Taxonomy as a Rigorous Science

[Dinwar]

What I have said is that the programs commonly (and traditionally) used for phylogeny reconstruction work equally well with morphological and genetic data...

I'm going further and saying that it necessarily must be that way. Characters are coded when they go into the equations, and once they're coded it doesn't matter what they were prior to coding, the math treats them all the same.

Surely, for fossilised hominid skulls, all we would have is precisely type specimens (or other specimens)?

Depends. The hominid fossil record is much better than most people realize; that said, we ARE talking terrestrial organisms who sometimes bury their dead and who lived in rift valleys. Not ideal locations for preservation. Still, we do have a fair number of specimens from some species, certainly far more than just the type specimens (to be fair, when I say "type" I mean holotype, not paratypes or the rest; you add those in and things get rather confusing). I've always admitted that the assumption that the hominid species weren't as diverse as modern humans was just that--an assumption. That said, when you hear hoofbeats, think horse, not zebra; unless there is some reason to assume that a species has an unusually high diversity in terms of morphology, we assume that it has similar diversity to other species in the taxa. Most apes do not have diversity anywhere NEAR that of humans, and a good chunk of human diversity is tied to the fact that we spread out. Anything that didn't spread out as far as H. sapiens sapiens, and which presents no evidence for being unusually diverse, will be treated as if it had normal morphological range in paleontology. That's the default. Sure, we're wrong on occasion--and hominids are no where near the most spectacular error we've made. However, we're right far more often than we're wrong, and we are always, always, always open to re-examining our conclusions in light of new data.

The alignment programs have an error rate of 1-2.5% depending on what you are looking at, how is that subjective?

As I have said repeatedly, and you have yet to even attempt to refute, the selection of the genes one wishes to examine is subjective. Even if you use the whole genome and enzymes to slice it into chunks (I forget what they're called) to run the bar-code tests, you've still got the choice of which enzyme to choose. I have seen that make significant differences. Unless you analyze the entire genome, base-pair by base-pair, you're dealing with subjectivity in the form of choice. We can mitigate that by re-running the tree based on what we learn from previous attempts, ie, learning from past attempts and removing obviously spurious data, but that still includes a great deal of subjectivity and--and this is the critical issue, which you insist on ignoring--this is true for ANY phylogenetic construction, based on ANY set of traits.

The ONLY difference between DNA and morphology in terms of phylogeny is that DNA has fewer characters to code if you go down to the base-pair level. After that, they're on a level playing feild as far as subjectivity/objectivity go. This has been demonstrated by the fact that they yield the same results.

I ask again: if morphology and genetics provide the same results (which they demonstrably do in at least some cases, particularly in the fossil record), why should we add genetic data? If the data will not give us any new information, why should we treat it as anything other than superfluous?

 

[Jodie]

And I'll say it again, you don't always get the same results with morphology versus genetics, that's why you can't always rely on morphology alone when it's not possible to obtain DNA. Convergent and parallel evolution can lead to similar morphology suggesting that certain animals are more closely related than they really are.

In the strands of DNA, you can figure out how the evolution of a particular organism happened. For example, the taxonomic status of the giant panda was contested for several years. Some placed it in the family Procyonidae, others placed it in the bear family. Red Pandas belong to neither family.

Look at skunks, they were in the wrong family up until 1997 when the mtDNA analysis was done that proved otherwise. The fossil record can be hit or miss. Fossils form only under certain conditions so we will always have an incomplete record if we rely on the fossil record alone to determine taxonomy and phylogeny. Here is an example that overlooks what the genetic evidence indicates:

http://news.nationalgeographic.com/news/2009/06/090623-humans-chimps-related.html

If these modern animals have results that aren't comparative with morphology versus DNA then it’s likely that some of the taxonomy of extinct organisms are wrong, hominids included. As you say, it depends on the sections of DNA you look at but with the new advancements in technology and the SNP chips, it ought to make that easier and more accurate to do. I don't see how you can put DNA in the same league as morphology when it comes to subjectivity. And even if DNA is better why does it matter? Isn't that a good thing? Why did you jump to the immediate conclusion that I was saying that morphology lacked rigor? For years it was all that was available. Anyway, my stance remains the same, I'm not budging because I know I'm right.

 

[Jodie]

Here are some articles, albeit a tad old, but I don't have access to anything more recent, that back up my what I'm saying:

http://depts.washington.edu/phylo/OlmsteadPubs/Scotland_et_al.SystBio.2003.pdf

http://www.scielo.br/pdf/bn/v10n2/35.pdf

 

[Kotatsu]

And I'll say it again, you don't always get the same results with morphology versus genetics, that's why you can't always rely on morphology alone when it's not possible to obtain DNA.

