Philosophy, DNA, Morphology and Taxonomy

[raglanroad]

I'd like to explore the difference between your two replies, guys.
Bas sees this case as a failure of DNA and Larry sees it as a failure of Barcoding.
Why is Bas seeing it as weakness for any DNA analyses, and Larrry sees it as a weakness apparently for Barcodig in particular ?

I'm not familiar with this case, so I'm not aware of how this relects poorly on anything.

[Bas Pels]

The most common way to analyse DNA - such as done in forensic research - is to use a few enzymes (proteins) which cut DNA on a special place, in a certain sequence.

If I have somewhere CCATGG (A is opposite to T C to G, all such enzymes cut on this kind of places: those which have been mirrored this way, often referred as palyndromic sequences) and your code is CCAAGG, mine will be cut, and yours not

This will (assuming the rest is the same) in your DNA having 1 fragment less, and 1 fragement which is precicely as heavy as one of mine, combined with the missing one

Then the DNA is put into a solution with a certain pH (it's not my fiel, so I can't say the value) in which the DNA fragments are charged. Finally, the solution is put into a gel - allowing the passage of the fragments, but slowing them down - the heavier the more this friction does.

Then electricity is put on, and the charged DNA moves - depending on the wieght faster or slower - through the gel. After a certain amount of time, the electricity is put off, and some chemical is added to produce color where a fragment is

The result will be something alike a bar-code: dark lines on a white field. (quite often the barcode is turned 90 degrees, and again separated with other conditions, resulting in a square pattern, but the idea is basically the same)

As in this example you have 1 heavy fragment where I have 2, our barcodes will be dissimilar, and therefore we can say this is me and that is you

For species the same method can be used.

However, other techniques are also possible, bar coding is one of the DNA techniques I referred to in a more generalized way

I hope this explains a lot. A book on DNA techniques will explain it much better - but perhaps the details will hinder your understanding

[raglanroad]

Thank you, but even if a true fact clouds my understanding, I'm sure others wouldn't be so hindered.

http://en.wikipedia.org/wiki/Palindromic_sequence
Something is seen as a failure. What failure. By what means. And how does the Cryptoheros illustrate this failure occuring ?

Is this subject, which you are alluding to, "Biased Gene Conversion" ? it's neat subject for a chemist - chemical preferences rather than Natural Selection - as they may like that in their universe better..
http://www.scientificblogging.com/news_ ... esearchers
but I don't see any connection.

And Larry ? Why Barcoding in particular, about Cryptoheros ?

NA barcoding and the mediocrity of morphology - May 01, 2009 ()
[Packer, L., Gibbs, J., Sheffield, C., & Hanner, R. 2009. Molecular Ecology Resources. 9(s1) 42-50.]

A small but vocal community of critics has questioned the epistemological value of DNA barcoding by suggesting that either it 'cannot work' for the identification or discovery of species or that it ignores the 'richness' inherent in traditional approaches. We re-examine these arguments through a comparison of DNA barcoding and morphological taxonomy in terms of their accuracy and diversity of characters employed. We conclude that morphology often does not work and that it is often nowhere near as 'rich' as has been argued. Morphology is particularly poor in numerous important situations, such as the association of larvae with adults and discrimination among cryptic species. The vehemence of some of the criticisms is surprising given that morphology alone is known to be inadequate to the task of species-level identification in many instances.

[apistomaster]

I think Bas and I are basically talking about the same thing. how, where and which few protein encoding are selected may catch different sections being looked at by one researcher and not another even if the are close in molecular weight or electrostatic charges that begin the sorting and selection process. I don't think the concept is bad but I think it isn't yet perfected and a universal standard hasn't been decided on that would/could be the most useful for species specific identification.
There were some Costa Rican flies studied which were considered to be one species by bar coding but had definite lifestyle specializations and slight morphological differences. I may not have perfectly described that case but I am not expected to be an authority. I'm just interested. That example mirrored Bas' comments about the results in the Crytptoheros Clichlidae.
I really think this reflects the kinds of things that can be overcome as techniques become standardized and compact computing power increases and the programming algorithms are refined.
We have several interdependent sciences and fields of computer engineering involved and each has room for more refinement and systematists and taxonomists have to agree on the standards. I see it as an evolving science and will be in a state of change for sometime but progress is being made rapidly in all these fields. I am not arguing against bar coding but just pointing out that it is not yet to the point where it is the gold standard or be all and end all method, yet.
In criminal forensics, there are some stricter guide lines and a few well recognized markers known to be unique to an individual.
If we can reach a similar level of standardization that works at the species and higher level are identified for fish we would have less disagreement. I just think we have to allow this science to mature and meanwhile use it in conjunction with other techniques.
Eventually, it may literally be true that a compact device could be taken to the field than could analyze the DNA, take the GPS coordinates and whatever other collection data is customary to record. Even give it an SKU type bar code fore tracking purposes. Nice to do before the fish is pickled, mislabeled, misplaced, and then mishandled to pieces by future workers. I doubt that any one method will ultimately be the only method because it also depends on things which have always been debated or defined differently by different interest groups; ie: what is the definition of a species? This is one of those subjects that has always been something well meaning and honorable people may agree to disagree about like politics and religion.

