Skip to Main content Skip to Navigation
Journal articles

The axolotl genome and the evolution of key tissue formation regulators

Abstract : Salamanders boast an illustrious history in biological research as the animal in which the Spemann organizer 1 and Sperry's chemoaffinity theory of axonal guidance 2 were discovered. Since 1768, when Spallanzani discovered tail and limb regeneration, researchers have probed this animal's remarkable regenerative capabilities with increasing molecular resolution. A. mexicanum (Fig. 1a) was first collected by von Humboldt, and has been cultivated in the laboratory since 1864 as a model for investigating phenomena such as nuclear reprogramming, the embryology of germ-cell induction, retinal neuron processing and regeneration 3. Owing to the ease with which A. mexicanum can be bred in the laboratory, a sophisticated molecular toolkit has been developed for this species, including germline transgenesis and CRISPR-mediated gene mutation as well as viral and other transfection methods. These tools have enabled discoveries such as the identification of the source cells of regeneration and molecular pathways that control regeneration 4,5. A full exploitation of the axolotl model, including understanding regeneration and why it is limited in other tetrapods, requires analysis of its genome regulation and evolution. However, efforts towards comprehensive assembly of salamander genomes have been challenging owing to their large genome sizes (14-120 Gb) and the large number of repetitive regions they contain; the 32-Gb axolotl genome is ten times the size of the human genome. Here we report the sequencing, assembly and analysis of the axolotl genome. A long-read assembler for large genomes Our aim was to generate a genome sequence assembly for the d/d axolotl strain (Fig. 1a), which is commonly used in laboratory regene ration studies owing to its compatibility with live imaging. Taking into consideration the expected challenge of assembling the complex 32-Gb genome 6 , we sequenced 110 million long reads (32× coverage, N50 read length 14.2 kb) using Pacific Biosciences (PacBio) instruments (Supplementary Information section 1) to avoid the read sampling bias that is often found when using other technologies and to span repeat-rich genomic regions that cause breaks in short-read assemblies (Fig. 1b, c). We developed an assembly algorithm (MARVEL) that integrates a two-phase read-correction procedure that keeps long PacBio reads intact for assembly (Supplementary Information section 2). MARVEL produced a contig assembly with an N50 of 218 kb. Next, we used 7× Illumina-based sequencing to correct sequence errors in 1% of the contig bases (Fig. 1b), which yielded a sequence accuracy of more than 99.2%. On the basis of the Illumina data, we estimated a heterozygosity of 0.47% (Supplementary Information section 2.2). To provide a scaffold for the contig assembly, we generated de novo optical maps using the Bionano Saphyr system (Supplementary Information section 2.3). The Bionano map resolved contig chimaeras, which were found in 1.7% of contigs, slightly reducing N50 contig length to 216 kb (Fig. 1d). The final hybrid assembly yielded an N50 scaffold length of 3 Mb. Compared to the short-read assembly of the 20-Gb spruce genome 7 or the 22-Gb loblolly pine genome 8 , which involved 12× long-read coverage, the axolotl assembly showed 56-and 29-fold improvements in contiguity, respectively (Table 1). To assess the completeness of the assembly (Supplementary Information section 4.1), we first determined the number of aligning non-exonic ultraconserved elements 9 (UCEs). We found that 194 (98.5%) of 197 non-exonic UCEs that are conserved across vertebrates align to the axolotl assembly. By comparison, 189 and 192 UCEs align to the Tibetan frog and Xenopus genomes, respectively, and 195 UCEs align to the coelacanth genome, indicating that the completeness of the axolotl genome assembly is comparable to or better than the two other amphibian genomes, which are smaller than 2.3 Gb 10. Salamanders serve as important tetrapod models for developmental, regeneration and evolutionary studies. An extensive molecular toolkit makes the Mexican axolotl (Ambystoma mexicanum) a key representative salamander for molecular investigations. Here we report the sequencing and assembly of the 32-gigabase-pair axolotl genome using an approach that combined long-read sequencing, optical mapping and development of a new genome assembler (MARVEL). We observed a size expansion of introns and intergenic regions, largely attributable to multiplication of long terminal repeat retroelements. We provide evidence that intron size in developmental genes is under constraint and that species-restricted genes may contribute to limb regeneration. The axolotl genome assembly does not contain the essential developmental gene Pax3. However, mutation of the axolotl Pax3 paralogue Pax7 resulted in an axolotl phenotype that was similar to those seen in Pax3 −/− and Pax7 −/− mutant mice. The axolotl genome provides a rich biological resource for developmental and evolutionary studies.
Complete list of metadatas

Cited literature [37 references]  Display  Hide  Download
Contributor : Bianca Habermann <>
Submitted on : Monday, October 28, 2019 - 12:39:25 PM
Last modification on : Wednesday, October 14, 2020 - 3:45:14 AM


The axolotl genome and the evo...
Publication funded by an institution


Distributed under a Creative Commons Attribution 4.0 International License




Sergej Nowoshilow, Siegfried Schloissnig, Ji-Feng Fei, Andreas Dahl, Andy Pang, et al.. The axolotl genome and the evolution of key tissue formation regulators. Nature, Nature Publishing Group, 2018, 554 (7690), pp.50 - 55. ⟨10.1038/nature25458⟩. ⟨hal-01874591⟩



Record views


Files downloads