Arrow Worms, Rewritten: How Evolution Remixed a Genome to Build a Brand-New Body Plan

If you’ve never heard of chaetognaths (a.k.a. arrow worms), you’re not alone. They’re transparent, torpedo-shaped ocean predators with a mouthful of chitinous grasping spines—and a body plan so strange that biologists have argued about where to put them on the animal family tree for more than a century. A new Nature study finally shows how these animals became so unique: they reorganized their genome, minted new genes, duplicated thousands more, rewired gene regulation, and then used that toolkit to reinvent themselves.

Meet the misfits of the sea

Chaetognaths are common in plankton, yet evolutionary outliers. Their heads carry raptorial spines; their bodies are mostly muscle; their nervous systems include compact ventral ganglia; their sensory gear—ciliary organs and papillae—doesn’t quite look like anyone else’s. For decades, their mix-and-match traits confused phylogeneticists: some features looked “deuterostome-ish,” others “protostome-ish.” Only in the last few years did a consensus emerge that arrow worms belong with Gnathifera, a group that also includes rotifers and gnathostomulids—sister to the rest of the spiralians (molluscs, annelids, etc.). That new placement came from large-scale genomic analyses and the discovery of a quirky Hox gene called MedPost, shared with rotifers.

But even after we knew where they belong, we didn’t know how they got so weird. That’s the riddle this new paper set out to solve.

The big reveal: a genome, a 3D chromosome map, and a single-cell atlas

The authors sequenced the genome of Paraspadella gotoi to chromosome scale, built a 3D contact map (Hi-C) to see how DNA folds, profiled open chromatin (ATAC-seq), mapped DNA methylation, and created a 30,000-cell single-cell RNA-seq atlas covering juveniles and adults. It’s like switching on the lights in a previously dim room: we can now follow the threads from genome architecture to gene expression to the cell types that make the animal.

A companion News & Views captures the punchline succinctly: arrow worms seem to have arrived at their distinctive anatomy through “bursts of gene emergence, duplication and loss”—a massive remix of the standard bilaterian toolkit.

A genome that evolved fast—and oddly

1) Fusions and rearrangements, but not chaos

Across bilaterian evolution, many species preserve chunks of ancestral chromosome neighborhoods, sometimes called bilaterian linkage groups. P. gotoi still shows statistical traces of these old neighborhoods, but stitched together differently: several chaetognath chromosomes look like fusions of two to four ancient blocks. That points to accelerated chromosomal evolution in the gnathiferan lineage—but not a total scramble. By contrast, rotifers seem far more scrambled. In short: arrow worms fused and re-arranged without melting down their genome map entirely.

2) Centromeres without the usual centromere protein

Many animals position centromeres (the “handles” for chromosome segregation) using a special histone variant, CenH3, and its friends. Chaetognaths have lost a big chunk of that classical toolkit. Yet the Hi-C data flag localized centromere-like regions enriched for transposons and heavily methylated DNA—neocentromeres made with alternative machinery. It’s a striking example of convergent solutions to a basic cellular problem. (Rotifers, their gnathiferan cousins, likely have holocentromeres that run along entire chromosome arms; arrow worms went the other way and re-localized.)

3) Limited “A/B” neighborhoods and no strong TADs

Unlike vertebrate genomes with pronounced active/inactive compartments and topologically associating domains (TADs), the arrow worm chromosomes show weak, if any, of those regularities—hinting that their 3D genome logic is simpler or at least different. That meshes with other regulatory quirks below.

The cell types behind the body: both ancient and brand-new

With ~30,000 single cells, the atlas resolves ~30 differentiated cell types. Some are familiar—neurons, muscles, gut, ciliated cells, germline—sharing conserved regulators with other bilaterians. Others are chaetognath-specific: the “chaete” cells associated with grasping spines, ciliary sense-organ cells, and papillae sensory cells scattered through the epidermis. The team validated marker genes with in situ hybridization, literally painting these cell types in the animal.