[Mike_Noren]

Barcoders sometimes get a bit blinkered, though.
Defining species based on raw DNA distance of the COI gene, or trying to pin down a "universal" distance between species, is downright silly, and for some groups barcoding fails spectacularly, e.g. Sebastes. In fact, it often turns out that molecular methods fail in the same places morphological methods fail, and probably for the same reasons: rapid radiations or hybridization.
Also while barcoding is relatively cheap and fast, it's still not as fast or cheap as simply looking at the animal.

It's a source of data, neither better nor worse than morphology. By all means use it, but be aware that it's not a magic bullet.

[raglanroad]

yeah, everything has limtiations. Though it is cheaper to just look at things, looking at a flock of birds that appear identical doesn't show reason for splitting the flock into more than the one species until barcoding catches it..then further study finds out there are mating preferences that match the differences pointed to by barcoding...biological species concept comes into force again.
Or this, depending on your usage of it, no reason to slag it out of hand either.

http://www.fda.gov/Food/FoodSafety/Prod ... 079804.htm
so what's so devastating about Sebastes ?

[Mike_Noren]

looking at a flock of birds that appear identical doesn't show reason for splitting the flock into more than the one species until barcoding catches it

I find that to be rare. Nearly always when I find "cryptic" species it is already named, but as "variants", "subspecies", "geographical races", and it's really just a case of splitting vs lumping.

so what's so devastating about Sebastes ?

Try entering Sebastes norvegicus COI sequence into FishBOL and watch the resulting mess: half a dozen species have identical COI. This is probably real and not a case of mistaken identity, as several researchers have got the same result.
If you open the tree view of Sebastes you'll also notice another thing: some sequences are suspiciously divergent. I haven't investigated closer, and a wide range of issues may cause a situation like that, but one might guess some researchers have sequenced numts (those tend to amplify when the primers don't have a good fit to the real COI gene - and there is no such thing as a universal COI primer).

When the primers don't fit well contamination also becomes an issue. I recently barcoded a fish and to my surprise found that the sequence I got was from a species from another family, a fish which had never been within a thousand miles of the lab. It turned out my specimen had been caught in the _same seine_ as fish of that second species, and obviously had a few cells from it clinging to it, and as the primers didn't fit my target, they instead amplified the contamination.

But those are practical problems. Primers can be redesigned, other genes can be made to bear on species where COI doesn't work.

My main problem with the barcoding crowd is the tendency to confuse identification of species with definition of species. If you describe a species based on it's COI distance (which, one should remember, is a gene not found in the organism itself, but in an obligate intracellular symbiont)... what species concept is that?
There are a number of papers which investigate COI distance between species and the two things they have in common is that they agree that greater than a certain percentage divergence means two studied individuals belong to different species, and that they disagree, wildly, on just what that percentage is.

I'm not saying barcoding and molecular data is useless, I consider it a very powerful tool, but that one should be aware that molecular data is not inherently truer or less likely to be wrong than morphological data. Molecular data do have pitfalls, they just happen to be different ones from those of morphology. No magic bullet.

[raglanroad]

http://www.fda.gov/Food/FoodSafety/Prod ... 085481.htm

anyone in a market should know Ocean Perch fillet. It has a odd bouquet of ammonia developing more quickly than other like market-fish. no need to barcode them
sniff test is cheaper, I'd know that fish blindolded. Smell actually produced by organisms other than the fish ? ......could be, Doc ! After all, we have heard of forensics that look at the state of insect populations on a corpse, to determine time of death, right ?
It's that kind of situation. These kinds of things might be very reliable markers, but care is needed of course, before drawing inferences.
I may have read something about this, that barcoding was helpful actually, not a dismal failure at all.

Why question distance over obligate symbiont genes but not question some distance over a desert or mountain, for your species concept ?

Is there some philosophical problem just popping into people's heads now, about cause and effect ? It's a mirage every time...so often the Dr. Pepper machine on the sand dune is an RC Cola machine.
If the big boys announced they were taking on Hypancistrus for any other kind of study, I doubt there would be any of this consternation, just yelps of joy.

