaffolding pre-assembled contigs using SSPACE. Bioinformatics. 2011;27: 578–579. pmid:21149342
- 81. Luo R, Liu B, Xie Y, Li Z, Huang W, Yuan J, et al. SOAPdenovo2: an empirically improved memory-efficient short-read de novo assembler. GigaScience. 2012;1: 18. pmid:23587118
- 82. Parra G, Bradnam K, Korf I. CEGMA: a pipeline to accurately annotate core genes in eukaryotic genomes. Bioinformatics. 2007;23: 1061–1067. pmid:17332020
- 83. The C. elegans Sequencing Consortium. Genome Sequence of the Nematode C. elegans: A Platform for Investigating Biology. Science. 1998;282: 2012–2018. pmid:9851916
- 84. Hillier LW, Coulson A, Murray JI, Bao Z, Sulston JE, Waterston RH. Genomics in C. elegans: so many genes, such a little worm. Genome Res. 2005;15: 1651–60. pmid:16339362
- 85. Bosco G, Campbell P, Leiva-Neto JT, Markow TA. Analysis of Drosophila Species Genome Size and Satellite DNA Content Reveals Significant Differences Among Strains as Well as Between Species. Genetics. 2007;177: 1277–1290. pmid:18039867
- 86. Gregory TR, Nathwani P, Bonnett TR, Huber DPW. Sizing up arthropod genomes: an evaluation of the impact of environmental variation on genome size estimates by flow cytometry and the use of qPCR as a method of estimation. Genome. 2013;56: 505–510. pmid:24168671
- 87. Perfus-Barbeoch L, Castagnone-Sereno P, Reichelt M, Fneich S, Roquis D, Pratx L, et al. Elucidating the molecular bases of epigenetic inheritance in non-model invertebrates: the case of the root-knot nematode Meloidogyne incognita. Invertebr Physiol. 2014;5: 211. pmid:24936189
- 88. DeSalle R, Gregory TR, Johnston JS. Preparation of Samples for Comparative Studies of Arthropod Chromosomes: Visualization, In Situ Hybridization, and Genome Size Estimation. In: Enzymology B-M in, editor. Academic Press; 2005. pp. 460–488. http://www.sciencedirect.com/science/article/pii/S0076687905950258
- 89. Hare EE, Johnston JS. Genome size determination using flow cytometry of propidium iodide-stained nuclei. Methods Mol Biol Clifton NJ. 2011;772: 3–12. pmid:22065429
- 90. Li H, Durbin R. Fast and accurate short read alignment with Burrows–Wheeler transform. Bioinformatics. 2009;25: 1754–1760. pmid:19451168
- 91. Quinlan AR, Hall IM. BEDTools: a flexible suite of utilities for comparing genomic features. Bioinformatics. 2010;26: 841–842. pmid:20110278
- 92. Foissac S, Gouzy J, Rombauts S, Mathe C, Amselem J, Sterck L, et al. Genome Annotation in Plants and Fungi: EuGene as a Model Platform. Curr Bioinforma. 2008;3: 87–97.
- 93. Degroeve S, Saeys Y, De Baets B, Rouze P, Van de Peer Y. SpliceMachine: predicting splice sites from high-dimensional local context representations. Bioinformatics. 2005;21: 1332–8. pmid:15564294
- 94. Danchin EGJ, Arguel M-J, Campan-Fournier A, Perfus-Barbeoch L, Magliano M, Rosso M-N, et al. Identification of Novel Target Genes for Safer and More Specific Control of Root-Knot Nematodes from a Pan-Genome Mining. PLoS Pathog. 2013;9: e1003745. pmid:24204279
- 95. Haas BJ, Papanicolaou A, Yassour M, Grabherr M, Blood PD, Bowden J, et al. De novo transcript sequence reconstruction from RNA-seq using the Trinity platform for reference generation and analysis. Nat Protoc. 2013;8: 1494–1512. pmid:23845962
- 96. Wu TD, Watanabe CK. GMAP: a genomic mapping and alignment program for mRNA and EST sequences. Bioinformatics. 2005;21: 1859–1875. pmid:15728110
- 97. Cotton JA, Lilley CJ, Jones LM, Kikuchi T, Reid AJ, Thorpe P, et al. The genome and life-stage specific transcriptomes of Globodera pallida elucidate key aspects of plant parasitism by a cyst nematode. Genome Biol. 2014;15: R43. pmid:24580726
- 98. Jurka J, Kapitonov VV, Pavlicek A, Klonowski P, Kohany O, Walichiewicz J. Repbase Update, a database of eukaryotic repetitive elements. Cytogenet Genome Res. 2005;110: 462–7. pmid:16093699
- 99. Altschul SF, Madden TL, Schaffer AA, Zhang J, Zhang Z, Miller W, et al. Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res. 1997;25: 3389–402. pmid:9254694
- 100. Lagesen K, Hallin P, Rødland EA, Stærfeldt H-H, Rognes T, Ussery DW. RNAmmer: consistent and rapid annotation of ribosomal RNA genes. Nucleic Acids Res. 2007;35: 3100–3108. pmid:17452365
- 101. Lowe TM, Eddy SR. tRNAscan-SE: A Program for Improved Detection of Transfer RNA Genes in Genomic Sequence. Nucleic Acids Res. 1997;25: 0955–0964.
