Modern Sciences is a premier science journal that bridges the gap between science and its application to society.
LATEST STORIES

Active and unusually expanded PIF/Harbinger transposable elements in the Caenorhabditis inopinata genome

Kazuki Sato1†, Xiaodan Jin1†, Shun Oomura2†, Kazuma Kawahara2, Simo Sun13, Akemi Yoshida4, Nami Haruta2, Asako Sugimoto2 and Taisei Kikuchi1*

    1Laboratory of Parasite Systems Biology, Department of Integrated Biosciences, Graduate School of Frontier Sciences, The University of Tokyo, Kashiwa, Chiba 277-8562, Japan

    2Laboratory of Developmental Dynamics, Graduate School of Life Sciences, Tohoku University, Sendai, 980-8577, Japan

    3Chinese Institute for Brain Research, Beijing 102206, China

    4Genomics and Bioenvironmental Science, Frontier Science Research Center, University of Miyazaki, Miyazaki, 889-1692, Japan

    *Corresponding author: taisei.kikuchi{at}edu.k.u-tokyo.ac.jp

      bioRxiv preprint DOI: https://doi.org/10.64898/2026.05.26.728016

      Posted: May 29, 2026, Version 1

      Copyright: † These authors contributed equally to this study.

      Abstract

      Background Transposable elements (TEs) serve as powerful drivers of genome innovation but also threaten genome integrity. The PIF/Harbinger superfamily is distinctive among DNA transposons because mobilisation typically requires proteins, a DDE transposase and a MADF DNA-binding protein. Caenorhabditis inopinata, the closest known relative of C. elegans, has a TE-rich genome and lacks multiple components of the ERGO-1–class endogenous small-RNA pathway, making it a useful system for examining TE dynamics in a distinct host context. We identified a spontaneous dumpy mutant of C. inopinata caused by insertion of a PIF/Harbinger-family element into the coding region of Cin-dpy-11. The inserted element, designated Harbinger-1M_cIno, belongs to the Turmoil2 lineage originally defined in C. elegans and retains a MADF domain but lacks a recognisable DDE transposase ORF. Genome-wide curation recovered 258 related copies, revealing a strongly asymmetric family structure. Short noncoding derivatives were predominant, MADF-bearing derivatives were expanded and only one DDE-bearing locus retained an apparently intact transposase gene, suggesting that DDE and MADF functions are partitioned across distinct elements and may be supplied in trans during mobilisation. We also identified a second PIF/Harbinger-derived family, Harbinger-2M_cIno, associated with the Turmoil1 lineage. This family comprises 1,376 copies and therefore records substantial past amplification, but it lacks a detectable DDE source, shows greater sequence divergence and more degraded terminal structures than Harbinger-1M_cIno. Together, these data indicate that the two PIF/Harbinger lineages in C. inopinata differ not in whether amplification occurred, but in when it occurred and whether present-day mobilisation competence has been retained.

      Introduction

      Transposable elements (TEs) are ubiquitous in eukaryotic genomes and are major drivers of genome evolution (Bourque et al., 2018). While their mobilisation can compromise genome integrity and fitness, TEs also generate regulatory sequences, promote genome rearrangements, and contribute to coding and noncoding innovation.

      Two major mechanistic classes of TEs are generally recognised. Class I elements (retrotransposons) mobilise via an RNA intermediate and are therefore often described as “copy-and-paste” elements. Class II elements (DNA transposons) mobilise as DNA, and many transpose through a canonical “cut-and-paste” mechanism. In canonical DNA transposons, the transposase recognises terminal inverted repeats (TIRs), excises the transposon sequence, and reinserts it elsewhere in the genome, typically generating short target-site duplications (TSDs) upon integration (Wicker et al., 2007). Transposable elements can broadly be classified as autonomous or non-autonomous according to whether they encode the machinery required for their own mobilisation. This distinction applies to both retrotransposons and DNA transposons, although the enzymatic machinery involved differs between them. Within DNA transposons, autonomous elements encode a transposase, whereas non-autonomous elements depend on trans-acting transposase supplied by related autonomous elements. Miniature inverted-repeat transposable elements (MITEs) are short, non-autonomous derivatives of TIR DNA transposons that lack a protein-coding open reading frame (ORF) but retain the terminal features required for mobilisation. They are generally thought to arise through internal deletion of autonomous progenitors, with abortive or interrupted gap repair representing one commonly proposed mechanism. Because direct sequence similarity to autonomous partners is often restricted to the terminal regions, the evolutionary origin of individual MITE families can be difficult to resolve. Once formed, however, such compact derivatives may undergo substantial copy-number amplification in the presence of a compatible trans-acting transposase (Rubin and Levy, 1997; Fattash et al., 2013).

      The PIF/Harbinger superfamily is widely distributed across eukaryotes. Harbinger elements were initially identified in plant genome analyses, including Arabidopsis thaliana (Kapitonov and Jurka, 1999), and were later unified with maize PIF elements within a single PIF/Harbinger superfamily. In nematodes, related elements were first characterised in C. elegans as Ce-PIF1 and Ce-PIF2 and were subsequently linked to the Turmoil elements (Jurka and Kapitonov, 2001; Zhang et al., 2001).

      PIF/Harbinger elements typically carry TIRs of highly variable length and generate 3-bp TSDs at a characteristic TWA target site upon insertion (Han et al., 2015). A characteristic feature of autonomous PIF/Harbinger elements is their two-gene organisation: one gene encodes a DDE transposase, typically with an associated helix-turn-helix-related DNA-binding region, and the other encodes a Myb/SANT-like DNA-binding protein, often referred to as a MADF- or trihelix-like factor (Wicker et al., 2007). Available genetic and heterologous reconstitution studies indicate that both proteins are required for efficient PIF/Harbinger transposition (Yang et al., 2007; Sinzelle et al., 2008; Hancock et al., 2010; Liu et al., 2024). In experimentally tractable systems, this Myb-like partner binds terminal or subterminal transposon sequences, interacts with the transposase, and promotes its accumulation in the nucleus, supporting assembly of a functional transposition complex at the element ends.

      In animals, TE mobilisation is typically restrained by host defence systems, particularly small-RNA pathways (Siomi et al., 2011; Iwasaki et al., 2025). In C. elegans, piRNAs and endogenous siRNAs act together to suppress TE expression, largely through secondary 22G-RNA amplification and WAGO-mediated silencing, with the ERGO-1–class 26G pathway providing an additional upstream trigger on selected targets (Das et al., 2008; Guang et al., 2008; Buckley et al., 2012; Billi et al., 2014). By contrast, C. inopinata—the closest known relative of C. elegans—has a larger, TE-enriched genome and lacks several genes of the ERGO-1–class 26G-siRNA pathway, including ergo-1, eri-6/7, and eri-9 (Kanzaki et al., 2018). Its genome also bears clear signatures of TE-associated restructuring, including disruption of the her-1 orthologue by an LTR insertion (Kanzaki et al., 2018). Although these observations do not on their own establish globally elevated ongoing TE activity, recent work has demonstrated active hAT-family transposition in C. inopinata, including repeated mobilisation of Ci-hAT1 and expansion of related non-autonomous derivatives (Hatanaka et al., 2024).

