GTDB currently assigns the name g__Limnothrix to a cluster represeted by s__Limnothrix_rosea. However, despite Limnothrix rosea being a valid name, there are signs that it isn’t a good match to the type species (true Limnothrix should be somewhere close to Psudanabaena, see PMID 36443823).
The NCBI taxonomy puts it under g__Picosynechococcus; x__unclassified Picosynechococcus; s__Picosynechococcus sp. PCC 7002, which is more reasonable. I partly corroborated this assignment by eyeing the branching order of the trees in PMID 32700350 vs PMID 36443823. It should be possible to get a better confirmation by making a 16S tree.
PS: PMID 36443823 makes the GTDB and ICNafp nomenclatures a lot more comparable, now that both agree on using a backbone based on phylogenomics. This opens a chance of hormonizing some names for “basically the same clade” like Cyanobiaceae/Prochlorococcaceae, Thermosynechococcales/Acaryochloridales if y’all want to.
PPS: PMID 32700350 has a name Limnothrix euryhalinus for what is currently called sp001693275. This invalid name is not indexed in LPSN, but is visible at Picosynechococcus sp. PCC 7002 - NCBI - NLM . Perhaps the naming pipeline should be changed to be able to pick this sort of names up?
Hi Artoria
Thanks for your feedback on this. Cyanobacteria are generally problematic because much of the classification is still phenotype based and we are lacking the genomes of many type strains, Limnothrix redekei NIES 847 being a case in point. Perhaps the simplest solution for the time being is to revert the name to a placeholder until the genome of the type is sequenced so we can unambiguously place it in the tree. I will confer with our cyanobacterial specialist Jan Mares on this.
Bw, Phil
1 Like
Oops, correction. The NCBI assignment for “Limnothrix rosea” should end with x__unclassified Picosynechococcus; s__[Limnothrix rosea] IAM M-220. The PCC 7002 stuff is for euryhalinus/sp001693275.
Instead of a placeholder I like the idea of putting it under g__Picosynechococcus (perhapss__Picosynechococcus_roseus) more as that’s what everyone else does based on 16S. It does come with a small risk of the type species Picosynechococcus fontinalis turning out too different from “L. rosea” when its genome eventually gets sequenced, but at least it will be close… ish.
From a more programmatic viewpoint: maybe the taxonomy filter should be able to handle the name-with-a-bracket thing better? NCBI curation policy and taxon pages (e.g. [Limnothrix rosea] IAM M-220 - NCBI - NLM ) state that
Square brackets (\[ \]) around a genus indicates that the name awaits appropriate action by the research community to be transferred to another genus.
So it would make sense for the name-decoration pipeline to (1) not use the genus name suggested by the binomial-in-bracket and (2) potentially use the genus name from a parent node instead. Would mean less human work.
We’ve run into problems in the past relying on 16S-based placements in the reference tree. I think for the time being we’ll go with a placeholder name until the genomes of the type species are sequenced and made public. Point taken on square brackets, we will avoid using those names in future.