PlanetCatfish.com

Does anyone know if anyone is performing ribosomal or similar work to potentially genetically identify the different Hypancistrus? It seems as though this type of work would "look past" the natural variation within a group if we found the right part to analyze.

Feel free to do the work

I'm waiting too for real study of the Hypancistrus complex.

Janne

You could try a simple COX1 or CytB barcoding, but in all probability there'll not be much variation, maybe not even with the mitochondrial control region/d-loop.
The sad fact is that while in theory DNA analysis would be great for untangling the relationships of groups where morphological methods doesn't work well, in practice it turns out that DNA analysis tends to give rubbish results in those same groups too.

That said, I don't know why there's no molecular studies on Hypancistrus. It does seem a pretty obvious and high-priority target.

Mike Noren wrote:The sad fact is that while in theory DNA analysis would be great for untangling the relationships of groups where morphological methods doesn't work well, in practice it turns out that DNA analysis tends to give rubbish results in those same groups too.

While I agree this is true in some instances (e.g. Symphysodon), it remains that almost every barcoding study, even on really well studied groups such as birds, reveal instances of divergences within species consistent with interspecific variation. This variation was overlooked by the current taxonomy, although it is not to say that differences could not be found should freshly collected material be re-examined.

I think that for these difficult groups, more sophisticated techniques such as SNP analysis will be used in future, as next generation sequencing gets cheaper. Here is a good example: http://www3.interscience.wiley.com/jour ... 0/abstract.

However, I am not sure why it is always assumed that Hypancistrus will be difficult to resolve with both morphological and molecular methods. The fact is, except H. zebra, there has been no work done on any Brazilian Shield Hypancistrus. There are no keys available, and we don't see characters, so how do we know?

The shipments we receive as aquarists are usually juvenile, sex biased, consolidated, split, and reconsolidated from different localities. This does not make it easy for us to interpret the diversity.

Sure, some of them look similar, but I think it may all make a lot more sense when scientists can work with lots of well preserved material, with capture locations.

Mike Noren wrote:That said, I don't know why there's no molecular studies on Hypancistrus. It does seem a pretty obvious and high-priority target.

I think the problem is more to do with the politics of carrying out science in Brazil, than the fishes themselves.

If anyone is particularly interested in some L numbers in particular, it costs me about $20 US to generate a COI sequence.

So if you feel like sending me some tissue, and 20 bucks, I can do it.

Probably not a bad idea actually, as it would be good to have some of the lower Xingu loricariid diversity represented on databases like GenBank before it is eradicated!

EDIT ...

There are already 22 specimens of Hypancistrus on BOLD, but some of these appear to be misidentified.

http://www.boldsystems.org/views/taxbro ... xid=182994.

Looking at other ancistrins, someone at Guelph appears to be putting a project together on ornamental loricariids. None of these sequences are public though, so I can't do anything with them yet.

Lets hope they don't publish with these mis-identifications!

Jeez I wish. I've also been playing with the idea of sequencing Hypancistrus genes donated by aquarists, but I don't have much lab experience yet.

racoll wrote:

Mike Noren wrote:That said, I don't know why there's no molecular studies on Hypancistrus. It does seem a pretty obvious and high-priority target.

I think the problem is more to do with the politics of carrying out science in Brazil, than the fishes themselves.

This is the primary reason. A secondary reason is the difficulty of collecting Hypancistrus; this is why they are rare in collections.

racoll wrote:There are already 22 specimens of Hypancistrus on BOLD, but some of these appear to be misidentified.

http://www.boldsystems.org/views/taxbro ... xid=182994.

Looking at other ancistrins, someone at Guelph appears to be putting a project together on ornamental loricariids. None of these sequences are public though, so I can't do anything with them yet.

Lets hope they don't publish with these mis-identifications!

Huh, I might have to contact whoever's doing this in the future. That's interesting, but indeed, those are some bad misidentifications.

Someone interested to corporate with UFPA here in Belém? Send a mail to jme@telia.com

Janne

Edit, my email address... I put my old one before ;)

I know that in a lab we used to colaborate with we were able to differentiate the various species of Prphyra from the east coast using rRNA amplification and RFLP. Are some thinking this would not work with Hyps?

racoll wrote:While I agree this is true in some instances (e.g. Symphysodon), it remains that almost every barcoding study, even on really well studied groups such as birds, reveal instances of divergences within species consistent with interspecific variation. This variation was overlooked by the current taxonomy, although it is not to say that differences could not be found should freshly collected material be re-examined.