What’s unusual is which genes these cell types use. Many of their defining markers are brand-new, phylum-specific genes—and many others come from lineage-specific duplications. That’s the developmental genome equivalent of building instruments nobody else has, then composing music that only those instruments can play.

Tandem duplication, not a whole-genome duplication

When you see thousands of extra gene copies in an animal, a tempting explanation is a whole-genome duplication (WGD). But several lines of evidence argue against a WGD here:

The duplicates don’t come in neatly paired, chromosome-wide sets.
Their synonymous substitution profiles (a molecular “age” signal) point to a single burst of copy-making, but the copies are scattered in patterns best explained by tandem duplication (back-to-back repeats on the same chromosome) and nearby dispersals—not a duplicated whole genome.
Introns haven’t vanished in a way that would scream retrotransposition.

So arrow worms most likely ran a massive tandem-duplication spree, then kept and specialized the copies they needed.

This matters because duplicate genes are raw material for innovation. One copy can keep the old job; the other can tinker—change expression timing, tweak protein domains, or shift cell-type specificity—without breaking essential functions. The authors show that many duplicates in P. gotoi touch ion channels, development, and neural functions—exactly the sorts of levers you’d pull to retune a predator’s senses and behaviors.

The Hox cluster: expanded and eccentric

Hox genes are the “postal codes” of animal body plans, patterning the head-to-tail axis. Outside vertebrates, most animals carry a relatively compact Hox set. Chaetognaths, by contrast, boast 14 Hox genes, with expansions in the middle and posterior groups, all arranged in a big (~2.4 Mb) cluster that’s still colinear (the physical order reflects the body axis order). They also carry MedPost, the gnathiferan hallmark sitting between median and posterior genes. In situ data show some Hox genes active in the nervous system and peripheral sensory cells—not just in broad embryonic stripes—hinting at redeployed roles. Recent preprint work also points to chaetognaths having one of the most extensive Hox repertoires among protostomes.

If you’re trying to build unusual structures (like arrow-worm papillae and ciliary organs) or to repattern a ventral nervous system, extra Hox “dials” plus duplicated regulators give you more ways to do it.

New genes—lots of them—and they’re doing real work

One of the study’s most surprising results is the catalog of phylum-specific orphan genes: thousands of gene families found in chaetognaths but not detectable elsewhere, many without recognizable protein domains. Crucially, these aren’t idle passengers. They’re expressed, often cell-type-specifically, sometimes forming the majority of markers in those novel sensory and epidermal cell types. That suggests two things:

The gnathiferan ancestors lost many ancestral bilaterian genes (especially in neural/sensory families);
Arrow worms replaced some of that lost functionality with newly evolved genes and specialized duplicates.

Evolution didn’t just simplify; it simplified and then reinvented, using different parts.

DNA methylation, retargeted: from genes to transposons

In most invertebrates, DNA methylation lightly coats gene bodies—often on stably expressed, housekeeping genes—while leaving many transposons untouched. P. gotoi flips that script. Gene bodies are largely unmethylated; transposons are heavily methylated, particularly LTR retrotransposons. The enzymatic toolkit is also unusual: multiple DNMT1 and UHRF1 copies with simplified domain architecture; six DNMT3s that have lost domains that usually steer methylation to gene bodies; and three TET genes lacking classical domain combinations. Together, these changes point to a repurposed methylome whose job is primarily transposon suppression, not transcriptional fine-tuning.

Why might that matter? If methylation vacates gene bodies, other mechanisms must maintain steady housekeeping expression. Enter trans-splicing and operons.

Trans-splicing and operons: a different way to run the transcriptome

The team shows that nearly half of arrow-worm genes receive a splice-leader (trans-splicing), and about 18% are organized into operons—arrays of adjacent genes transcribed together and then processed into individual mRNAs. Multiple distinct splice leaders circulate; there aren’t operon-specific leaders, unlike in nematodes. Operonic/trans-spliced genes are enriched for metabolic and housekeeping functions, are expressed more broadly (less cell-type-specific), and are associated with fewer distal regulatory elements—in other words, a compact, streamlined regulatory style. They’re especially prominent in the germline.