      C. inopinata therefore provides a valuable comparative framework for examining how divergence in small-RNA pathways and TE dynamics shapes genome structure and function in Caenorhabditis.

      Here, we analysed a spontaneously arising dumpy (Dpy) mutant of C. inopinata and identified a novel insertion of a PIF/Harbinger DNA transposon, providing direct evidence of ongoing mobilisation. The newly inserted copy retains an intact MADF-like DNA-binding domain but lacks a DDE transposase ORF, indicating that it is a derivative in which only the MADF module has been retained. Genome-wide analysis revealed more than twenty highly similar copies of this MADF-only form, as well as a distinct locus encoding a DDE transposase but lacking a cognate MADF ORF, a genomic configuration consistent with mobilisation in trans. Phylogenetic and structural analyses place these MADF-only derivatives within the Turmoil2 lineage. Turmoil1-related sequences are also present in the genome; however, no Turmoil1-type DDE transposase was detected, and the remaining MADF-containing loci have degraded TIRs, together suggesting that this lineage is no longer mobile. Thus, C. inopinata contains two Turmoil-related lineages with contrasting current states: a Turmoil2 lineage that retains evidence of ongoing mobilisation and a Turmoil1 lineage that appears to have degenerated. This contrast provides a tractable system for investigating how the modular MADF–DDE architecture influences lineage-specific TE evolution.

      Materials and Methods

      Nematode strains and maintenance

      C. inopinata was maintained under modified C. elegans growth conditions as previously described (Kanzaki et al., 2018; Oomura et al., 2022). A spontaneous Dpy mutant, SA1635 [tj217/tj217], was isolated from laboratory cultures of the wild-type strain NKZ35 based on its shortened, stout body morphology at the L4 or adult stage. Because homozygous progeny from homozygous mothers were sterile, the allele was maintained in the heterozygous state by crossing to wild type. A wild-type population (SA1689), derived within several generations of the original isolate, was used as the control. Images were acquired directly from NGM plates with a Nikon SMZ1270 stereomicroscope and a Leica MC170 HD camera using LAS-EZ software.

      Whole-genome sequencing and variant calling

      Genomic DNA was extracted from single adult males of SA1635 and SA1689, with five biological replicates per strain. DNA from individual worms was used to prepare sequencing libraries with the Illumina DNA Prep Kit (Illumina). Libraries were sequenced on an Illumina NovaSeq 6000 platform to generate 150-bp paired-end reads.

      Reads were trimmed to remove adapter and low-quality sequences and mapped to the C. inopinata reference genome (v7.11) with SMALT v0.7.4 (https://www.sanger.ac.uk/resources/software/smalt/). Alignments were processed with SAMtools (Danecek et al., 2021) and Picard tools (http://broadinstitute.github.io/picard/). Small variants were called with GATK HaplotypeCaller (https://gatk.broadinstitute.org/hc/en-us) using filtering criteria of QUAL > 30 and depth ≥ 10×, and structural variants were identified using Manta v1.3.1 (Chen et al., 2016). Variants were annotated with ANNOVAR (Wang et al., 2010). Variants of interest were visualised in Artemis (Carver et al., 2012) to confirm breakpoint structures.

      PCR validation of the Cin-dpy-11 insertion

      Genomic DNA from SA1635 and SA1689 was amplified with primers located in intron 2 and the 3′ UTR of Cin-dpy-11 (forward, 5′-gaagcagtcgtgaaaacaagagagg-3′; reverse, 5′-gcatatcacaggtacaatcccgtc-3′). The expected wild-type product size was 975 bp. When SA1635 genomic DNA was used as the template, an approximately 1-kb PCR artefact was also detected, likely because annealing between TIR sequences formed a hairpin-like secondary structure, resulting in skipping of the TE insertion. PCR products were analysed by Sanger sequencing. Sanger reads were aligned to the Cin-dpy-11 genomic region to define the breakpoint structure and insertion coordinates by confirming TIR sequences at both ends.

      Repeat identification

      RepeatMasker v4.1.8 was run with the parameters -a -s -e rmblast on C. inopinata and other Caenorhabditis genomes using either a C. inopinata-specific TE library (Kawahara et al., 2023) or a 14-species Caenorhabditis TE library (Sun et al., 2022). Assemblies were obtained from WormBase Parasite release 19 for C. elegans (PRJNA13758), C. nigoni (PRJNA384657), C. sinica (PRJNA194557), C. brenneri (PRJNA20035), C. remanei (PRJNA577507), C. quiockensis (PRJEB11354), C. latens (PRJNA248912), C. angaria (PRJNA51225), C. bovis (PRJEB34497), C. briggsae (PRJNA784955), C. niphades (PRJEB53466), and C. drosophilae (PRJEB67483). The C. japonica v2 assembly was generated in our laboratory.

      Self-alignments were generated with LAST v1542-1 (Camacho et al., 2009) using lastdb – uNEAR and lastal. MAF (Multiple Alignment Format) output was parsed and visualised with ggplot2 v4.0.0 (https://ggplot2.tidyverse.org/) implemented in R v4.4.1 (https://www.r-project.org/). Forward–forward matches were plotted along the main diagonal, whereas forward–reverse matches were plotted along the anti-diagonal.

      Genome screening for PIF/Harbinger elements

      PIF/Harbinger elements were identified in the C. inopinata genome with seqkit locate (Shen et al., 2016) by searching for lineage-specific 15-bp TIR sequences, allowing up to two mismatches (-m 2). Candidate elements were defined as oppositely oriented TIR-like pairs located within 4 kb of each other. Candidates lacking identical 3-bp flanking sequences at both ends were excluded. When nested or overlapping pairs were detected within the same interval, the shortest candidate was retained. Terminal sequences of curated elements were aligned with MAFFT v7.505 (Katoh and Standley, 2013) (–auto –adjustdirection) and used to summarise TIR motifs with WebLogo v3.7.9 (Crooks et al., 2004). Terminal windows comprised 80 nt from each end for Turmoil2-like elements and 30 nt from each end for Turmoil1-like elements. TSD motifs were summarised from the 3-bp flanking sequences.

      ORFs within identified PIF/Harbinger elements were predicted with EMBOSS getorf (v6.6.0.0) (Rice et al., 2000). Predicted proteins were scanned with pfam_scan.pl (v1.6.4) (Mistry et al., 2021) for DDE transposase domains (DDE_Tnp_4; PF13359.11) and MADF DNA-binding domains (MADF_DNA_bdg; PF10545.14). Elements were classified as MADF-only when a MADF_DNA_bdg-containing ORF was detected in the absence of a DDE_Tnp_4-containing ORF, as DDE-only when the reciprocal pattern was observed, and as non-autonomous MITEs when neither domain was detected.

