The brittle star genome illuminates the genetic basis of animal appendage regeneration

The study addresses whether the gene expression programs controlling animal regeneration are evolutionarily conserved across distant lineages. Echinoderms, particularly brittle stars, possess remarkable regenerative capabilities but lack high-quality genomic resources, impeding mechanistic insights. Given echinoderms’ phylogenetic proximity to vertebrates compared to classical invertebrate regeneration models, the authors aim to generate a chromosome-scale genome for Amphiura filiformis and profile gene expression during adult arm regeneration. They seek to contextualize brittle star genome evolution (chromosome rearrangements, Hox/ParaHox organization, gene family dynamics) and to test for conserved temporal gene expression modules by comparing with vertebrate (axolotl) and invertebrate (Parhyale) appendage regeneration datasets, thereby clarifying the conservation and timing of wound healing, proliferation, and differentiation phases.

Previous echinoderm genomics centered on sea urchins revealed local Hox cluster reorganizations but later sea star genomes showed intact clusters, indicating Hox reorganization is not required for pentameral symmetry. Chromosome-scale genomes now exist for sea stars, cucumbers, and urchins, but not brittle stars, leaving a gap given ~480–500 Myr divergence among echinoderm classes. Prior comparative regeneration studies identified conserved pathways across metazoans but lacked explicit temporal comparisons between invertebrates and vertebrates and suffered from orthology detection challenges. Brittle stars are underrepresented in regeneration transcriptomics despite their relevance and high regeneration frequency in nature. This context motivates a comprehensive brittle star genome and comparative time-series expression analyses to probe conservation of regeneration programs.

• Genome sequencing and assembly: High-coverage Oxford Nanopore long reads (PromethION, ~160.6 Gb ~100×) and 10x Genomics linked reads (~86 Gb) were generated from sperm DNA. Assembly with Flye, polishing with Racon (Nanopore and Illumina reads), haplotig purging (purge_dups), quality assessed by Merqury and Inspector. Proximity ligation (Hi-C; Omni-C) scaffolding using YAHS (validated with 3D-DNA) produced 20 chromosome-scale scaffolds covering 93.5% of the 1.57 Gb assembly (N50 68.86 Mb). BUSCO completeness was 92.7%.
• Repeat annotation: De novo repeat library with RepeatModeler2 and masking with RepeatMasker. Unclassified repeats were classified using a re-trained DeepTE CNN model (validated accuracy 0.98 at P≥0.55). Repeat landscapes and ages were inferred from Kimura divergence and neutral substitution rate estimation.
• Gene annotation: Evidence-based pipeline combining assembled transcriptomes (18 samples), protein homology from 27 metazoans, and ab initio predictions. PFAM domain assignments and BUSCO completeness evaluated. Curated gene lists for immunity, stemness, signalling, neuronal, kinase, and transcription factors were compiled.
• Macrosynteny and ancestral linkage groups: Chromosome-level synteny analyses among brittle star, sea star, sea cucumber, and sea urchins with outgroups to reconstruct Eleutherozoa linkage groups (ELGs). Orthologs identified by reciprocal best hits (DIAMOND), Circos visualization, statistical tests for conserved linkage, and inference of rearrangements (fusions, fissions, translocations).
• Hox/ParaHox analysis: Identification via sequence comparison and phylogenetics, cluster organization mapping, repeat enrichment near breakpoints, and expression profiling across development from published datasets.
• Gene family evolution: Orthogroups built with Broccoli across 28 metazoans (10 Ambulacraria). CAFE5 tested expansions/contractions in 10,367 pre-Ambulacraria families on a fossil-calibrated dated tree. Selected families linked to regeneration and glycosaminoglycan metabolism were highlighted.
• Brittle star regeneration transcriptomics: Time-course RNA-seq across eight conditions (control; 48 hpa, 72 hpa; stages 3–5; 50% proximal and distal), each from pooled individuals. Kallisto quantification, TMM normalization, MFuzz soft-clustering (major clusters A1–A9). GO and curated gene list enrichment (clusterProfiler); TF motif enrichment around TSS (HOMER; −5 kb to +1 kb).
• Explant experiments (regenerative vs non-regenerative response): Arms were amputated proximally (non-regenerative wound closure) and distally (regeneration). Distal, medial (control), and proximal segments sampled at 3 and 5 dpa (3–4 replicates). Kallisto quantification, DESeq2 differential expression versus medial controls per segment/timepoint, identification of DEGs and oppositely regulated candidates.
• Comparative regeneration analyses: Axolotl limb regeneration RNA-seq (12 time points) reprocessed and clustered (Ax1–Ax9); Parhyale limb regeneration clusters adopted from prior work (renamed P1–P8). Broccoli orthogroups used to map homologs and permutation tests assessed overrepresentation of shared genes across co-expression clusters between species pairs. Functional enrichment of conserved genes was evaluated with tailored backgrounds controlling for homology detectability and cluster biases.
• Repetitive element transcription: SalmonTE used to quantify repeat family expression in early (immune) vs mid (proliferative) regeneration phases; DESeq2 tested for differential activity; associations with repeat age (divergence) and genomic context were examined.