PS when fishermen step on a netted fish and squirt it's guts all over another fish, maybe it can get contaminated. As you say, some fish will not match correctly. Kind of like the ornamental fish that have had females described as different species from the males, after morphological assessment .

Any "view" of a subject comes with a "point" of view. Many different points of view and perhaps different chronological times of viewing offer a possibly much wider appreciation of reality, naturally.

and...everyone bemoans the paucity of Magic Bullets. But something that can spray a sheet of regular non magical hot metal at everything, is not so bad to have, when you need it. It's a powerul tool, and more importantly, more available for the public's usages and application.

I'm taking another look at BOLD because I think I recall Canadian sites were commenting on the success of barcoding on these fish.

[Suckermouth]

It seems that we are in agreement in that DNA barcoding should be used in conjunction with other techniques. Part of the issue of course is its novelty in the field, so its fairly debatable, for the same reason we use various phenetic and cladistic trees in our analysis. Many scientists you will see combine their molecular data with more traditional morphological study, and there is continuing study into how to best combine data from these different methods to determine species identities.

Why question distance over obligate symbiont genes but not question some distance over a desert or mountain, for your species concept ?

These are apples and oranges. A more comparable analogy would be why question the difference of genes and not the difference between morphological traits. The answer is that both are questioned. There is no definition of species distance in either morphological or molecular traits. This definition of species problem is not something that's inherent to molecular systematics.

Although, similar to how genetic differences will differ between different species, different barriers and different geographical distances will differ in separating species, so it's somewhat comparable.

I can tell you one thing, though, some scientists, especially the morphological systematists, are opposed to the idea of cryptic species. I'm not exactly opposed to it, but I would prefer if species defined through molecular systematics had further supporting evidence reflected in morphological traits so you can be "sure".

[raglanroad]

It seems that we are in agreement in that DNA barcoding should be used in conjunction with other techniques. Part of the issue of course is its novelty in the field, so its fairly debatable, for the same reason we use various phenetic and cladistic trees in our analysis. Many scientists you will see combine their molecular data with more traditional morphological study, and there is continuing study into how to best combine data from these different methods to determine species identities.

Why question distance over obligate symbiont genes but not question some distance over a desert or mountain, for your species concept ?

These are apples and oranges. A more comparable analogy would be why question the difference of genes and not the difference between morphological traits. The answer is that both are questioned. There is no definition of species distance in either morphological or molecular traits. This definition of species problem is not something that's inherent to molecular systematics.

It's better in that it is a neater riposte. I had then thought the question for me to reply to, had to be about the notion of dubious results from assessing difference in "the organism proper", by looking at difference in a symbiont gene. Indirect assessment is actually necessary anyway, for morphological study, because a phene difference cannot be taken as a direct measurement of a gene difference. So then I had to drag in the corpse and maggots anyway.

Although, similar to how genetic differences will differ between different species, different barriers and different geographical distances will differ in separating species, so it's somewhat comparable.

I can tell you one thing, though, some scientists, especially the morphological systematists, are opposed to the idea of cryptic species. I'm not exactly opposed to it, but I would prefer if species defined through molecular systematics had further supporting evidence reflected in morphological traits so you can be "sure".

It makes sense I think not to buy into individual incidents of "cryptic species" too easily.

[Suckermouth]

It's better only in that it is a neater riposte , but I thought the question for me to reply to, had to be about the notion of dubious results from assessing difference in "the organism proper", by looking at difference in a symbiont gene. Indirect assessment is actually necessary anyway, for morphological study, because a phene difference cannot be taken as a direct measurement of a gene difference. So then I had to drag in the corpse and maggots.

Okay, I see what you're getting at, again I'm arguing something different than what you were getting at. There are certain issues with relying on mitochondrial DNA (mtDNA), but the use of mtDNA is gaining acceptance in science, so I don't really see too much of a problem with it. However, again it is not a bad thing for morphological comparisons to be used as well to further validate a "species" found through molecular techniques, if only to skirt the debate that we are having right now about morphology vs molecules.

I think I will read the Packer paper you mentioned, and maybe discuss it with some people in my department.

[Mike_Noren]

I may have read something about this, that barcoding was helpful actually, not a dismal failure at all.

I'm not sure who you're arguing against. I certainly have not argued that barcoding in general is " a dismal failure".

Why question distance over obligate symbiont genes but not question some distance over a desert or mountain, for your species concept ?

I am not aware of anyone wanting to define species solely based on geographical distance.