And as you don't always get the same results with nuclear and mitochondrial DNA, what's your solution then? Yo don't always get the same result with two different sections of the same mitochondrial genome, even, so what can you rely on in those cases? If you get the following results:

Morphology:
((A, B) (((C, D) E) (F, G) H))

Mitochondrial gene #1:
(((A, B) (F, G)) (((C, D) E) H))

Mitochondrial gene #2:
((A, B) (((C, D) E) (F, H) G))

Nuclear gene:
(((((A, B) F) G) H) ((C, E) D))

What do you recommend we should do?

Convergent and parallel evolution can lead to similar morphology suggesting that certain animals are more closely related than they really are.

While I am glad that you seem to belong to the small group of people (like myself) who see parallelism and convergence as two different processes, I would like to emphasize that both Dinwar and I have extensive training and education in these matters behind us, and do not need these terms explained.

In the strands of DNA, you can figure out how the evolution of a particular organism happened. For example, the taxonomic status of the giant panda was contested for several years. Some placed it in the family Procyonidae, others placed it in the bear family. Red Pandas belong to neither family.

No. In the strands of DNA, you can figure out how the evolution of the particular genes or regions you are looking at happened, which may or may not be the same as the evolution of the organisms. That is why there is such a thing as the "gene tree vs. species tree" discussion, and why we have things like coalescence methods and algorithms for working out the more probable evolutionary history of a group of organisms when the different genetic data sets say different things. Even if different genetic data sets only differ in branching dates, this can be very tricky to handle, especially if, as in my field, you're also dealing with host-parasite coevolution, where you have to juggle not only branching dates for two sets of organisms in various genetic data sets, but also have to take into account that e.g. host switches cannot happen between hosts that don't exist at the same time.

Look at skunks, they were in the wrong family up until 1997 when the mtDNA analysis was done that proved otherwise.

Systematics note: no analysis can even prove that a taxon belongs to a specific higher taxon. That is always a judgement call by the systematists and taxonomists involved. It is always possible to split a family down to monotypic families. What can be shows (not proved) is that a taxon is closely related to other taxa commonly placed in a certain family, based on a certain data set and a certain set of assumptions. That is not the same.

As you say, it depends on the sections of DNA you look at but with the new advancements in technology and the SNP chips, it ought to make that easier and more accurate to do.

Doesn't help when different genes say different things, though.

I don't see how you can put DNA in the same league as morphology when it comes to subjectivity.

Mainly because genetic data is just rapidly produced morphological data extracted from molecular structures, and because there are so many subjective elements that are the same regardless of what kind of data you are looking at.

Here's another matter: morphological specimens are typically deposited at museums or other institutions so that the character matrix can be checked independently. This is not necessarily the case with genetic vouchers, which may be extracted for DNA in their entirety (as with e.g. Gastrotrichs) or be deposited in such a way that you cannot get more DNA out of the specimen (for instance when slide mounting them).

 

[Jodie]

While I am glad that you seem to belong to the small group of people (like myself) who see parallelism and convergence as two different processes, I would like to emphasize that both Dinwar and I have extensive training and education in these matters behind us, and do not need these terms explained.

If I don't fully flesh something out with the assumption that any of you understand what I'm posting, the meaning gets twisted here, or I'm accused of ignorance.

Mainly because genetic data is just rapidly produced morphological data extracted from molecular structures, and because there are so many subjective elements that are the same regardless of what kind of data you are looking at.

I researched this from the forensic approach versus the genetic uses regarding disease processes ( which is what I'm familiar with) or taxonomy. I see what you are saying, this is frightening, to say the least. I guess more is published about the issues with databases from this perspective because it has more immediate serious consequences for individual people:

http://hstlj.org/articles/concerns-associated-with-expanding-dna-databases/

Here's another matter: morphological specimens are typically deposited at museums or other institutions so that the character matrix can be checked independently. This is not necessarily the case with genetic vouchers, which may be extracted for DNA in their entirety (as with e.g. Gastrotrichs) or be deposited in such a way that you cannot get more DNA out of the specimen (for instance when slide mounting them).

Shouldn't there be other records of the source specimen linking it to the profile? I would think that would definitely be a quality control issue. So how this relates back to bigfoot and the latest fiasco with the supposed bigfoot DNA project, Ketchum was pooh pooh'ed for lack of expertise in interpreting the data. Is it possible those criticizing her are wrong about their own interpretations that contradict what she found?

Yet here are some articles, the most recent that I could find, that back up what I've been saying about morphology versus genetics:

http://www.eurekalert.org/pub_releases/2013-03/ajob-dbf030413.php

http://www.agriculturejournals.cz/publicFiles/51736.pdf

But this has more relevance for the bear results that Sykes supposedly got when looking for proof of the Yeti, most likely you can't distinguish the difference between bear species living in the Himalayas just by looking at them:

http://www.biomedcentral.com/biome/...andro-minelli-on-identifying-cryptic-species/

Now what, how can we trust anything as reliable for evolutionary theory?