• Genome resource: Produced the first chromosome-scale brittle star (Amphiura filiformis) genome (1.57 Gb) with 20 chromosomes (93.5% of assembly), scaffold N50 68.86 Mb, and 30,267 protein-coding genes (BUSCO 92.7% complete). Repeat content is 59.3% with a pronounced repeat activity burst ~10–15 Ma, primarily DNA transposons.
• Chromosome evolution: Reconstructed 23 Eleutherozoa ancestral linkage groups (ELGs), derived from a fusion of bilaterian B2 and C2. Observed very few interchromosomal events between sea star and sea cucumber (one event over ~500 Myr), but extensive reshuffling in A. filiformis with 26 interchromosomal rearrangements (estimated 0.052/Myr), making it the most rearranged echinoderm genome among those compared.
• Hox/ParaHox organization: A. filiformis Hox cluster is locally reorganized: Hox1–Hox3 inverted at 3’ cluster end and Hox8 inverted and displaced between Hox9/10 and Hox11/13a. Expanded repeats within the Hox cluster (notably SINE/tRNA-Deu-L2) are associated with breakpoints (BH-corrected permutation P<0.05) and have 18–22% divergence, suggesting activity ~100 Ma. One breakpoint occurs near Hox4, paralleling sea urchins. ParaHox cluster disruptions include a tandem duplication of Gsx (two paralogues) separated by >5 Mb from Xlox–Cdx, while Xlox–Cdx remain linked (sea urchins have full dispersion). Developmental expression shows low early expression of anterior Hox (Hox1, Hox3–Hox6) and higher of Hox7, Hox11/13a, Hox11/13b; brittle star Hox2 is expressed early, unlike sea urchins.
• Gene family dynamics: Echinoderm gene complements are relatively stable overall (790 expanded/contracted of 10,367 families). In A. filiformis, significant expansions involve regeneration-linked families (plasminogen-like, carboxypeptidase B-like, coagulation factor-like, ficolin-like, TRIM-like/MAPKKK-like, acyl-CoA thioesterase, UDP-glucuronosyltransferase). Several of these regulate coagulation/clotting in vertebrates, suggesting roles in wound closure and immune defense. Genes involved in keratan sulfate metabolism are overrepresented among both expanded and contracted families, aligning with known glycosaminoglycan roles in arm regeneration.
• Regeneration transcriptomics in brittle star: Soft-clustering defined nine major temporal clusters (A1–A9), revealing three phases: (1) wound healing (immune response, cell migration/protection; clusters A1–A4), with NF-κB motif enrichment around TSS (A2); (2) proliferation (clusters A9, A5–A7), enriched for stemness, cell division and translation; TF motif enrichments include NRF1, p53, PRDM14, YY1, RORα and zinc-finger TFs; early activation of proliferation-related signaling (VEGF, AKT, insulin-like, JAK-STAT) from ~48 hpa; and (3) differentiation and appendage morphogenesis (A8) enriched for TFs including tbx3-1/3-2, ngn1-like, heyl-like.
• Cross-species conservation: Comparative clustering showed broad conservation of co-expression modules, especially during proliferation. Five A. filiformis clusters significantly overlap axolotl (926 shared genes), six overlap Parhyale (913 genes), and four are conserved across all three (154 genes). Conservation is limited between axolotl and Parhyale alone (two clusters, 370 genes), indicating brittle star bridges detection across distant models. Conserved clusters predominantly correspond to proliferative and early wound-healing phases and are deployed in a consistent temporal order across species (one noted heterochrony: Ax3 vs A5). Conserved genes are enriched for proliferation-related processes (translation, DNA replication, chromosome segregation, intracellular transport) and for kinase and stemness categories, with depletion of immune genes. Only two TFs among the 154 three-way conserved genes (Id2-like, Wdhd1-like) point to potential conserved regulators of regeneration; YY1 and NRF1 motifs align across corresponding clusters in axolotl and brittle star.
• Repetitive element transcription: Early immune-phase shows higher transcriptional activity of repeats versus the proliferative phase; immune-up repeats are younger (lower divergence) and often intergenic, supporting a model where transposon control may influence the transition from immune response to regeneration.
• Regenerative versus non-regenerative responses (explant): Distal regenerating tips show many more DEGs than proximal wound-healing-only tips at 3 and 5 dpa. Most proximal DEGs overlap with distal DEGs (61%), while most distal DEGs are distal-specific (82%). Five genes exhibit opposite regulation between distal and proximal: Agrin-like-1 and AF133635 (down in wound healing, up in regeneration), AW-SPI, AF118858 (TRIM-like), and Gdf8/myostatin (up in wound healing, down in regeneration), implicating them in steering outcomes from wound closure to regenerative proliferation.