      Clustering analysis of PIF/Harbinger elements

      Exact duplicate sequences were removed with vsearch v2.18.0 (Rognes et al., 2016) using –derep_fulllength –sizeout –minuniquesize 1, and dereplicated sequences were clustered at 80% identity with –cluster_size –id 0.80 –strand both –sizein. Clusters containing at least 10 sequences were retained. Sequences within each cluster were aligned with MAFFT v7.505 (Katoh and Standley, 2013) (–auto –adjustdirection) and trimmed with trimAl v1.4 (Capella-Gutierrez et al., 2009) (-gt 0.8). Pairwise nucleotide divergence was calculated under the Kimura two-parameter model with EMBOSS distmat (v6.6.0.0) (Rice et al., 2000) (-nucmethod 2). Summary statistics were calculated from unique pairwise distances in the upper triangle of each distance matrix. The weighted median element length for each cluster was calculated from dereplication copy counts. Sequence similarity to Ci_Harb_255 was assessed by BLASTN with BLAST+ v2.12.0 (Camacho et al., 2009) using -dust no -evalue 1e-5 -word_size 10 -outfmt 6.

      Phylogenetic analysis

      Amino-acid sequences were aligned with MAFFT v7.505 (–auto) (Katoh and Standley, 2013) and trimmed with trimAl v1.4 (-automated1) (Capella-Gutierrez et al., 2009). Maximum-likelihood phylogenies were inferred with IQ-TREE v2.1.4 (Minh et al., 2020), with substitution models selected by ModelFinder (Kalyaanamoorthy et al., 2017) and branch support assessed using 1,000 SH-aLRT replicates and 1,000 ultrafast bootstrap replicates. Sequences used in the analysis are listed in Supplementary Table 1. For conserved-domain comparisons within Caenorhabditis, Pfam-annotated domain regions were extracted from the protein sequences, aligned with MAFFT (–auto), and visualised with ggmsa v1.10.0 (Zhou et al., 2022) in R v4.4.1.

      RNA-seq analysis

      RNA-seq data from six libraries (two biological replicates each for egg, male, and female samples) were retrieved from GenBank (BioProject PRJDB5687). Reads were aligned to the C. inopinata reference genome using STAR (v2.7.11b) (Dobin et al., 2013) with a curated gene annotation (v7.11; Supplementary File 1). To retain multi-mapping reads, particularly those derived from repetitive TE-related sequences, mapping thresholds were relaxed (–outFilterMultimapNmax 200 and –winAnchorMultimapNmax 200). Read counts were quantified using TEcount (TEtranscripts v2.2.3) (Jin et al., 2015), which incorporates multi-mapping reads during expression estimation. Although TE annotations generated from RepeatMasker output using a C. inopinata-specific TE library were used during quantification, only gene-level count data were used in the present analyses. Raw gene count tables were imported into R, where library sizes were normalised using calcNormFactors (edgeR v4.2.2), and expression levels were converted to counts per million (CPM).

      MADF domain diversity

      MADF-domain regions were extracted from protein sequences, aligned with MAFFT v7.505 (– localpair –maxiterate 1000), and trimmed with trimAl (v1.4; -automated1) to obtain domain-only alignments. Pairwise protein distances under the Kimura model were computed using PHYLIP protdist (PHYLIPNEW v3.69.650; EMBOSS v6.5.7.0) (Alzohairy, 2009). To avoid overweighting identical copies, zero-distance clusters within each family were collapsed by single-linkage hierarchical clustering (cut height h ≤ 1×10⁻¹²), retaining one representative per cluster (yielding n = 7 for each family). We then performed bootstrap resampling with replacement (10,000 replicates; each replicate sampled k = 7 sequences). For each replicate, the mean of the upper triangle of the symmetrised Kimura distance matrix was calculated; uncertainty was summerised as 95% percentile bootstrap confidence intervals.

      Results

      Isolation of a spontaneous Dpy mutant in C. inopinata

      A spontaneous Dpy mutant of C. inopinata with reduced body size was isolated from a laboratory culture of the wild-type strain NKZ35 (Fig. 1A). Genetic crossing demonstrated that this phenotype was caused by a single recessive mutant allele, designated tj217. Whereas homozygous tj217 animals derived from heterozygous parents displayed a Dpy phenotype, homozygous animals derived from homozygous mothers exhibited a more severe Dpy phenotype (extreme Dpy, or Ex-Dpy) (Fig. 1A). These results indicate that maternal contribution of the relevant gene influences the severity of the Dpy phenotype.

      Fig. 1.Identification of a PIF/Harbinger-family insertion in the coding region of Cin-dpy-11.

      (A) Isolation of a spontaneous dumpy (Dpy) mutant in C. inopinata. Adult worms of wild-type (WT), Dpy (tj217 homozygotes derived from heterozygous mothers), Ex-Dpy (extreme-Dpy; a tj217 homozygote derived from a tj217 homozygous mother) are shown. Scale bars, 200 μm. (B) Schematic structures of the WT and the tj217 allele of Cin-dpy-11. A 1,631-bp insertion in exon 6 of the tj217 allele is indicated (exons are shown as grey boxes). Arrows indicate the positions of the primers used in (C). (C) PCR analysis of the Cin-dpy-11 (Sp34_50214510) gene regions in WT and Dpy. An asterisk indicates a PCR-artifact. (D) Sequences of the WT and tj217 allele around the insertion site. Three-base target site duplications (TSDs) flank the insertion, and the terminal inverted repeats (TIRs) of the PIF/Harbinger element are present at both ends (only the terminal 10 bp of each TIR are shown). The C-terminal ER-retention motif (KKTK) of DPY-11 is shown in orange and is disrupted by the insertion. (E) Self dot-plot alignment of Harbinger-1M_cIno sequences.

      Insertion of PIF/Harbinger element in Cin-dpy-11

      To identify the mutation underlying the tj217 Dpy phenotype, we performed whole-genome sequencing on individual Dpy and wild-type worms. Sequencing of five single-worms from each genotype yielded genome coverage of 130–380× per sample. We then screened for variants that were consistently homozygous in Dpy animals and absent from wild-type controls. Among these variants, we identified an insertion in Cin-dpy-11 (Sp34_50214510), a one-to-one orthologue of C. elegans dpy-11 (Fig. 1B). Because loss of dpy-11 causes a recessive dumpy phenotype in C. elegans (Ko and Chow, 2002; Vora et al., 2025), the insertion in Cin-dpy-11 is a strong candidate for the causal lesion underlying tj217. Cin-dpy-11 comprises six exons, spans approximately 3.2 kb of genomic sequence and shares 93.9% amino acid identity (231/246 residues) with C. elegans DPY-11 (Supplementary Fig. 1).

      We validated the candidate insertion by PCR using genomic DNA from wild-type and Dpy worms. The wild-type template yielded a single ∼1-kb amplicon, whereas the Dpy template yielded a larger ∼2.6-kb product (Fig. 1C). Sequencing of the mutant amplicon revealed a 1,631-bp insertion within exon 6 of Cin-dpy-11. The insertion disrupts the C-terminal KKTK motif of DPY-11, which is predicted to contribute to ER retention, thereby providing a plausible molecular basis for loss of function (Fig. 1D; Supplementary Fig. 1). Sequence analysis further showed that the inserted sequence corresponds to a PIF/Harbinger-family DNA transposon with TIRs and a flanking 3-bp TSD (TTA) (Fig. 1D and 1E). We designated this element Harbinger-1M_cIno. Members of the PIF/Harbinger superfamily typically encode two proteins required for transposition: a DDE-family transposase and a protein with a MADF DNA-binding domain. Annotation of the predicted ORFs within Harbinger-1M_cIno identified a MADF DNA-binding domain but did not detect a recognisable DDE-transposase domain, suggesting that Harbinger-1M_cIno is non-autonomous.