 

[Kotatsu]

Shouldn't there be other records of the source specimen linking it to the profile? I would think that would definitely be a quality control issue.

Well, as I said, there are many groups of organisms where it is simply not possible to get both a DNA sample and a voucher, because the organisms are so small, that in order to get DNA from them, you need to dissolve the entire specimen. The interview with professor Minelli you link to mentions this as well, for small slugs, but this is true for a range of organisms. Sure, you can take photos of the organism before sequencing it, but that does not necessarily solve the problem either, as there is no way to know what character sets may prove to be important for the systematics of the group in future studies. You'd need to take enormous amounts of very detailed photos and deposit these at a museum or something, and this evidently does not happen as often as it should.

In other cases, there is the problem of preservation of the voucher. When I sequence bird lice, for instance, I cut them in half and place both halves in the extraction fluid, then let this simmer for about 24 hours. This dissolves all the soft tissues inside the lice, but preserves the exoskeleton, which I can then mount on a microscopy slide and deposit as a voucher. However, once I've used up all the extracted DNA, there is no way of getting more. The exoskeleton is in Canada balsam, and is in any case empty. I would need to extract a second individual and somehow make sure that it is the same species as the first one, but if the reason I needed to extract a second individual is that there was some problem or ambiguity with the genetic data from the first, how do I ensure that they are the same species with genetic data?

It works different with larger organisms, of course, where you can take multiple tissue samples without exhausting the supply, but the majority of organisms -- especially when you're talking about cryptic organisms -- are too small to do that.

So how this relates back to bigfoot and the latest fiasco with the supposed bigfoot DNA project, Ketchum was pooh pooh'ed for lack of expertise in interpreting the data. Is it possible those criticizing her are wrong about their own interpretations that contradict what she found?

I have no idea, as I don't know what you're talking about. Is this that thread you liked to before, with the purported polar bears?

Yet here are some articles, the most recent that I could find, that back up what I've been saying about morphology versus genetics:

http://www.eurekalert.org/pub_releases/2013-03/ajob-dbf030413.php

I don't know whether this article back up what you're saying or not. As I see it, this appears to be a poorly studied group of plants that occur over large areas and in various biomes. When studied properly, it was found that one morphological character set (flowers) was found to be more readily adaptable than other morphological character sets ("vegetative and fruiting traits"), and these -- I assume -- were more in line with the genetic data. There is no way to assess the accuracy of these statements without reading the actual paper, which I don't have access to from home (and in any case, I hardly think I am qualified to assess this for plants...).

Taking this article at face value, what I see is that some morphological characters are in correspondence with some genetic characters, while other morphological characters do not give the same results as those particular genetic regions. They do not mention what genetic regions they have studied, and therefore it cannot be ruled out -- knowing, as we do, that different gene regions very often give different results -- that:
1. A different gene region that is directly relevant to inflorescence morphology may give results that are more in line with the morphological characters of the inflorescence;
2. A different gene region may give results that are at variance with the morphological data from vegetative and fruiting traits data, but still not in line with the morphological data from inflorescences.

Like I've said several times, it is very common that different gene regions give contrasting results. The gene regions they choose to work with in this article corresponded to some morphological data, but not other morphological data. From the link, we don't know how many regions they looked at (and it's not certain that we would find out by looking in the actual paper either, as we cannot rule out that they tested more regions than they published, but only some gave publishable results), and if all of these gave the same results or not.

I didn't read your other link.

But this has more relevance for the bear results that Sykes supposedly got when looking for proof of the Yeti, most likely you can't distinguish the difference between bear species living in the Himalayas just by looking at them:

Most likely, we cannot yet tell the difference by looking at them. I have no idea how well know the bear fauna in the Himalayas are, but for too many groups of purportedly cryptic species, it is just a matter of no one having looked at the material thoroughly enough to find the characters.

Now what, how can we trust anything as reliable for evolutionary theory?

This is a pretty big leap from what we've discussed before, given that phylogeny is only one of dozens of different sets of evidence for evolutionary theory. I don't see how anything we've discussed has any bearing on the reliability of evolutionary theory at all.

 

[DeiRenDopa]

I've been following this thread with considerable fascination; thank you Dinwar and Kotatsu, in particular, for really putting the E in JREF!