The chromosome-level A. filiformis genome reveals extensive karyotype reshuffling relative to other echinoderms, with a repeat-associated history suggesting episodes of genomic instability. Local Hox and ParaHox rearrangements, with shared breakpoint vicinity near Hox4, and limited embryonic expression of anterior Hox support relaxed regulatory constraints within echinoderms. Functionally, expansions in coagulation/immune-associated gene families and glycosaminoglycan metabolism may underpin efficient wound closure that precedes regeneration. Transcriptomic time courses delineate wound response, proliferation, and differentiation phases, and motif enrichments suggest conserved regulators (NF-κB in immune response; NRF1, p53, PRDM14, YY1, RORα in proliferation). Cross-species comparisons demonstrate conserved deployment of an ancient proliferative program during appendage regeneration across vertebrates and invertebrates, while wound-response and differentiation modules are more lineage- and tissue-specific. This supports scenarios where either distinct wound programs converge on a conserved proliferation machinery or homologous proliferation modules have been retained amid divergent immune gene evolution. The brittle star’s phylogenetic position enables detection of long-range expression conservation that is harder to detect in direct vertebrate–arthropod comparisons.

This work delivers the first chromosome-scale brittle star genome and a comprehensive regeneration transcriptomic atlas, resolving echinoderm chromosome evolution and Hox/ParaHox dynamics, and revealing extensive conserved proliferation gene expression across distant animal appendage regeneration models. It nominates candidate gene families and regulators (for example, coagulation-related genes, Id2, Wdhd1, YY1, NRF1) for functional follow-up. Future directions include sequencing additional brittle star genomes to resolve the timing and mechanisms of karyotype evolution, detailed analyses of repeat dynamics and chromatin architecture (especially at Hox clusters), higher-resolution temporal sampling of early wound responses, and single-cell atlases to test the hypothesis of a conserved regenerating proliferative cell type across animals.

• Temporal resolution at the earliest wound-response phase in brittle star is limited; denser sampling is needed to refine conservation of early-response gene dynamics.
• Conclusions are based on one brittle star species; additional ophiuroid genomes are required to generalize karyotype and Hox/ParaHox reorganization patterns and to parse intra-class variability.
• Motif enrichment and TF inference are biased toward vertebrate TF binding data; functional validations in echinoderms are needed.
• Cross-species comparisons depend on orthology detection and harmonization of staging; some homologs are undetectable or ambiguously assigned, potentially underestimating conservation.
• Bulk RNA-seq may obscure cell-type-specific programs; single-cell approaches would clarify conserved cell lineages and trajectories.