      Genome-wide distribution of the Harbinger-1M_cIno family

      Having identified Harbinger-1M_cIno as the element inserted in Cin-dpy-11, we next asked whether it belongs to a broader element family in the C. inopinata genome. To address this, we searched the genome using the terminal 15-bp TIR motif of Harbinger-1M_cIno together with filters for inverted terminal matches, TSDs, and element length. This search recovered 258 related elements, which we hereafter refer to as the Harbinger-1M_cIno family (Supplementary Table 2). For convenience, the recovered elements were indexed by size as Ci_Harb_1 to Ci_Harb_258, with Ci_Harb_258 denoting the longest element. These elements ranged from 180 to 2,752 bp in length, with a median of 394 bp, and most (191/258) were shorter than 500 bp (Fig. 2A). A distinct peak was present at approximately 1.63 kb, corresponding to the size of Harbinger-1M_cIno itself. Conserved terminal blocks were evident at both ends, and the elements were flanked by characteristic 3-bp TSDs with a TWA consensus, most commonly TTA or TAA (Fig. 2B and 2C). The family was distributed predominantly on the autosomes, with only three copies detected on the X chromosome (Fig. 2D). This bias resembles the genomic distribution previously reported for the C. inopinata DNA transposons mCi-hAT1 and mCi-hAT4 (Hatanaka et al., 2024). Together, these results show that the element inserted in Cin-dpy-11 is one member of a genome-wide PIF/Harbinger-derived family in C. inopinata.

      Fig. 2.Characterisation of Harbinger-1M_cIno–like transposable elements in the C. inopinata genome.

      (A) Histogram of lengths for 258 Harbinger-1M_cIno–like transposable elements (TEs) extracted from the C. inopinata genome. The x-axis shows TE length, and the y-axis shows number of TEs per size bin. Elements with a DDE transposase domain are marked by blue arrowheads. Those near 1.63 kb, corresponding to Harbinger-1M_cIno is indicated by black arrowheads. (B) Sequence logos for 80-bp TIRs at both ends. Black traces indicate query sequences used for TE extraction. (C) Sequence logos for 3-bp TSDs. (D) Chromosomal distribution of Harbinger-1M_cIno-like TEs. TE counts per chromosome are shown at the top right.

      A notable feature of this size distribution was that coding capacity was concentrated in the relatively small subset of long elements, particularly those around 1.6 kb, whereas the much more abundant short elements were predominantly noncoding derivatives. To examine this explicitly, we annotated predicted ORFs across all 258 family members. Pfam searches identified five elements carrying only DDE transposase domains (Ci_Harb_250, Ci_Harb_252, Ci_Harb_253, Ci_Harb_255, and Ci_Harb_256), 19 elements carrying only MADF domains, and one element, Ci_Harb_258, carrying both DDE and MADF domains (Fig. 3A; Supplementary Tables 2 and 3). Thus, long-elements clustered around 1.63 kb and above retained recognisable protein-coding potential, whereas most shorter elements appear to represent non-autonomous or degenerate derivatives. After manual curation, however, we found that only one DDE-bearing element, Ci_Harb_252, retained an apparently intact transposase gene, whereas the remaining DDE-like loci were pseudogenised. By contrast, 14 of the 19 MADF-bearing loci retained intact MADF domains. Ci_Harb_258 was particularly informative because, as the longest element and the only family member carrying both DDE- and MADF-related loci, it provides the clearest structural reference for the family; however, both loci in Ci_Harb_258 were pseudogenised. In particular, the MADF-related locus carries a 10-bp deletion in exon 2 that causes a frameshift and disrupts most of the MADF domain (Fig. 3A and 3B; Supplementary Fig. 2). RNA-seq analysis further showed that the intact DDE transposase gene Sp34_30019800 and several MADF-domain genes were expressed across eggs, adult males, and adult females, whereas most pseudogenes showed little or no expression (Fig. 3C and 3D). These observations indicate that the Harbinger-1M_cIno family contains only a small number of potentially functional coding loci embedded within a much larger population of short, noncoding derivatives.

      Fig. 3.Structural organisation and expression of DDE- and MADF-domain loci within the Harbinger-1M_cIno family

      (A) Schematic structures of Harbinger-1M_cIno family members carrying DDE transposase- or MADF-domain loci. Red and gold gene models indicate DDE- and MADF-domain loci, respectively; filled models denote predicted intact genes and open models denote pseudogenes. Grey boxes indicate putative TIRs. (B) Alignment of the MADF-coding region in Ci_Harb_240 (Sp34_20054020) and Ci_Harb_258 (Sp34_20090510). A 10-bp deletion in exon 2 of Sp34_20090510 causes a frameshift predicted to disrupt most of the MADF domain. (C) Expression of DDE-bearing loci across eggs (E), adult males (M), and adult females (F). (D) Expression of MADF-bearing loci across the same stages. In (C) and (D), each facet represents a single locus; filled circles indicate predicted genes and open circles indicate pseudogenes. All panels share the same y-axis scale (counts per million, CPM).

      We next examined how these coding and noncoding derivatives are related to one another. Sequence-similarity clustering of the 258 elements resolved 74 groups, indicating substantial diversification within the family. Because most groups were small, we focused on the four largest clusters for further analysis. One major cluster comprised elements closely similar to Harbinger-1M_cIno in both size and sequence and showed relatively low divergence, whereas the other large clusters consisted of shorter and more heterogeneous elements with substantially greater divergence (Fig. 4A). We then used Ci_Harb_258 as a structural reference to compare similarity across the family. Most elements retained highly similar terminal regions, but similarity across the internal region was partitioned. One group preferentially retained the segment corresponding to the MADF-associated region of Ci_Harb_258, whereas the DDE-bearing elements retained the segment corresponding to the transposase-encoding portion (Fig. 4B). Together, these patterns are consistent with diversification from a common Harbinger-1M_cIno–like ancestor through internal deletion and related structural rearrangements.

      Fig. 4.Divergence and structural partitioning within the Harbinger-1M_cIno family.

      (A) Kimura two-parameter distances within the four largest sequence clusters. Cluster labels indicate the number of elements and median length. The long-element cluster (clu_9; median 1,628 bp) shows markedly lower divergence than the shorter clusters. (B) BLAST similarity of Harbinger-1M_cIno family members to Ci_Harb_258. Each horizontal line represents one query sequence; colors indicate nucleotide identity. Most elements retain high similarity at the termini, whereas internal similarity is partitioned between the DDE-transposase and MADF-associated regions. Gray and purple side bars indicate DDE-bearing elements and clu_9 members, respectively.