Plants have come up a few times, most recently in a post by Jodie. IIRC, neither Dinwar nor Kotatsu (nor anyone else?) claims particular expertise in plants (and both have explicitly said you need to dig into any particular taxonomy field in some detail to understand it). My limited understanding of plants (well, land plants anyway, and mostly angiosperms) bears this out; in particular:

a) mitochondrial, nuclear, and chloroplast phylogenies can be - and often - are inconsistent, at the two lowest taxonomic levels

b) detailed examination of morphologically-based taxonomies (?) with DNA tools has revealed quite a few inconsistencies (a recent example)

c) 'species' is a very tricky classification, for land plants, and even 'genus' can be hard to pin down, especially compared with, say, mammals (I think you know why, don't you Jodie?)

d) 'morphological parallelism' is far more common in land plants than it is in multi-cellular animals.

If we were able to get a plant systematics expert here, I'm sure they'd confirm what Dinwar and Kotatsu have been saying. They would also surely say that while molecular genetics turns up some really interesting results when applied to plants, the fact that leads to reclassification of many species (and even some genera), but essentially no higher classes, is exactly what you'd expect, given evolution and the history of (land) plant taxonomy.

 

[Kotatsu]

If we were able to get a plant systematics expert here, I'm sure they'd confirm what Dinwar and Kotatsu have been saying. They would also surely say that while molecular genetics turns up some really interesting results when applied to plants, the fact that leads to reclassification of many species (and even some genera), but essentially no higher classes, is exactly what you'd expect, given evolution and the history of (land) plant taxonomy.

Even worse: I have heard plant systematists say (on conferences and similar, so no links, sadly) that often mitochondrial DNA does not work at all to solve any kind of systematic puzzle, as it is apparently evolving either too slowly or too quickly (I forget which) to give any useful results. I have no idea if this has been limited to the specific groups of plants they were working on, or if this is a general problem for plant systematists, though.

What should be realised is that plant and fungi work in markedly different ways from most animals, and we should not expect that for instance genetic analyses of plant material should be directly comparable to that of animal material. Things like auto- and allopolyploidization, for example, is much more common in plant speciation than in animal speciation.

 

[Jodie]

So I guess that knocks out Mendelian genetics with sweet peas?

 

[DeiRenDopa]

So I guess that knocks out Mendelian genetics with sweet peas?

No.

Why not take an hour or so to research this yourself, Jodie? It's not that hard to understand why (the answer to your question is no), and by doing the work yourself you may gain a better grasp of the topic than reading what others have posted.

 

[Dinwar]

And I'll say it again, you don't always get the same results with morphology versus genetics

Two issues with this:

1) You continue to insist that we treat morphology as only part of the data, yet put no such restrictions on genetics despite clear and unambiguous evidence that genetics are every bit as problematic as morphology.
2) You have yet to explain why we must treat morphology as more tentative than genetics in those situations where the DO give the same answer.

You are still over-emphasizing genetics, and at this point it has gone beyond all rationality.

IIRC, neither Dinwar nor Kotatsu (nor anyone else?) claims particular expertise in plants (and both have explicitly said you need to dig into any particular taxonomy field in some detail to understand it).

Very much so, particularly across kingdoms. I've studied a number of animal phyla, but the extent of my plant knowledge is some intro courses in college and helping my grandfather on his farm. I know just enough to know that plants work very differently from animals!

They would also surely say that while molecular genetics turns up some really interesting results when applied to plants, the fact that leads to reclassification of many species (and even some genera), but essentially no higher classes, is exactly what you'd expect, given evolution and the history of (land) plant taxonomy.

Possibly. I'm not sure. Kotatsu does raise a good point--when you have multiple genomes, you can have multiple rates of evolution. Animals have mitochondrial, nuclear, and rhibosomal DNA if I recall correctly (could be rhibosomal RNA). Plants add a whole other genome to it. There's no reason to assume that they'd all evolve at a constant rate--which means you may have an organism with the same mitochondrial, chloroplast, or rhibosomal DNA as its ancestor. Or you could get multiple groups with different mitochondrial DNA. The fact that these aren't impacted by sexual reproduction (ie, mitochondrial DNA comes almost exclusively form the mother) makes the issues even more complex, particularly when you start with hybredization.

I'm sure a plant expert would be able to give answers to this sort of thing; I can't be the first to consider it! I just don't know enough to know what those answers are.

So I guess that knocks out Mendelian genetics with sweet peas?

Even if it did, there's still de Vrise' work on plants (he actually rediscovered Mendel's concepts independently). Second, the fact that plants work in ways that are weird to humans doesn't negate Mendel's work. Mendel selected traits that were more or less determined by single genes, likely through careful observation and careful selection of traits that could breed true (a key aspect of his experiments). The real issue with Mendellian genetics is that it ignores mutations.

Taking this article at face value, what I see is that some morphological characters are in correspondence with some genetic characters, while other morphological characters do not give the same results as those particular genetic regions.