Put these pieces together and a picture emerges: when gene-body methylation steps back, trans-splicing and operons take on a bigger role in stabilizing housekeeping expression. Meanwhile, cell-type-specific regulation leans heavily on the new and duplicated genes with more bespoke enhancers. Different jobs, different regulatory strategies.

A plausible evolutionary storyline

Fossils suggest the chaetognath body plan has been remarkably stable since the Cambrian, but the lineage’s deeper history sits among gnathiferans that experienced accelerated gene loss and fast chromosome evolution. The authors propose a two-phase arc:

Simplification: Gnathiferan ancestors lost many conserved bilaterian genes—including neural/sensory families—and reshuffled chromosomes extensively.
Reinvention: The chaetognath branch then duplicated thousands of genes (mostly by tandem duplication), invented thousands of new genes, retargeted methylation to transposons, embraced trans-splicing/operons for house-keeping, and re-deployed Hox and other regulators to sculpt novel sensory and structural cell types suitable for fast, ambush predation in the plankton.

The end result is the arrow worm we see today: a predator assembled from old and new parts, with a regulatory system that keeps housekeeping simple and puts the evolutionary “creativity budget” into cell-type innovation.

Why this matters beyond arrow worms

1) Body plans can be rebuilt with different genetic bricks.
We tend to talk about conserved toolkits as if they’re the only way to make animals. Chaetognaths show the opposite: even after widespread gene loss, lineages can “re-complexify” using new genes and duplicated variants—and still land on stable, functional anatomies for hundreds of millions of years.

2) Multiple regulatory logics are compatible with complex life.
Gene-body methylation isn’t obligatory. Strong A/B compartments and TADs aren’t inevitable. Operons and trans-splicing—sporadic across animals—can be center stage in some phyla. Chaetognaths expand the menu of viable genomic “operating systems.”

3) Evolutionary history can include simplification and innovation.
The chaetognath story underlines that “more complex over time” is too simple. Lineages can shed genes and then innovate anew. For Gnathifera broadly, that might have included small-bodied branches (rotifers) with different strategies (e.g., horizontal gene transfer), and larger predatory branches (chaetognaths) that leaned on duplication and de novo gene birth.

A closer look at a few standout findings

The papillae and ciliary sense organs

Arrow worms hunt by detecting fluid disturbances—a lifestyle that depends on exquisitely tuned mechanosensation. The atlas recovers specialized papillae and ciliary organ cell types whose marker genes include piezo channel expansions (mechanosensors) and other duplicated ion channels. Functionally, that’s exactly what you’d evolve if swimming and predation are your day job.

The epidermis and spines

Chaetognaths sport a pluristratified epidermis (unusual outside vertebrates) and iconic grasping spines. Corresponding clusters express distinct chitin synthases and chitinases—again, a case where structural novelty rides on both new genes and paralog specialization.

The nervous system and Hox

Hox expression in arrow worms goes beyond a simple anterior–posterior code; some Hox genes map to peripheral sensory neurons in the epidermis. That’s an intriguing redeployment that could help pattern those peppered sensory fields. The cluster’s expansion gives extra “slots” to tune these domains.

Open questions this work unlocks

What, biochemically, are the neocentromeres made of?
If CenH3 is gone, what proteins take over? Can we reconstitute that architecture in vitro?
How do brand-new genes arise and specialize so quickly?
De novo gene birth is still poorly understood. Arrow worms offer a living laboratory to test how new ORFs gain regulatory hooks and cell-type specificity.
Are operons a general solution when methylation shifts to transposons?
Nematodes and tunicates also use operons and trans-splicing; ctenophores and some nematodes show gene-body methylation loss. Is there a recurring evolutionary “swap” here?
How widespread is Hox redeployment into sensory cell types?
Comparative single-cell work in gnathostomulids and micrognathozoans could reveal whether arrow worms are outliers or part of a gnathiferan trend.