      Turmoil2 affiliation of the Harbinger-1M_cIno family

      Because the 258-element search was anchored on the TIR sequence of Harbinger-1M_cIno, it was designed to recover the family related to the inserted element rather than to comprehensively identify all PIF/Harbinger lineages in C. inopinata. We therefore next asked how the Harbinger-1M_cIno family relates to the PIF/Harbinger elements in C. elegans (named Turmoil1 and Turmoil2). DDE transposase phylogeny across Caenorhabditis and related PIF/Harbinger elements resolved two major clades corresponding to Turmoil1 and Turmoil2 and recovered the conserved D-D-E catalytic motif in both groups (Fig. 5). Within this framework, the coding component of the Harbinger-1M_cIno family is associated with the Turmoil2 lineage.

      Fig. 5.Phylogenetic relationships of PIF/Harbinger superfamily transposases.

      (A) Maximum-likelihood phylogeny inferred from DDE transposase sequences of Caenorhabditis species together with representative PIF/Harbinger elements from other taxa. Numbers above branches indicate SH-aLRT and ultrafast bootstrap (UFBoot) support values. Transposases from the ISL2 and ISL2EU groups were used as outgroups. (B) Multiple sequence alignment of the DDE domain in Turmoil1– and Turmoil2-derived transposases. Numbers in parentheses indicate amino acid positions in the full-length proteins. Conserved DDE catalytic residues are marked by arrowheads.

      For comparative purposes, we then surveyed 14 Caenorhabditis genomes using a common repeat library containing Turmoil1– and Turmoil2-type consensuses. Copy number and family composition varied markedly among species (Supplementary Fig. 3; Supplementary Table 4). C. inopinata, C. bovis, and C. japonica each contained at least 80 detectable PIF/Harbinger hits, whereas the remaining 11 species contained 15 or fewer. Within this comparative survey, most detectable hits in C. inopinata matched the Turmoil2-type consensus, whereas C. bovis and C. japonica were dominated by Turmoil1-type hits. Outside C. inopinata, PIF/Harbinger insertions typically retained both DDE- and MADF-related genes within the same TIR-bounded locus or were represented by DDE-only loci; by contrast, MADF-only loci were not detected. Representative examples from C. niphades and C. nigoni showed the canonical head-to-head arrangement of DDE and MADF genes, closely resembling the structure of Ci_Harb_258 (Supplementary Fig. 3E). Together, these results indicate that the extensive proliferation of MADF-only derivatives associated with the Harbinger-1M_cIno family is unusual among the Caenorhabditis genomes examined here and is especially pronounced in C. inopinata.

      A Turmoil1-derived PIF/Harbinger family in C. inopinata

      The Turmoil-based comparison suggested that C. inopinata might retain a second PIF/Harbinger-derived family distinct from Harbinger-1M_cIno. To examine this directly, we analysed a species-specific repeat library for C. inopinata, which identified three PIF/Harbinger-derived consensus families. One showed clear similarity to C. elegans Turmoil1, whereas the other two were derived from the Harbinger-1M_cIno family (Supplementary File 2). TBLASTN searches using C. niphades Turmoil1 transposase (CAI2355069.1) and MADF (CAI2355070.1) as queries recovered seven MADF-like loci but no DDE transposase-like loci, and all seven MADF loci retained full-length MADF domains. Thus, unlike the Harbinger-1M_cIno family, which includes a small subset of long elements with DDE and/or MADF coding potential, this second family was represented in the current genome only by MADF-bearing loci.

      We next examined whether these MADF loci retained recognisable terminal structures. Inverted-repeat scans of the ±1 kb regions flanking each MADF gene detected matching flanking repeats at four of the seven loci, although these were shorter and less conserved than those associated with the Harbinger-1M_cIno family. These observations indicate that C. inopinata retains a distinct Turmoil1-like MADF family, whereas the corresponding DDE transposase has been lost and the flanking TIRs have become degraded. To define this family more explicitly, we reconstructed a consensus from the conserved terminal sequences shared by the seven MADF-coding loci and by additional noncoding loci that retained the same terminal architecture. We designated the resulting Turmoil1-derived family Harbinger-2M_cIno, to distinguish it from the Turmoil2-associated Harbinger-1M_cIno family.

      Contrasting histories of the two C. inopinata PIF/Harbinger families

      We then compared the evolutionary histories of the two PIF/Harbinger-derived families in C. inopinata, Harbinger-1M_cIno and Harbinger-2M_cIno. Reconstruction of a Turmoil1-derived consensus followed by genome-wide searching identified 1,376 Harbinger-2M_cIno copies ranging from 80 to 2,878 bp in length, with a median size of 138 bp, distributed predominantly on autosomes (Fig. 6; Supplementary Table 5). Although these elements retained 3-bp TWA target-site duplications, their terminal and internal sequences were less conserved than those of Harbinger-1M_cIno (Fig. 6B and 6C). Divergence analyses likewise showed broader sequence variation and older substitution profiles, lacking the low-divergence peak characteristic of Harbinger-1M_cIno (Fig. 6E). Together, these features indicate that Harbinger-2M_cIno is older and more degenerate than Harbinger-1M_cIno.

      Fig. 6.Genome-wide characterisation of the Harbinger-2M_cIno family in C. inopinata.

      (A) Length distribution of 1,376 Harbinger-2M_cIno copies. (B) Sequence logos of the 5′ and 3′ TIRs. (C) Sequence logo of the 3-bp TSD. (D) Chromosomal distribution of Harbinger-2M_cIno copies across the six chromosomes of C. inopinata. Numbers indicate total copies per chromosome. (E) Kimura two-parameter distances within the major sequence clusters of the family. Cluster labels indicate cluster ID, number of unique elements [total number of elements], and median length.

      A direct comparison of Turmoil1– and Turmoil2-derived MADF genes supported the same conclusion. Phylogenetic analysis of 21 MADF loci, including seven Turmoil1-derived and 14 Turmoil2-derived sequences, resolved two reciprocally monophyletic clades, with markedly longer branches in the Turmoil1-derived clade (Supplementary Fig. 4A and 4B). Bootstrap estimates of mean pairwise Kimura protein distance were 0.143 (95% CI, 0.082–0.189) for Turmoil1 and 0.026 (95% CI, 0.016–0.033) for Turmoil2, indicating substantially greater divergence in the Turmoil1-derived MADF set (Supplementary Fig. 4C). RNA-seq analysis further detected transcription of all seven Turmoil1-derived MADF genes, without a strong stage-specific bias (Supplementary Fig. 4D). Together, these results indicate that the two PIF/Harbinger-derived families in C. inopinata have distinct histories: Harbinger-1M_cIno is comparatively homogeneous and consistent with more recent proliferation, whereas Harbinger-2M_cIno is more deeply diverged and structurally degraded, although its MADF genes remain transcriptionally retained.

      Discussion

      The main advance of this study is not simply the identification of a transposon insertion in Cin-dpy-11, but the recognition that this lesion reveals an unusual evolutionary configuration of a PIF/Harbinger-derived family in C. inopinata. Canonical autonomous PIF/Harbinger elements generally encode both DDE transposase and MADF functions within the same element (Kapitonov and Jurka, 1999; Sinzelle et al., 2008). Against that background, the Turmoil2-associated family Harbinger-1M_cIno is notable because these two functions are retained in separate descendant elements, yet only the MADF-bearing branch has undergone broad genome-wide expansion, whereas DDE-bearing descendants remain restricted to a small number of copies. We suggest that the key signal here is not modularity itself, which is a general feature of PIF/Harbinger biology, but a markedly asymmetric pattern of descendant-family expansion. In this sense, the tj217 mutant reveals not only a causal lesion but also a naturally occurring example of how a canonical two-ORF PIF/Harbinger family can diversify into unequally expanded descendant branches.