I've never met anyone who does morphological analysis that would argue otherwise. Jodie is attempting to ignore the "derived" part of "shared, derived traits". What that term means is that the traits folks doing morphology look for are those that are passed down. In paleontology we focus on those that are passed from species to species, but you can't get passed from species to species without being passed from parent to child. What this means is that we need to focus on traits that are in some way heritable--often traits that are tied to genetics, but that's not absolutely necessary (epigenetics and behavioral traits can be useful). When we find that some traits aren't shared or derived, we tend to discard them because they're not useful and they clutter the dataset (this is all dictated by the math).

Assuming your summary is accurate (not saying I doubt you, just saying that I haven't read it), this amounts to little more than scientists doing science. This is empiricism in a nutshell: someone finds evidence that an assumption we've been making is flawed, and we change the assumption. The only other option is dogmatism. Jodie has neatly constructed her world-view such that nothing scientists do can be considered rational.

Most likely, we cannot yet tell the difference by looking at them.

Oh, I doubt that. "Just looking at them"? Sure, it'd be impossible. However, morphological analysis is not "just looking at them". It's a careful and rigorous process of measurement and analysis. I'm pretty sure that if we had bear specimens from that region we could determine which were from which species.

Though this does go to my earlier point: humans are absurdely diverse for a species.

 

[Kotatsu]

Possibly. I'm not sure. Kotatsu does raise a good point--when you have multiple genomes, you can have multiple rates of evolution.

This point is very important, Jodie, as it has several implications for lower-level systematics that may not be very obvious.

The most obvious of these implications is that if an organism has three genomes, each of which evolves essentially independently from each other, this may mean that the genome that evolves the fastest will be more likely to adapt to local circumstances over short evolutionary times, and these adaptations may become fixed much more quickly. When surveying a widely distributed species like this, we may therefore see some highly divergent results in one genome (in animals typically the mitochondrial) but no or very little differentiation in another (in animals typically the nuclear). The mitochondrial genome may suggest that these organisms constitute (say) 14 different, but closely related, allospecies, while the nuclear genome suggests that they are all the same. What would you do with this kind of problem? You could argue that the species are on their way to become separate species, of course, but taxonomy does not really allow for that sort of treatment (subspecies is not really an option here...), or you could chose one tree over the other, depending on if you are more or less conservative (a splitter or a lumper).

Having genomes that evolve at different rates within the same organism may also mean that the individual organisms may still be compatible sexually. If the mitochondrial genome is faster than the nuclear genome, but the mitochondrial genome is inherited entirely from one parent (usually the mother), and the nuclear genome is still more or less identical, two individuals from two different mitochondrial lineages that differ substantially may still be able to interbreed perfectly, because their nuclear genomes are still compatible. If this happens regularly but seldom enough, this may cause the phylogeny of the group to become entirely chaotic, with nuclear and mitochondrial data showing very different results. This is true even if such interactions happened only in the past, and the nuclear genome of the different populations has since actually diverged somewhat.

Lastly (and overlapping much with the two former points), there is the issue of incomplete lineage sorting. This is the case when a population of very mobile organisms (including organisms where only one life stage is very mobile) is split perhaps very suddenly. Prior to the split, various nuclear and mitochondrial lineages would be distributed in a more or less random fashion across the range of the species, so that when the split comes (perhaps from an external source, such as tectonic movements) the same lineages occur more or less randomly on all sides of the split. Given that a particular genetic marker has no or little impact on the relative survivability of the various lineages, this may mean that descendants of the same lineage will persist in the populations of both sides of the split over evolutionary time. When we do a genetic study of these organisms, we may thus find that certain individuals on population A are more closely related to certain individuals in population B than they are to any other members of population A for certain genes, but not for others. This effect is exacerbated by the different genomes evolving at different speeds, and can occur multiple times, so that we get multiple layers of incomplete lineage sorting. Looking at such a data set can be very problematic (I've fortunately never had it happen to my organisms, but I've been to enough workshops and conferences where they've shown examples...).

These, and many other factors, are things that forces the systematists to make subjective decisions. They may certainly be aided by all kinds of probability calculations and theory, but in the end it does come down to a subjective decision, and this is every bit as subjective as morphology.

 

[Jodie]

Two issues with this:

1) You continue to insist that we treat morphology as only part of the data, yet put no such restrictions on genetics despite clear and unambiguous evidence that genetics are every bit as problematic as morphology.
2) You have yet to explain why we must treat morphology as more tentative than genetics in those situations where the DO give the same answer.

You are still over-emphasizing genetics, and at this point it has gone beyond all rationality.

Very much so, particularly across kingdoms. I've studied a number of animal phyla, but the extent of my plant knowledge is some intro courses in college and helping my grandfather on his farm. I know just enough to know that plants work very differently from animals!