TL;DR (but make it sticky)

Arrow worms didn’t just inherit a weird body plan—they engineered it. Massive tandem duplications and thousands of novel genes gave them a custom parts list.
Their genome runs on a different regulatory logic. Methylation targets transposons; operons and trans-splicing stabilize housekeeping; 3D genome compartmentalization is subdued.
Result: a predator with unique sensory gear and structures, built from an evolutionary remix rather than a by-the-book toolkit.

Glossary (quick and painless)

Gnathifera: A clade including rotifers, gnathostomulids, micrognathozoans—and chaetognaths—sister to the rest of spiralians.
Hox genes: Master regulators that pattern the head-to-tail axis; often arranged in clusters that mirror body order.
MedPost: A gnathiferan-signature Hox-like gene wedged between middle and posterior Hox genes.
Operon: A gene neighborhood transcribed as a single unit, then processed into individual mRNAs (common in bacteria; found in some animals).
Trans-splicing: A process that adds a short “splice-leader” sequence to the 5′ end of mRNAs; often coupled to operons.
Neocentromere: A centromere formed at a new chromosomal location without the classical centromeric protein (CenH3).

Want to dive deeper?

The study itself: Piovani et al., “The genomic origin of the unique chaetognath body plan” (Nature, 2025). It’s open access and packed with figures, extended data, and methods.
Context & commentary: “The perplexing body plan of arrow worms decoded” (News & Views, Nature, 2025). Great plain-language framing.
Background on placement of chaetognaths: Marlétaz et al., “A new spiralian phylogeny places the enigmatic arrow worms among Gnathiferans” (Current Biology, 2019).
Hox repertoire note: Ordoñez et al., preprint on chaetognath Hox expansion (bioRxiv, 2025).

Final thought

We often talk about “the” animal toolkit as if evolution wrote it once and then only annotated the margins. Arrow worms remind us that toolkits can be torn apart and reassembled—that body plans can be re-authored by duplication, invention, and regulatory reinvention. If nature can build a chaetognath out of ancestral scraps and brand-new parts, it means the space of possible animals is larger—and more creative—than we dared imagine.

Keywords: chaetognath body plan, arrow worms, Paraspadella gotoi, Gnathifera, spiralians, rotifers comparison, bilaterian linkage groups, chromosomal fusions, genome evolution, gene loss, gene family contraction, tandem gene duplication, Hox cluster expansion, MedPost Hox gene, posterior Hox genes, de novo gene birth, orphan genes, neocentromeres, centromere evolution, transposon repression, DNA methylation retargeting, methylome, trans-splicing, operons in animals, splice leader, single-cell RNA-seq, cell-type innovation, sensory papillae, ciliary sense organs, pluristratified epidermis, ATAC-seq, Hi-C chromatin architecture, regulatory evolution, plankton predator, evolutionary novelty, Nature 2025

#Chaetognaths, #ArrowWorms, #Gnathifera, #Spiralians, #Bilateria, #GenomeEvolution, #EvoDevo, #ComparativeGenomics, #Phylogenomics, #GeneLoss, #GeneDuplication, #TandemDuplication, #HoxGenes, #HoxCluster, #MedPost, #DeNovoGenes, #OrphanGenes, #Neocentromeres, #Transposons, #DNAmethylation, #Methylome, #TransSplicing, #Operons, #SingleCell, #scRNAseq, #ATACseq, #HiC, #CellTypeAtlas, #SensoryBiology, #CiliaryOrgans, #MarineBiology, #Plankton, #Predation, #RegulatoryGenomics, #BodyPlan, #NaturePaper, #OpenAccess

Original study:

Piovani, L., Gavriouchkina, D., Parey, E. et al. The genomic origin of the unique chaetognath body plan. Nature (2025). https://doi.org/10.1038/s41586-025-09403-2

The RNA Blog

About Us

Monday, August 18, 2025

How Evolution Remixed a Genome to Build a Brand-New Body Plan