Is there some philosophical problem just popping into people's heads now, about cause and effect ? It's a mirage every time...so often the Dr. Pepper machine on the sand dune is an RC Cola machine.
If the big boys announced they were taking on Hypancistrus for any other kind of study, I doubt there would be any of this consternation, just yelps of joy.

I have no idea what you're trying to say here.

PS when fishermen step on a netted fish and squirt it's guts all over another fish, maybe it can get contaminated. As you say, some fish will not match correctly. Kind of like the ornamental fish that have had females described as different species from the males, after morphological assessment .

And neither have I argued that morphological methods are infallible.

Any "view" of a subject comes with a "point" of view. Many different points of view and perhaps different chronological times of viewing offer a possibly much wider appreciation of reality, naturally.

THAT, on the other hand, is what I have argued.

[Bas Pels]

Personally I think studying both DNA and moprphological differences is not enough.

Say, one finds within 3 species (or samples) 2 to be quite alike, and 1 being much more different. Does that mean the differing species is less related?

That depends. If the differing species lifes in a way other environment - say not in a lake, but in a fast moving river, the truth MAY bey that the found differences are a direct result of adaptations needed to survive

Surviving in fast moving rivers means more muscle power, a bodyshape wich catches less current, perhaps another diet (certain foodsources are not available, such as Daphina) while others might be more abundent (shrimp, to do a suggestion) but lastly, the demands put on young fishes are higher, and often fish living in high current have larger fry

As we have seen evolution can go very rapidly, in a hundred years species can change a lot (see, for example the Lake Victoria disaster) many found differences can be reasoned away

So a good studie should answer that as well

[racoll]

There's a really good discussion I nearly missed here about modern systematics.

Mods, would it be possible to take the time to separate this from the "nonsense" about this Suzuki character, which is not particularly relevant?

Would be nice, so the evaluation of the science doesn't get buried.

Lots of remarks to discuss though. I will sleep, and then try to address some of the critical points.

[raglanroad]

Personally I think studying both DNA and moprphological differences is not enough.

Say, one finds within 3 species (or samples) 2 to be quite alike, and 1 being much more different. Does that mean the differing species is less related?

That depends. If the differing species lifes in a way other environment - say not in a lake, but in a fast moving river, the truth MAY bey that the found differences are a direct result of adaptations needed to survive

Surviving in fast moving rivers means more muscle power, a bodyshape wich catches less current, perhaps another diet (certain foodsources are not available, such as Daphina) while others might be more abundent (shrimp, to do a suggestion) but lastly, the demands put on young fishes are higher, and often fish living in high current have larger fry

As we have seen evolution can go very rapidly, in a hundred years species can change a lot (see, for example the Lake Victoria disaster) many found differences can be reasoned away

So a good studie should answer that as well

Bas, your argument is valid, so I'd like to point this paper out which gives answer to possible replies to your post that if they adapted to the fast river, they may indeed be a different species no matter what reasoning is used to justify the need for adaptation of the PHENOTYPE.
The two legged goat is a nice example because of the individual's phentoypic changes, so obviously possible without genetic change. However, genes doing the following instead of precursing a change, is acknowledged as an important element in evolution.

http://www.springerlink.com/content/pt4p488768710v01/

http://www.pnas.org/content/102/suppl.1/6543.full

[raglanroad]

There's a really good discussion I nearly missed here about modern systematics.

Where ? Link ?

Mods, would it be possible to take the time to separate this from the "nonsense" about this Suzuki character, which is not particularly relevant?

Would be nice, so the evaluation of the science doesn't get buried.

Lots of remarks to discuss though. I will sleep, and then try to address some of the critical points.

Mods, please, this is the "Speak Easy" section; I plead for not mangling the fun thread - and reply that those that can't take the wine with the gravy, need not slurp it up so noisily.
A slight problem lies with the subject title, that's all.
Perhaps we can have another thread with a poll, in order that all members could have a chance to give input as to the most dynamic and suitable title.


Sleep well, my friend, and let imagination run wild.

[MatsP]

I will edit the title to something more suitable - "Philosophy, DNA, Morphology and Taxonomy" would make some sort of sense, and perhaps get those interested in the thread to see the thread for what it is.

--
Mats

[raglanroad]

Thanks Mats.
Do you have something you could add in there, like a crown or pope's mitre ?

[Suckermouth]

There's a really good discussion I nearly missed here about modern systematics.

Where ? Link ?

racoll is referring to the discussion that we are engaged in.