      The interpretation that the Cin-dpy-11 insertion underlies the Dpy phenotype is biologically compelling. In C. elegans, dpy-11 encodes a membrane-associated thioredoxin-like protein required for normal body and sensory-organ morphogenesis (Ko and Chow, 2002; Vora et al., 2025), making disruption of its coding region fully consistent with the strong body-shape defect observed here. Although formal proof by rescue or targeted reversion remains to be obtained, the convergence of genetic mapping, structural validation, and functional correspondence to C. elegans dpy-11 strongly supports the insertion as the causal lesion.

      The abundance of short, domainless derivatives in the Harbinger-1M_cIno family is also consistent with a MITE-like mode of family expansion. MITEs are short non-autonomous DNA transposons that commonly arise through internal deletion of larger autonomous elements while retaining terminal features required for mobilisation (Yang et al., 2009). In this context, the many short Harbinger-1M_cIno derivatives are most naturally interpreted as MITE-like noncoding descendants of a more complete ancestral element, and their predominance over the longer coding-capable copies suggests that amplification of such noncoding derivatives has been a major component of Turmoil2-associated family evolution in C. inopinata. Against this background, the coding derivatives represent only a minor subset of the family. Within that subset, MADF-bearing descendants appear to have persisted and expanded more successfully than DDE-bearing descendants, whereas the great majority of copies are short noncoding derivatives rather than coding descendants.

      By contrast, the Turmoil1-derived family in C. inopinata appears to represent a later stage of erosion rather than a parallel active lineage. It is retained as MADF-bearing loci without a detectable DDE counterpart, its terminal structures are more degraded, and its sequences are more diverged. For that reason, we view the Turmoil1-derived family mainly as a useful contrast that defines the range of evolutionary states present in the same genome. Its retained transcription is intriguing, but transcription alone does not establish present-day function. At most, these loci may represent substrates for subsequent host co-option, as seen for PIF/Harbinger-derived proteins in other animal and plant lineages (Kapitonov and Jurka, 1999; Sinzelle et al., 2008; Velanis et al., 2020). More importantly, the Turmoil1-derived family no longer behaves like the Turmoil2-associated family and therefore is unlikely to contribute to the ongoing activity inferred for Harbinger-1M_cIno.

      Comparative analysis also suggests that Turmoil elements have followed different structural trajectories in C. elegans and C. inopinata. Relative to the compact terminal structures typical of PIF/Harbinger elements, the Turmoil1 and Turmoil2 elements described in C. elegans appear to carry unusually extended, composite TIR structures, whereas the Turmoil2-associated family in C. inopinata retains shorter termini resembling those seen in other Caenorhabditis species. Together with the TE-rich genome of C. inopinata and its loss of several small-RNA pathway components, this contrast raises the possibility that lineage-specific host–TE interactions have shaped Turmoil architecture and activity differently in the two species. At present, however, this remains a hypothesis rather than a demonstrated mechanism.

      The spontaneous insertion into Cin-dpy-11 indicates that at least one Turmoil2-associated PIF/Harbinger element has been active under laboratory conditions in C. inopinata. Together with the TE-rich genome of this species (Kanzaki et al., 2018) and the recent activity of other DNA transposon families, including the Ci-hAT1/mCi-hAT1 system (Hatanaka et al., 2024), this observation suggests that C. inopinata provides a host background permissive for recent transposon activity. However, this permissiveness should not be interpreted as a simple loss of TE control. Although C. inopinata has lost multiple components of the ERGO-1/26G-siRNA pathway (Kanzaki et al., 2018), many core small-RNA pathway genes remain conserved, implying that complementary silencing systems are still present. We therefore infer that TE surveillance in C. inopinata has been reconfigured rather than abolished. Within this altered host background, the two PIF/Harbinger-derived families appear to occupy different temporal states: Harbinger-1M_cIno remains linked to recent activity, whereas Harbinger-2M_cIno preserves the genomic signature of substantial past amplification but now appears to have lost present-day mobilisation competence. In this framework, host-side permissiveness may help explain why recent transposition is still detectable in C. inopinata, whereas family-specific properties, such as compatibility between DDE-bearing and MADF-bearing derivatives and recognition of family-specific terminal sequences, are more likely to determine which descendant classes continue to expand.

      Several limitations define the next steps. First, the causal role of the Cin-dpy-11 insertion, although strongly supported by the convergence of phenotype, structure, and gene identity, remains to be verified by rescue or targeted reversion. Second, the inference that DDE and MADF functions may be supplied in trans within the Harbinger-1M_cIno family remains hypothetical rather than mechanistically demonstrated. Third, the evolutionary placement of the shortest noncoding derivatives necessarily relies on indirect evidence from terminal structure and sequence similarity. These limitations are experimentally tractable. Direct mobilisation assays, DDE–MADF cross-complementation tests, and terminal-sequence recognition assays should determine whether the expanded MADF-bearing derivatives are still functionally linked to the residual DDE source.

      In summary, our results support a model in which two PIF/Harbinger-derived families with a shared origin have followed sharply different trajectories in C. inopinata. The Turmoil2-associated family remains linked to recent activity, as evidenced by the Harbinger-1M_cIno insertion into Cin-dpy-11, whereas the Turmoil1-derived family appears more eroded, having lost a detectable DDE component and retained only degraded terminal structures. A distinguishing feature of C. inopinata is the broad proliferation of MADF-bearing derivatives, particularly in the Turmoil2-associated family, whereas a comparable genome-wide expansion of DDE-only derivatives is not evident. Together with the TE-rich genome of C. inopinata, the loss of multiple small-RNA pathway components, and recent activity of other DNA transposon families in this species, these findings are most consistent with a host background permissive for transposon activity. Within such a background, family-specific properties, including compatibility between DDE- and MADF-bearing derivatives and recognition of terminal sequences, are likely to influence which descendant classes remain mobilisation-competent and therefore expand. This framework makes C. inopinata a useful system for testing how host genome defense and PIF/Harbinger family architecture interact to shape lineage-specific transposon trajectories.

      Funding

      This work was supported by Japan Society for the Promotion of Science (JSPS) KAKENHI Grant Numbers 19H03212 and 25H01307, and JST CREST Grant Numbers JPMJCR18S7 and JPMJCR23B1 to T.K. and A.S.

      Authors’ contributions

      A.S. and T.K. conceptualised the study. S.O., K.K. and N.H. isolated and photographed a spontaneous mutant and performed genetic crosses. X.J. prepared a whole-genome sequencing library, and A.Y. performed whole-genome sequencing. S.O., K.K. and N.H. carried out PCR and validated the TE insertion by Sanger sequencing. X.J., K.S., and T.K. primarily conducted the computational analyses with support from S.S. X.J. and K.S. drafted the initial version of the manuscript, which was reviewed and edited by all authors. T.K. finalised the manuscript.