Well sure they do, and in my mind, you use genetics for identification if you don't have anything but blood, a small piece of tissue, that kind of thing. I wouldn't think you could use it all by itself since you can't tell what something looks like based on a genetic code without a type specimen. I've said that repeatedly.

Possibly. I'm not sure. Kotatsu does raise a good point--when you have multiple genomes, you can have multiple rates of evolution. Animals have mitochondrial, nuclear, and rhibosomal DNA if I recall correctly (could be rhibosomal RNA). Plants add a whole other genome to it. There's no reason to assume that they'd all evolve at a constant rate--which means you may have an organism with the same mitochondrial, chloroplast, or rhibosomal DNA as its ancestor. Or you could get multiple groups with different mitochondrial DNA. The fact that these aren't impacted by sexual reproduction (ie, mitochondrial DNA comes almost exclusively form the mother) makes the issues even more complex, particularly when you start with hybredization.

I thought mtDNA had a higher rate of mutation than the nuclear DNA based on what I've read. But I see Kotatsu has responded to this and will read what he says after I finish reading your post.

I'm sure a plant expert would be able to give answers to this sort of thing; I can't be the first to consider it! I just don't know enough to know what those answers are.

Even if it did, there's still de Vrise' work on plants (he actually rediscovered Mendel's concepts independently). Second, the fact that plants work in ways that are weird to humans doesn't negate Mendel's work. Mendel selected traits that were more or less determined by single genes, likely through careful observation and careful selection of traits that could breed true (a key aspect of his experiments). The real issue with Mendellian genetics is that it ignores mutations.

I doubt Mendel even considered that mutations existed at the time, wasn't it a new hypothesis?

I've never met anyone who does morphological analysis that would argue otherwise. Jodie is attempting to ignore the "derived" part of "shared, derived traits". What that term means is that the traits folks doing morphology look for are those that are passed down. In paleontology we focus on those that are passed from species to species, but you can't get passed from species to species without being passed from parent to child. What this means is that we need to focus on traits that are in some way heritable--often traits that are tied to genetics, but that's not absolutely necessary (epigenetics and behavioral traits can be useful). When we find that some traits aren't shared or derived, we tend to discard them because they're not useful and they clutter the dataset (this is all dictated by the math).

They do the same with genetics.

Assuming your summary is accurate (not saying I doubt you, just saying that I haven't read it), this amounts to little more than scientists doing science. This is empiricism in a nutshell: someone finds evidence that an assumption we've been making is flawed, and we change the assumption. The only other option is dogmatism. Jodie has neatly constructed her world-view such that nothing scientists do can be considered rational.

No, that is what you are saying, not me.

Oh, I doubt that. "Just looking at them"? Sure, it'd be impossible. However, morphological analysis is not "just looking at them". It's a careful and rigorous process of measurement and analysis. I'm pretty sure that if we had bear specimens from that region we could determine which were from which species.

I actually think this sounds like a good idea, it would be interesting to see how it compares to whatever it is Sykes is claiming in his peer reviewed paper. What little has been shared on the TV series mentioned something about a 40,000 year old polar bear jaw but that was in reference to a genetic match with the mtDNA, if I remember correctly.

Though this does go to my earlier point: humans are absurdely diverse for a species.

I don't disagree with you there.

 

[Jodie]

This point is very important, Jodie, as it has several implications for lower-level systematics that may not be very obvious.

The most obvious of these implications is that if an organism has three genomes, each of which evolves essentially independently from each other, this may mean that the genome that evolves the fastest will be more likely to adapt to local circumstances over short evolutionary times, and these adaptations may become fixed much more quickly. When surveying a widely distributed species like this, we may therefore see some highly divergent results in one genome (in animals typically the mitochondrial) but no or very little differentiation in another (in animals typically the nuclear). The mitochondrial genome may suggest that these organisms constitute (say) 14 different, but closely related, allospecies, while the nuclear genome suggests that they are all the same. What would you do with this kind of problem? You could argue that the species are on their way to become separate species, of course, but taxonomy does not really allow for that sort of treatment (subspecies is not really an option here...), or you could chose one tree over the other, depending on if you are more or less conservative (a splitter or a lumper).

Having genomes that evolve at different rates within the same organism may also mean that the individual organisms may still be compatible sexually. If the mitochondrial genome is faster than the nuclear genome, but the mitochondrial genome is inherited entirely from one parent (usually the mother), and the nuclear genome is still more or less identical, two individuals from two different mitochondrial lineages that differ substantially may still be able to interbreed perfectly, because their nuclear genomes are still compatible. If this happens regularly but seldom enough, this may cause the phylogeny of the group to become entirely chaotic, with nuclear and mitochondrial data showing very different results. This is true even if such interactions happened only in the past, and the nuclear genome of the different populations has since actually diverged somewhat.