Commonly such variation was not overlooked, but known but not recognized in classification. I'm not sure I've ever encountered truly cryptic species, species which really could not be told apart morphologically, but dozens of cases along the lines of Sundadanio axelrodi where molecular evidence, shockingly, shows that the neon blue "form" is genetically distinct from the red "form".
On the other hand I've seen quite a lot of cases like with icelandic charr, where species are easily told apart morphologically and ecologically, but not molecularly.

I think that for these difficult groups, more sophisticated techniques such as SNP analysis will be used in future, as next generation sequencing gets cheaper.

Sure, but then you're getting in to "what's a species" territory. If you need to use methods so sensitive that every single individual is unique on a continuous spectrum, by what yardstick do you decide that what you're looking at is a species? When I read papers like that I often feel the researchers are interpreting ink blots: they find support for whatever they think they'll find.

However, I am not sure why it is always assumed that Hypancistrus will be difficult to resolve with both morphological and molecular methods.

I'd say it's a fair guess it'll be messy, given what's been seen in other young groups in the midst of an adaptive radiation, such as Victoria or Malawi cichlids.

Mike Noren wrote:That said, I don't know why there's no molecular studies on Hypancistrus. It does seem a pretty obvious and high-priority target.

I think the problem is more to do with the politics of carrying out science in Brazil, than the fishes themselves.

That would be my guess, and I've said how I feel about that on more than one occasion.

As for BOLD, I've yet to search it without finding egregious identification errors. I used to report errors, but now I only do so when it messes things up for me, e.g. misidentified Anguilla sequences which screwed up identification for a conservation project. For a database specifically set up to provide standardized, well-identified, curated, well-documented sequences it's remarkable that it manages to be less accurate than a simple BLAST search at GenBank.

alga wrote:I know that in a lab we used to colaborate with we were able to differentiate the various species of Prphyra from the east coast using rRNA amplification and RFLP. Are some thinking this would not work with Hyps?

You mean Porphyra? Markers evolve differently depending on the organism; what could be useful to clarify some groups may be useless in other groups. For example, mitochondrial genomes evolve much more slowly in plants than they do in animals, if I'm not mistaken. There is also variation in evolutionary rate with genes in the nuclear genome between organisms.

Sorry, did not notice the Porphyra typo

We were able to differentiate the single cell thick blades of different species. So, basically no morphological differnces due to overlapping ranges for size and width but different after cutting the rRNA segment.

I repeat my self, any one interested... mail me, if money is a problem... let solve that, time is running...

Janne

Mike Noren wrote:Sure, but then you're getting in to "what's a species" territory. If you need to use methods so sensitive that every single individual is unique on a continuous spectrum, by what yardstick do you decide that what you're looking at is a species? When I read papers like that I often feel the researchers are interpreting ink blots: they find support for whatever they think they'll find.

Every taxonomic decision gets into "what's a species" territory. Surely, providing you can demonstrate an independent evolutionary history through corroborating lines of evidence, this should be what is required. I don't think they were "interpreting ink blots", they investigated evolutionary processes, as part of a hypothesis testing enquiry:

Mims et al. (2010) wrote: However, Bayesian analysis of nuclear single nucleotide polymorphism (SNP) data revealed two distinct genetic clusters corresponding perfectly to morphologically diagnosed L. fuelleborni and M. zebra

The mtDNA rejected the morphological hypothesis of two species, but nuclear SNP data supported the classification, and, they demonstrated the likely cause of the shared mtDNA haplotypes. Although this wasn't a taxonomic study, it is standard scientific procedure, in my opinion.

alga wrote:we were able to differentiate the various species of Prphyra from the east coast using rRNA amplification and RFLP. Are some thinking this would not work with Hyps?

Which rRNA genes? Generally, for species identification purposes, the mitochondrial protein coding genes such as cytochrome b and cytochrome oxidase I are most appropriate in vertebrates. These are easier to work with, as they are simple to align, don't have indels, and suffer less from contamination issues. They also have faster mutation rates, so contain more information at the species/population level.