      Data Availability

      The raw whole-genome sequencing reads generated in this study for Caenorhabditis inopinata strains SA1635 and SA1689 (five biological replicates each) have been deposited in the NCBI Sequence Read Archive (SRA) under BioProject accession number [PRJDB14908]. The wild-type (NKZ35, SA1689) and mutant (SA1635 [tj217]) nematode strains utilized in this work are available upon request from the corresponding author.

      Figure legends

      Supplementary Fig. 1. Conservation of the dpy-11 gene in C. elegans and C. inopinata.

      Amino-acid alignment of DPY-11 proteins from C. elegans (F46E10.9) and C. inopinata (Sp34_50214510.t1), highlighting sequence conservation and the ER-retention signal (KKTK).

      Supplementary Fig. 2. Pseudogenisation of a MADF-domain gene by frameshift.

      (A) Pfam annotation identifies an intact MADF domain in Sp34_20054020.t1 (red), whereas the domain is truncated in Sp34_20090510.t1. (B) Nucleotide alignment of the MADF-coding regions of Sp34_20054020 and Sp34_20090510, showing a 10-bp deletion and the resulting disrupted reading frame. Exons of Sp34_20054020 are highlighted in green. The amino-acid translation of Sp34_20054020 is shown beneath the alignment.

      Supplementary Fig. 3. Detection of Turmoil1– and Turmoil2-type transposable elements in Caenorhabditis.

      (A) Copy numbers of PIF/Harbinger-like elements in 14 species based on cross-species RepeatMasker library. The families rnd-6_family-38341 and rnd-6_family-25689 are classified as Turmoil1-type, whereas rnd-6_family-8036 is Turmoil2-type. (B–D) Numbers of elements with DDE or MADF domains within 2 kb upstream or downstream of TE hits. Genome assemblies not at chromosome-scale are marked with asterisks. (E) Representative Turmoil2-like loci in C. niphades and C. nigoni, showing opposite-orientation DDE (red) and MADF (gold) genes flanked by grey TIRs.

      Supplementary Fig. 4. Greater divergence of MADF proteins in Turmoil1 than in Turmoil2.

      (A) Maximum-likelihood phylogeny of MADF domains from Caenorhabditis species and representative PIF/Harbinger elements. Numbers above branches indicate SH-aLRT/UFBoot support for key nodes. Domesticated PIF/Harbinger-derived Myb/SANT proteins—NAIF1 (Homo sapiens) and ALP2 (Arabidopsis thaliana)—and a Pangu transposon from Magallana gigas were used as outgroups. (B) Multiple sequence alignment of MADF domains from C. inopinata Turmoil1 (T1; purple, n = 7) and Turmoil2 (T2; light blue, n = 14). Numbers in parentheses indicate amino acid positions in the full-length proteins. (C) Mean pairwise Kimura protein distances of MADF domains from Turmoil1 and Turmoil2. Violin plots show bootstrap distributions of mean distances (10,000 replicates; n = 7 sequences per family after deduplication). Black points indicate bootstrap means; error bars indicate 95% percentile confidence intervals. A bootstrap test indicated greater divergence in Turmoil1 than in Turmoil2 (p < 0.001). (D) Expression of Turmoil1-derived MADF-domain genes across developmental stages (E, egg; M, male; F, female). Each facet represents one locus; points indicate two replicates per stage. All facets share the same y-axis scale (CPM).

      Acknowledgements

      Genome data analyses were performed in part using the DDBJ supercomputer system. We thank members of the Kikuchi laboratory for helpful discussions.

      This pre-print is available under a Creative Commons License (Attribution 4.0 International), CC BY 4.0, as described at http://creativecommons.org/licenses/by/4.0/