[raglanroad]

Oh, OK. thanks.
BTW, I am trying to get a working link to the West/ Eberhard "Phenotypic Accommodation" paper.

there we go.

http://www.isnature.org/Files/West-Eber ... vation.pdf

[apistomaster]

Brook Trout are a species which seems to have the ability to react very differently to different environments even from the same strain used as initial stock. They become very stunted and grow to only six inches or so in low productivity waters, like alpine lakes, where they are able to breed. Same fish in a rich environment and no overpopulation may grow to 6 lbs. That is quite a bit of built-in plasticity.
Many of the Salmonids which evolved along advancing and receding glaciation have traits that allow them to adapt to great changes which occur within a short geological time frame.

[raglanroad]

At the Toronto zoo the curator witnessed their snail-eater African cichlidae have soft-food eater jaws instead of snail-cruncher jaws, within 2 generations of captive-raising.

[raglanroad]

This is from our discussion with Heiko on Cichlidae Forumae

Heiko:

As far as I know, there are no DNA tests done on P. altum - simply because the type material can not be used for DNA. One has to collect material at the Type locality in the Atabapo (and no place else). I shall do that also (I hope)...
...I have been in Paris and we searched for the type material of P. altum Pellegrin, 1903 and photographed it.

so about our Type Material for Hypancistrus zebra;

How was it collected by the biologists ?

In this paper a new genus of the family Loricariidae (subfamily Ancistrinae, tribe Ancistrini) is established for a new species, Hypancistrus zebra. It was collected for commercial (aquarium) purposes by staff of Tropicarium Para (Belem).

...ahhhh...wonder what kind of deal they got per box ?

[Suckermouth]

so about our Type Material for Hypancistrus zebra;

How was it collected by the biologists ?

In this paper a new genus of the family Loricariidae (subfamily Ancistrinae, tribe Ancistrini) is established for a new species, Hypancistrus zebra. It was collected for commercial (aquarium) purposes by staff of Tropicarium Para (Belem).

...ahhhh...

What's your point? Please be clear.

[raglanroad]

Point is that institutions take specimens for vouchering on faith -when they want to. Chain of possession could start with "local fish market" .

[raglanroad]

I will edit the title to something more suitable - "Philosophy, DNA, Morphology and Taxonomy" would make some sort of sense, and perhaps get those interested in the thread to see the thread for what it is.

--
Mats

Seeing as how Larry's interest in trout has led to almost a trout takeover in this thread ( an almost-takeover of which I do approve ) I'd like to suggest the title " Philosophy, DNA, Morphology, Trout, and Taxonomy" {crown or pope's mitre smilie}.

[MatsP]

Nah, I'm sure that's fine as it is.

--
Mats

[Mike_Noren]

This is from our discussion with Heiko on Cichlidae Forumae
Heiko:

As far as I know, there are no DNA tests done on P. altum - simply because the type material can not be used for DNA. One has to collect material at the Type locality in the Atabapo (and no place else).

This is inaccurate. The reason there's no published DNA sequences for P. altum is because the species is a royal pain in the ass to sequence. I know.

Also even if it wasn't so diverged that hardly any primers fit it, it still wouldn't matter that the type specimen can't be sequenced - most type material can't, it's too old or it's been treated with formalin, or both - as you simply have to show that the fresh fish you've captured (ideally at the type locality) conforms to the type, then sequence the fresh fish. That's pretty much what type material is for.

[Bas Pels]

Another example of plasticity:

An adult male Gymnogeophagus labiatum, photographed at the collection locality



I took 5 home, @ 5 cm. A year later, I took this picture, of a wild-caugt, tank raised male:



No big lips, much more flesh on the body, larger and a lower forehead (probably due to less sexual competition)

[raglanroad]

This is from our discussion with Heiko on Cichlidae Forumae
Heiko:

As far as I know, there are no DNA tests done on P. altum - simply because the type material can not be used for DNA. One has to collect material at the Type locality in the Atabapo (and no place else).

This is inaccurate. The reason there's no published DNA sequences for P. altum is because the species is a royal pain in the ass to sequence. I know.

Also even if it wasn't so diverged that hardly any primers fit it, it still wouldn't matter that the type specimen can't be sequenced - most type material can't, it's too old or it's been treated with formalin, or both - as you simply have to show that the fresh fish you've captured (ideally at the type locality) conforms to the type, then sequence the fresh fish. That's pretty much what type material is for.

Heiko was talking about DNA testing and lost, degraded, or never existant original materials; it's in Paris, in this case ( and it's not suitable for testing using usual techniques ) . And in this context, treating the ideal as a "must have", in order to resolve the question. That's entirely logical.

Logically, the question coming next would be "How did you determine your fish to be Pterophyllum altum?"
Also, anything with "D-loop" or microsatellites ?