- 102. Burge SW, Daub J, Eberhardt R, Tate J, Barquist L, Nawrocki EP, et al. Rfam 11.0: 10 years of RNA families. Nucleic Acids Res. 2013;41: D226–D232. pmid:23125362
- 103. Finn RD, Bateman A, Clements J, Coggill P, Eberhardt RY, Eddy SR, et al. Pfam: the protein families database. Nucleic Acids Res. 2014;42: D222–D230. pmid:24288371
- 104. Mitchell A, Chang H-Y, Daugherty L, Fraser M, Hunter S, Lopez R, et al. The InterPro protein families database: the classification resource after 15 years. Nucleic Acids Res. 2015;43: D213–D221. pmid:25428371
- 105. McCarthy FM, Wang N, Magee GB, Nanduri B, Lawrence ML, Camon EB, et al. AgBase: a functional genomics resource for agriculture. BMC Genomics. 2006;7: 229. pmid:16961921
- 106. Kurtz S, Phillippy A, Delcher AL, Smoot M, Shumway M, Antonescu C, et al. Versatile and open software for comparing large genomes. Genome Biol. 2004;5: R12. pmid:14759262
- 107. Edgar RC. MUSCLE: a multiple sequence alignment method with reduced time and space complexity. BMC Bioinformatics. 2004;5: 113. pmid:15318951
- 108. Edgar RC. MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Res. 2004;32: 1792–7. pmid:15034147
- 109. Suyama M, Torrents D, Bork P. PAL2NAL: robust conversion of protein sequence alignments into the corresponding codon alignments. Nucleic Acids Res. 2006;34: W609–W612. pmid:16845082
- 110. Castresana J. Selection of Conserved Blocks from Multiple Alignments for Their Use in Phylogenetic Analysis. Mol Biol Evol. 2000;17: 540–552. pmid:10742046
- 111. Schliep KP. phangorn: phylogenetic analysis in R. Bioinformatics. 2011;27: 592–593. pmid:21169378
- 112. Guindon S, Dufayard J-F, Lefort V, Anisimova M, Hordijk W, Gascuel O. New Algorithms and Methods to Estimate Maximum-Likelihood Phylogenies: Assessing the Performance of PhyML 3.0. Syst Biol. 2010;59: 307–321. pmid:20525638
- 113. Zhang H, Gao S, Lercher MJ, Hu S, Chen W-H. EvolView, an online tool for visualizing, annotating and managing phylogenetic trees. Nucleic Acids Res. 2012;40: W569–W572. pmid:22695796
- 114. Nei M, Gojobori T. Simple methods for estimating the numbers of synonymous and nonsynonymous nucleotide substitutions. Mol Biol Evol. 1986;3: 418–426. pmid:3444411
- 115. Pond SLK, Frost SDW, Muse SV. HyPhy: hypothesis testing using phylogenies. Bioinformatics. 2005;21: 676–679. pmid:15509596
- 116. Kopylova E, Noé L, Touzet H. SortMeRNA: fast and accurate filtering of ribosomal RNAs in metatranscriptomic data. Bioinforma Oxf Engl. 2012;28: 3211–3217. pmid:23071270
- 117. Schmieder R, Edwards R. Quality control and preprocessing of metagenomic datasets. Bioinforma Oxf Engl. 2011;27: 863–864. pmid:21278185
- 118. Dobin A, Davis CA, Schlesinger F, Drenkow J, Zaleski C, Jha S, et al. STAR: ultrafast universal RNA-seq aligner. Bioinformatics. 2013;29: 15–21. pmid:23104886
- 119. Li B, Dewey CN. RSEM: accurate transcript quantification from RNA-Seq data with or without a reference genome. BMC Bioinformatics. 2011;12: 323. pmid:21816040
- 120. Smit AFA, Hubley R, Green P. RepeatMasker Open-4.0. In: RepeatMasker [Internet]. 2015 2013. http://www.repeatmasker.org
- 121. Flutre T, Duprat E, Feuillet C, Quesneville H. Considering transposable element diversification in de novo annotation approaches. PLoS One. 2011;6: e16526. pmid:21304975
- 122. Quesneville H, Nouaud D, Anxolabéhère D. Detection of new transposable element families in Drosophila melanogaster and Anopheles gambiae genomes. J Mol Evol. 2003;57 Suppl 1: S50–59. pmid:15008403
- 123. Ellinghaus D, Kurtz S, Willhoeft U. LTRharvest, an efficient and flexible software for de novo detection of LTR retrotransposons. BMC Bioinformatics. 2008;9: 18. pmid:18194517
- 124. Bao Z, Eddy SR. Automated de novo identification of repeat sequence families in sequenced genomes. Genome Res. 2002;12: 1269–1276. pmid:12176934
- 125. Edgar RC, Myers EW. PILER: identification and classification of genomic repeats. Bioinformatics. 2005;21 Suppl 1: i152–8.
- 126. Huang X. On global sequence alignment. Comput Appl Biosci. 1994;10: 227–35. pmid:7922677
- 127. Hoede C, Arnoux S, Moisset M, Chaumier T, Inizan O, Jamilloux V, et al. PASTEC: An Automatic Transposable Element Classification Tool. PLoS ONE. 2014;9: e91929. pmid:24786468
- 128. Jurka J, Klonowski P, Dagman V, Pelton P. Censor—A program for identification and elimination of repetitive elements from DNA sequences. Comput Chem. 1996;20: 119–121. pmid:8867843
- 129. Piriyapongsa J, Rutledge MT, Patel S, Borodovsky M, Jordan IK. Evaluating the protein coding potential of exonized transposable element sequences. Biol Direct. 2007;2: 31. pmid:18036258
- 130. Steinbiss S, Willhoeft U, Gremme G, Kurtz S. Fine-grained annotation and classification of de novo predicted LTR retrotransposons. Nucleic Acids Res. 2009;37: 7002–7013. pmid:19786494