Lastly (and overlapping much with the two former points), there is the issue of incomplete lineage sorting. This is the case when a population of very mobile organisms (including organisms where only one life stage is very mobile) is split perhaps very suddenly. Prior to the split, various nuclear and mitochondrial lineages would be distributed in a more or less random fashion across the range of the species, so that when the split comes (perhaps from an external source, such as tectonic movements) the same lineages occur more or less randomly on all sides of the split. Given that a particular genetic marker has no or little impact on the relative survivability of the various lineages, this may mean that descendants of the same lineage will persist in the populations of both sides of the split over evolutionary time. When we do a genetic study of these organisms, we may thus find that certain individuals on population A are more closely related to certain individuals in population B than they are to any other members of population A for certain genes, but not for others. This effect is exacerbated by the different genomes evolving at different speeds, and can occur multiple times, so that we get multiple layers of incomplete lineage sorting. Looking at such a data set can be very problematic (I've fortunately never had it happen to my organisms, but I've been to enough workshops and conferences where they've shown examples...).

These, and many other factors, are things that forces the systematists to make subjective decisions. They may certainly be aided by all kinds of probability calculations and theory, but in the end it does come down to a subjective decision, and this is every bit as subjective as morphology.

Well that's the best explanation I've read yet, and exactly what this Texas veterinarian claimed. Todd Disotell actually said her conclusions made the most sense, but I couldn't understand why. There were other factors involved that made it utter nonsense but now I understand why she went in that direction. Do you care if I link this to another forum?

 

[MikeG]

......... the mitochondrial genome is inherited entirely from one parent (usually the mother),.........

I was going to ask if it wasn't always the mother, but I guess I was being large-animal-centric. I assume you are talking about hermaphrodites? Are there other circumstances in which mtDNA comes down through a non-wholly-female line?

 

[Dinwar]

Well sure they do, and in my mind, you use genetics for identification if you don't have anything but blood, a small piece of tissue, that kind of thing. I wouldn't think you could use it all by itself since you can't tell what something looks like based on a genetic code without a type specimen. I've said that repeatedly.

Okay. So you've completely abandoned the insane notion that morphology is more subjective than genetics. Good--we are making headway!

Now, please address the myriad of issues the resident taxonomist raised regarding genetics.

I doubt Mendel even considered that mutations existed at the time, wasn't it a new hypothesis?

There's a reason I mentioned de Vrise. Mendel wasn't the only person doing such work--and until de Vrise, his work languished in obscurity. de Vrise' primrose experiments demonstrated quite conclusively that mutations are possible (due to self-fertilization and alterations of traits that bred true for generations, along with some of the most rigorous documentation in science).

They do the same with genetics.

See, now, when I point that out--with the data to support it--you scream your head off. I suggest that you go back and re-read the first two pages of this discussion.

No, that is what you are saying, not me.

~sigh~ Jodie, this is a medium that records conversations. If you wish to pretend you haven't made the arguments you have, you will only undercut your own credibility.

I actually think this sounds like a good idea, it would be interesting to see how it compares to whatever it is Sykes is claiming in his peer reviewed paper.

"Sounds like a good idea"? You have GOT to be kidding me!!!! What I described is morphological analysis, that thing you've been saying is too subjective to be usedful. Now that you're in full evasion mode, attempting to make it seem that you've always been arguing what you've only recently come to accept, I doubt you'll acknowledge your previous statements, but that's the fact of the matter. And if you think that what I described "sounds like a good idea", you are too ignorant of how morphology is analyzed to comment on anything about the field.

I don't disagree with you there.

There is no way to disagree with that statement and even pretend to be rational. What you have failed to consider are the implications of it.

 

[Kotatsu]

I wouldn't think you could use it all by itself since you can't tell what something looks like based on a genetic code without a type specimen.

Again, this has nothing to do with having a type specimen. A type specimen is a very particular (usually preserved) individual of a species. It is possible to identify all kind of organisms without looking at a type specimen (whether the holotype or some other form of type). That is, after all, why we taxonomists write and publish species descriptions, keys, and revisions: to allow people to identify organisms without having to visit the place where the types are located. What you mean, I think, is still a voucher specimen (as well as a general reference collection).

Well that's the best explanation I've read yet, and exactly what this Texas veterinarian claimed. Todd Disotell actually said her conclusions made the most sense, but I couldn't understand why. There were other factors involved that made it utter nonsense but now I understand why she went in that direction. Do you care if I link this to another forum?

I have no idea who this Texas veterinarian is, nor who Todd Disotell is, but I have no problems with linking to this.

I was going to ask if it wasn't always the mother, but I guess I was being large-animal-centric. I assume you are talking about hermaphrodites? Are there other circumstances in which mtDNA comes down through a non-wholly-female line?