      References

      1. Alzohairy, A., 2009 Phylip and Phylogenetics. Genes, genomes and genomics.
      2. Billi, A. C., S. E. J. Fischer and J. K. Kim, 2014 Endogenous RNAi pathways in C. elegans. WormBook: The Online Review of C. elegans Biology: 1–49.
      3. Bourque, G., K. H. Burns, M. Gehring, V. Gorbunova, A. Seluanov et al., 2018 Ten things you should know about transposable elements. Genome Biology 19: 199–z.
      4. Buckley, B. A., K. B. Burkhart, S. G. Gu, G. Spracklin, A. Kershner et al., 2012 A nuclear Argonaute promotes multigenerational epigenetic inheritance and germline immortality. Nature 489: 447–451.
      5. Camacho, C., G. Coulouris, V. Avagyan, N. Ma, J. Papadopoulos et al., 2009 BLAST+: architecture and applications. BMC Bioinformatics 10: 421–421.
      6. Capella-Gutierrez, S., J. M. Silla-Martinez and T. Gabaldon, 2009 trimAl: a tool for automated alignment trimming in large-scale phylogenetic analyses. Bioinformatics (Oxford, England) 25: 1972–1973.
      7. Carver, T., S. R. Harris, M. Berriman, J. Parkhill and J. A. McQuillan, 2012 Artemis: an integrated platform for visualization and analysis of high-throughput sequence-based experimental data. Bioinformatics (Oxford, England) 28: 464–469.
      8. Chen, X., O. Schulz-Trieglaff, R. Shaw, B. Barnes, F. Schlesinger et al., 2016 Manta: rapid detection of structural variants and indels for germline and cancer sequencing applications. Bioinformatics (Oxford, England) 32: 1220–1222.
      9. Crooks, G. E., G. Hon, J. Chandonia and S. E. Brenner, 2004 WebLogo: a sequence logo generator. Genome Research 14: 1188–1190.
      10. Danecek, P., J. K. Bonfield, J. Liddle, J. Marshall, V. Ohan et al., 2021 Twelve years of SAMtools and BCFtools. GigaScience 10: giab008.
      11. Das, P. P., M. P. Bagijn, L. D. Goldstein, J. R. Woolford, N. J. Lehrbach et al., 2008 Piwi and piRNAs act upstream of an endogenous siRNA pathway to suppress Tc3 transposon mobility in the Caenorhabditis elegans germline. Molecular Cell 31: 79–90.
      12. Dobin, A., C. A. Davis, F. Schlesinger, J. Drenkow, C. Zaleski et al., 2013 STAR: ultrafast universal RNA-seq aligner. Bioinformatics 29: 15–21.
      13. Fattash, I., R. Rooke, A. Wong, C. Hui, T. Luu et al., 2013 Miniature inverted-repeat transposable elements: discovery, distribution, and activity. Genome 56: 475–486.
      14. Guang, S., A. F. Bochner, D. M. Pavelec, K. B. Burkhart, S. Harding et al., 2008 An Argonaute transports siRNAs from the cytoplasm to the nucleus. Science (New York, N.Y.) 321: 537–541.
      15. Han, M., C. Xiong, H. Zhang, M. Zhang, H. Zhang et al., 2015 The diversification of PHIS transposon superfamily in eukaryotes. Mobile DNA 6: 12–17. eCollection 2015.
      16. Hancock, C. N., F. Zhang and S. R. Wessler, 2010 Transposition of the Tourist-MITE mPing in yeast: an assay that retains key features of catalysis by the class 2 PIF/Harbinger superfamily. Mobile DNA 1: 5–5.
      17. Hatanaka, R., K. Tamagawa, N. Haruta and A. Sugimoto, 2024 The impact of differential transposition activities of autonomous and nonautonomous hAT transposable elements on genome architecture and gene expression in Caenorhabditis inopinata. Genetics 227.
      18. Iwasaki, Y. W., K. Shoji, S. Nakagwa, T. Miyoshi and Y. Tomari, 2025 Transposon-host arms race: a saga of genome evolution. Trends Genet 41: 369–389.
      19. Jin, Y., O. H. Tam, E. Paniagua and M. Hammell, 2015 TEtranscripts: a package for including transposable elements in differential expression analysis of RNA-seq datasets. Bioinformatics 31: 3593–3599.
      20. Jurka, J., and V. V. Kapitonov, 2001 PIFs meet Tourists and Harbingers: a superfamily reunion. Proceedings of the National Academy of Sciences of the United States of America 98: 12315–12316.
      21. Kalyaanamoorthy, S., B. Q. Minh, T. K. F. Wong, A. von Haeseler and L. S. Jermiin, 2017 ModelFinder: fast model selection for accurate phylogenetic estimates. Nature Methods 14: 587–589.
      22. Kanzaki, N., I. J. Tsai, R. Tanaka, V. L. Hunt, D. Liu et al., 2018 Biology and genome of a newly discovered sibling species of Caenorhabditis elegans. Nature Communications 9: 3216–3215.
      23. Kapitonov, V. V., and J. Jurka, 1999 Molecular paleontology of transposable elements from Arabidopsis thaliana. Genetica 107: 27–37.
      24. Katoh, K., and D. M. Standley, 2013 MAFFT multiple sequence alignment software version 7: improvements in performance and usability. Molecular Biology and Evolution 30: 772–780.
      25. Kawahara, K., T. Inada, R. Tanaka, M. Dayi, T. Makino et al., 2023 Differentially Expressed Genes Associated with Body Size Changes and Transposable Element Insertions between Caenorhabditis elegans and Its Sister Species, Caenorhabditis inopinata. Genome Biology and Evolution 15: evad063.
      26. Ko, F. C. F., and K. L. Chow, 2002 A novel thioredoxin-like protein encoded by the C. elegans dpy-11 gene is required for body and sensory organ morphogenesis. Development (Cambridge, England) 129: 1185–1194.
      27. Liu, P., K. Panda, S. A. Edwards, R. Swanson, H. Yi et al., 2024 Transposase-assisted target-site integration for efficient plant genome engineering. Nature 631: 593–600.
      28. Minh, B. Q., H. A. Schmidt, O. Chernomor, D. Schrempf, M. D. Woodhams et al., 2020 IQ-TREE 2: New Models and Efficient Methods for Phylogenetic Inference in the Genomic Era. Molecular Biology and Evolution 37: 1530–1534.
      29. Mistry, J., S. Chuguransky, L. Williams, M. Qureshi, G. A. Salazar et al., 2021 Pfam: The protein families database in 2021. Nucleic Acids Research 49: D412–D419.
      30. Oomura, S., K. Tsuyama, N. Haruta and A. Sugimoto, 2022 Transgenesis of the gonochoristic nematode Caenorhabditis inopinata by microparticle bombardment with hygromycin B selection. microPublication Biology 2022: doi:10.17912/micropub.biology.000564. eCollection 002022.
      31. Rice, P., I. Longden and A. Bleasby, 2000 EMBOSS: the European Molecular Biology Open Software Suite. Trends in Genetics : TIG JID507085 16: 276–277.
      32. Rognes, T., T. Flouri, B. Nichols, C. Quince and F. Mahe, 2016 VSEARCH: a versatile open source tool for metagenomics. PeerJ 4: e2584.
      33. Rubin, E., and A. A. Levy, 1997 Abortive gap repair: underlying mechanism for Ds element formation. Molecular and Cellular Biology 17: 6294–6302.
      34. Shen, W., S. Le, Y. Li and F. Hu, 2016 SeqKit: A Cross-Platform and Ultrafast Toolkit for FASTA/Q File Manipulation. PLoS One 11: e0163962.
      35. Sinzelle, L., V. V. Kapitonov, D. P. Grzela, T. Jursch, J. Jurka et al., 2008 Transposition of a reconstructed Harbinger element in human cells and functional homology with two transposon-derived cellular genes. Proceedings of the National Academy of Sciences of the United States of America 105: 4715–4720.
      36. Siomi, M. C., K. Sato, D. Pezic and A. A. Aravin, 2011 PIWI-interacting small RNAs: the vanguard of genome defence. Nature Reviews.Molecular Cell Biology 12: 246–258.
      37. Sun, S., N. Kanzaki, M. Dayi, Y. Maeda, A. Yoshida et al., 2022 The compact genome of Caenorhabditis niphades n. sp., isolated from a wood-boring weevil, Niphades variegatus. BMC Genomics 23: 765–768.
      38. Velanis, C. N., P. Perera, B. Thomson, E. de Leau, S. C. Liang et al., 2020 The domesticated transposase ALP2 mediates formation of a novel Polycomb protein complex by direct interaction with MSI1, a core subunit of Polycomb Repressive Complex 2 (PRC2). PLoS Genetics 16: e1008681.
      39. Vora, M., J. Dietz, Z. Wing, K. George, J. Kelly Liu et al., 2025 Genome-wide analysis of Smad and Schnurri transcription factors in C. elegans demonstrates widespread interaction and a function in collagen secretion. eLife 13: doi:10.7554/eLife.99394.
      40. Wang, K., M. Li and H. Hakonarson, 2010 ANNOVAR: functional annotation of genetic variants from high-throughput sequencing data. Nucleic Acids Research 38: e164.
      41. Wicker, T., F. Sabot, A. Hua-Van, J. L. Bennetzen, P. Capy et al., 2007 A unified classification system for eukaryotic transposable elements. Nature Reviews.Genetics 8: 973–982.
      42. Yang, G., D. H. Nagel, C. Feschotte, C. N. Hancock and S. R. Wessler, 2009 Tuned for transposition: molecular determinants underlying the hyperactivity of a Stowaway MITE. Science (New York, N.Y.) 325: 1391–1394.
      43. Yang, G., F. Zhang, C. N. Hancock and S. R. Wessler, 2007 Transposition of the rice miniature inverted repeat transposable element mPing in Arabidopsis thaliana. Proceedings of the National Academy of Sciences of the United States of America 104: 10962–10967.
      44. Zhang, X., C. Feschotte, Q. Zhang, N. Jiang, W. B. Eggleston et al., 2001 P instability factor: an active maize transposon system associated with the amplification of Tourist-like MITEs and a new superfamily of transposases. Proceedings of the National Academy of Sciences of the United States of America 98: 12572–12577.
      45. Zhou, L., T. Feng, S. Xu, F. Gao, T. T. Lam et al., 2022 ggmsa: a visual exploration tool for multiple sequence alignment and associated data. Briefings in Bioinformatics 23.

      Journal of Modern Sciences

      info@modernsciences.org

      About
      Modern Sciences is a media company that aims to provide the academics, researches, and science enthusiasts with the latest news and stories in the world of Applied and Pure Science.