Fig. 1. Gene nomenclature for duplicated and novel genes. a. Two Cat loci arose by gene duplication. Comparison with the human genome reveals a single copy of CAT in the human genome. Thus, Cat1 and Cat2 are paralogs in the axolotl genome. b. Annotation of orthologs and paralogs. Geneb1 is annotated based on its homology to GENEB in another organism (indicated by the pattern of exons), while a paralogous gene is annotated as Geneb2 since it has a lower sequence similarity. Grey shaded area indicates the chromosome to highlight that Geneb1 is in a different locus c. A putative novel gene family member in axolotl that has some (BLAST e-value 1e-16) similarity to a gene in another organism. d. A putative novel gene that does not have any homologs, but has a long (1941aa) open reading frame.

Gene nomenclature in axolotl

The chromosome-scale axolotl genome assemblies^2,3 made it possible to study not only the coding sequences, but also the evolution and synteny at the whole-genome level and accurate gene annotation is vital for those purposes. However, to ensure accurate and unambiguous communication between scientists, it is crucial to work out a set of rules for how the genes and proteins should be named, how to distinguish paralogs and orthologs, and how to name pseudogenes.

4.     Orthologs and paralogs

Unlike zebrafish and Xenopus laevis, the axolotl genome was not structured by whole-genome duplication events. However, axolotl-specific (i.e. in-paralogs) as well as salamander-specific (i.e. out-paralogs) gene duplicates are known (Fig. 1a) and can be shared with other salamander taxa ⁴². In order to define the evolutionary relationship between the genes, it is crucial to define whether the genes in different species derive from a common ancestor (orthologs), or whether the genes in the same species arose from the same ancestor (paralogs) ⁴³. While it is important to standardize the gene names for orthologous genes, it is also important to keep the gene naming compatible to that used across vertebrates, in order to make the analyses of the axolotl data comparable with the large body of available human and mouse datasets. Similar to the human and mouse orthologous gene nomenclature (HUMOT) project ⁴⁴, we propose to rely on a mixed approach that combines both the orthology information with the synteny and also integrates the data from expert orthology resources. The importance of the synteny can be demonstrated by a following example. Imagine a situation that in the axolotl, a gene homologous to a gene GENEB in another species was duplicated (Fig. 1b), while the copy that is more similar to the ancestral gene (indicated by the exon pattern in Fig. 1b) moved out of the locus. In this case, the copy that stayed in the locus should be annotated as Geneb2, while the one that moved out should be annotated as Geneb1. In contrast, if the copy that moved out is less similar to GENEB than the one that stayed, then the latter should be annotated as Geneb1 and the former as Geneb2. If synteny is not conserved then phylogenetic analysis should be used to identify the relationships between the genes. As these relationships can be complex it is better to deliberately exclude ambiguous orthology information than to propagate incorrect assumptions about the genes. 

In the  case of paralogous genes, one must be very cautious in order to avoid name collisions across species. Imagine, in the above example, there were also a gene GENEB2 in the same species. However, this GENEB2 is not orthologous to the newly annotated Geneb2 in axolotl. In this case, we suggest using the next available number in the series, following consultation with other gene nomenclature groups. In the outlined example, it would be Geneb3. Hence, we always recommend contacting the axolotl gene nomenclature committee when naming a gene.

Paralogs:

Paired related homeobox 1 and Paired related homeobox 2: Prrx1 and Prrx2

Duplicated genes:

Catalases – Cat and Cat2

5.     Novel genes.

While paralogous genes should be treated as described above, potential novel genes should be annotated differently. Ideally, novel genes should be characterized functionally and named based on their function, whenever possible. However, for the vast majority of the novel genes this is not the case. We therefore suggest to examine the orthologous relationships to other functionally characterized or unambiguously annotated genes in axolotl or other organisms. If a novel gene can be assigned to a certain protein family but fails to fulfill the criteria outlined in 4 to be annotated unambiguously, we suggest adding a suffix ‘-like’ to the gene symbol of the closest ortholog based solely on the sequence homology. This annotation may be changed later when its existence is confirmed in the lab and the functional data become available. Finally, if none of those approaches can be applied, the novel gene should be annotated as Locxxxx (Fig. 1d), where xxxx are the NCBI gene IDs. In this case, the gene sequence should be submitted to NCBI first to get the NCBI gene ID, which ensures that this gene symbol is not arbitrary.

Example:

Prothymosin-alpha-like : Prothymosin-alpha-like – Ptal (Fig. 1c) and

Loc12345 between Tmem79 and Smg5 (Fig. 1d). 

5.1  Pseudogenes

Some genes may lose their open reading frame (ORF) in the course of the time and, thus, also their function. However, since the gene sequences are still present in the genome, they must be accurately annotated as they may have undesired effects on transgenics, transcript quantification and other analyses that rely on the gene sequence. To stay consistent with the nomenclature guidelines in other organisms and particularly in humans ²⁴, we suggest appending the suffix ‘p’ (pseudogene) followed by a number to the gene symbol of the ancestral gene if the gene is processed, while naming unprocessed genes as new members of the family with the suffix ‘p’.

6.     Transcript variants

We propose that the transcript variant names are formed after the following schema: GeneName.TranscriptNumber. All predicted transcripts of a gene are numbered by the order in which they were annotated. While in organisms with a well-established genome annotation, the nomenclature does not specifically deal with the transcript annotations, we feel that in axolotl the annotation is frequently changed at the moment and therefore it is vital to keep track of which isoforms were proven to be wrong. For example, imagine that a gene has three annotated isoforms, Gene.1, Gene.2 and Gene.3. However, it turns out that Gene.3 is just an artifact. Nevertheless, another isoform, Gene.4, is shown to be very tissue-specific. At this point, it is better to have the isoforms annotated as Gene.1, Gene.2 and Gene.4, to indicate that Gene.3 does not exist and avoid confusion in case any works referred to Gene.3 before it was excluded.

Example: transcript variants of Prrx1

Prrx1.1, Prrx1.2

7.     Non-coding transcripts

Similar to the guidelines for the human genome outlined in Seal et al, 2020⁴⁵, we propose that non-coding transcripts are annotated according to their RNA type. MicroRNAs should be annotated as “mir-XXX”, where XXX is the submission ID in the miRbase database⁴⁶. Transfer RNAs should receive gene names following the pattern tRNA-XXX-YYY- GtRNAdbID, where XXX is the three-letter amino acid code, YYY is the anticodon and GtRNAdbID is the gene ID in GtRNAdb database⁴⁷. Other classes of non-coding RNAs should be named after consulting the gene naming committee as very little work has been done on non-coding RNAs in axolotl so far and, thus, the exact requirements will be met later.

Similar to the guidelines in other vertebrates, we suggest to use the gene symbol with the suffix ‘-as’ for non-coding transcripts that originate from the same promoter as an annotated protein-coding gene on the opposite strand, e.g. Dio3-as for a non-coding RNA that originates from the Dio3 promoter.

8.     Mitochondrial genes

In order to stay consistent with the well-annotated species, we propose to use the human annotation of the mitochondrial genes (NCBI Reference Sequence: NC_012920.1) in axolotl. However, in agreement with the remainder of the nomenclature, only the first letter should be capital, e.g. Mtnd2 for the mitochondrially encoded NADH dehydrogenase 2, while the human counterpart would be MTND2.