I don't know how it works in non-animals (I also tend to ignore plants and stuff...), but I have heard from several sources that the muscles that close clam shells have different mitochondria from the rest of the animal, in that they are inherited from the father, while the mitochondria of the rest of the clam is inherited from the mother. I believe I have heard other examples (Ctenophores?) where the inheritance is from both parents, but I can't remember at the moment. As you say, any hermaphroditic organism may be a good place to start, but there may be other groups.

 

[DeiRenDopa]

I was going to ask if it wasn't always the mother, but I guess I was being large-animal-centric. I assume you are talking about hermaphrodites? Are there other circumstances in which mtDNA comes down through a non-wholly-female line?

Google is one of my BFFs.

In less than a minute, I found "Inheritance and recombination of mitochondrial genomes in plants, fungi and animals." (a.k.a. Barr et al. (2005), linky):

It is generally assumed that mitochondrial genomes are uniparentally transmitted, homoplasmic and nonrecombining. However, these assumptions draw largely from early studies on animal mitochondrial DNA (mtDNA). In this review, we show that plants, animals and fungi are all characterized by episodes of biparental inheritance, recombination among genetically distinct partners, and selfish elements within the mitochondrial genome, but that the extent of these phenomena may vary substantially across taxa. We argue that occasional biparental mitochondrial transmission may allow organisms to achieve the best of both worlds by facilitating mutational clearance but continuing to restrict the spread of selfish genetic elements. We also show that methodological biases and disproportionately allocated study effort are likely to have influenced current estimates of the extent of biparental inheritance, heteroplasmy and recombination in mitochondrial genomes from different taxa. Despite these complications, there do seem to be discernible similarities and differences in transmission dynamics and likelihood of recombination of mtDNA in plant, animal and fungal taxa that should provide an excellent opportunity for comparative investigation of the evolution of mitochondrial genome dynamics.

True, it does not actually answer your question (but it's pretty cool, eh?)

I have absolutely no idea how solid a reference this is ...

 

[Jodie]

Okay. So you've completely abandoned the insane notion that morphology is more subjective than genetics. Good--we are making headway!

Now, please address the myriad of issues the resident taxonomist raised regarding genetics.

My take on this is it depends on what you are using the genetic codes for, be it medical research, forensics, or what have you. Until I looked at the forensic research I didn't find much about the issues with DNA identification. When I did go that route, it more or less said the same things you and Kotatsu said about it's uses with taxonomy. I read further down to Kotatsu's post where he corrects me, yes, I meant voucher specimen. I wouldn't think genetic code would help you much in trying to describe the morphology of a cryptid without a voucher specimen.

There's a reason I mentioned de Vrise. Mendel wasn't the only person doing such work--and until de Vrise, his work languished in obscurity. de Vrise' primrose experiments demonstrated quite conclusively that mutations are possible (due to self-fertilization and alterations of traits that bred true for generations, along with some of the most rigorous documentation in science).

See, now, when I point that out--with the data to support it--you scream your head off. I suggest that you go back and re-read the first two pages of this discussion.

Dinwar, I don't think at any point that I screamed my head off, but you got rather nasty at times.

~sigh~ Jodie, this is a medium that records conversations. If you wish to pretend you haven't made the arguments you have, you will only undercut your own credibility.

In my field, genetic code is considered reliable for research regarding genetic diseases. Now I am left wondering about Angelina Jolie's decision to get a double mastectomy just because her genetic screening revealed she had the BRCA1 mutation. Would I make that same decision in her place? I'm not so sure.

"Sounds like a good idea"? You have GOT to be kidding me!!!! What I described is morphological analysis, that thing you've been saying is too subjective to be usedful.

Hold up there Mister, I meant in conjunction with genetic analysis. You are taking what I said out of context, but since this is a discussion with lapses of a day or two while I try to catch up on the reading, I won't hold it against you for loosing the train of the conversation.

Now that you're in full evasion mode, attempting to make it seem that you've always been arguing what you've only recently come to accept, I doubt you'll acknowledge your previous statements, but that's the fact of the matter. And if you think that what I described "sounds like a good idea", you are too ignorant of how morphology is analyzed to comment on anything about the field.

You just can't help yourself can you? You have tried to twist what I've said since the thread started, but I still don't think you can look at ancient hominid skulls, without comparative DNA from those skulls, and get a full picture of who is related to who. So shoot me.

There is no way to disagree with that statement and even pretend to be rational. What you have failed to consider are the implications of it.

Implications of what? Is the world going to come to an end because I have a different opinion? Don't think so.

 

[Kotatsu]

I wouldn't think genetic code would help you much in trying to describe the morphology of a cryptid without a voucher specimen.

... which would then, assuming this is an actual cryptid, most likely become the holotype specimen...