What is a species? (2 Viewers)

[njlarsen]

But the counter to this is that if mutations occur per x replications, the larger the populations the more chance of a mutation occurring. There is probably a complex model that balances chance of mutation verses likelihood of becoming influential

The best estimate if i remember correctly is that each of us human beings are born with 20-40 sequence changes compared to our parents. Most of these means nothing to our outcomes, but occasionally there is one that does. But for the population, it’s a big difference between that being one in a million or one in one thousand individuals carrying the change that causes a difference.
Niels

 

[njlarsen]

Presumably also some gene mutations are dominant? I can’t remember the ‘rules’ that determines human eye colour, but I seem to remember brown is dominant. If my unscientific understanding is correct, then a mutation of a dominant gene should spread more quickly than a mutation of a non-dominant one.

The spread in the population should happen at about the same rate, unless the effect of the mutation is very large on survival or fertility. The effect of the mutation should be detectable earlier with a dominant mutation.
Niels

 

[THE_FERN]

Interesting that rates vary. Doesn’t this make the calculation of when a species diverged highly complex? I.e a calculation of what mutations have occurred in which parts of the chromosome and at what assumed rate?

Different genes (and parts of genes) evolve at different rates. For example, those bits which are responsible for a function tend to be more conserved than the bits in between which produce no functional result (introns). Those may evolve more or less at random.

An individual is a bundle of lots of bits of genetic material which have diverged to a greater or lesser extent from those of other individuals. The same is true of species. So 2 species may have some genes which have not or have hardly diverged at all and others which have diverged a lot.

We define a "species" with reference to some notional divergence criterion. Depending on the species concept you use that might put more weight on some changes more than others. Biological species are those which can't interbreed with other such species so this concept weights genes involved in mate selection very highly.

The studies you read are only able to look at a relatively few genes in detail. So they use this to estimate (guesstimate) how divergent individuals are overall. Sometimes the differences in rates of divergence are very clear in these studies—for example when mitochondria and nuclear genes indicate different relatedness patterns.

All in all, this just underlines the fact that defining what is or isn't a species will remain an arbitrary and subjective thing. That's even more true as you go to higher biological categories like genera and families. It's an inevitable consequence of applying a categorical taxonomy to something which varies more or less continuously.

 

[nartreb]

Interesting that rates vary. Doesn’t this make the calculation of when a species diverged highly complex? I.e a calculation of what mutations have occurred in which parts of the chromosome and at what assumed rate?

Presumably also some gene mutations are dominant? I can’t remember the ‘rules’ that determines human eye colour, but I seem to remember brown is dominant. If my unscientific understanding is correct, then a mutation of a dominant gene should spread more quickly than a mutation of a non-dominant one.

"Dominant" refers to the gene's effect on phenotype, and does not imply anything about the rate of spread. Eye color is a good example.* Blue eyes have no pigment at all (they diffract blue, like the sky). Brown eyes come from a dark pigment in the iris (it's a melanin, somewhat similar to the pigment that darkens human skin). Recall that vertebrates have two copies of each chromosome (let's ignore the sex chromosome) in every somatic cell. If either chromosome carries a working gene for that pigment**, then the individual will have brown eyes. Only if both chromosomes have a nonfunctional version of that gene, will your eyes be blue. So, just one copy of the brown version "dominates" the blue version.

As for how quickly the gene will spread, well, let's assume that two brown-eyed people mate. Unbeknownst to them, they are each heterozygous - they each carry one copy of the blue-eye allele. Then 25% of their offspring will be homozygous blue (and have blue eyes), 25% will be homozygous brown (have brown eyes, and pass on only the brown-eye gene, meaning all of their children will be brown-eyed) and 50% will be heterozygous (brown-eyed but carrying the blue-eye allele). Those are exactly the proportions you get for any couple where both are heterozygous: each child has a 50% chance of getting one allele or the other from each of its parents. You can work out the odds when two blue-eyed people mate (100% homozygous blue-eyed children) or a brown-eyed homozygous person mates with a blue-eyed person (all brown-eyed, heterozygous children) and so on, but the point is that the spread of the genes has nothing to do with which allele dominates the phenotype. The proportion of an allele in a population is subject to the Hardy-Weinberg Law, and "will continue unchanged after the second generation" as Hardy said.

Unless, of course, the phenotype is important for survival. A helpful allele can spread faster if it is dominant: more of the children express it, and survive to have even more children of their own, at least half of which also express it. Conversely, a harmful dominant allele can be extirpated more quickly than a recessive one. If all carriers of a gene die young, then that gene will cease to exist within one generation. Whereas recessive alleles can "hide", unexpressed, in the population, and be expressed in a proportion of the next generation. (That's one reason you shouldn't marry your sibling.)

* It's of course not really just determined by one gene with two alleles one of which is dominant. But it's a good first approximation.

** A majority of blue-eyed people have a mutation in a gene called OCA2, on chromosome 15. It's been surmised that they all share a common ancestor who lived about 10,000 years ago.

 

[njlarsen]

All of the above is true except that eye color is more complex than described. The paler eyed persons can be blue, gray or green, and there is a definite difference in how dark brown a dark-eyed person is (hazel, brown, almost black, etc. ) Therefore, the best description of eye color is multifactorial.
Niels

 

[Jacana]

Interesting that rates vary. Doesn’t this make the calculation of when a species diverged highly complex? I.e a calculation of what mutations have occurred in which parts of the chromosome and at what assumed rate?

Presumably also some gene mutations are dominant? I can’t remember the ‘rules’ that determines human eye colour, but I seem to remember brown is dominant. If my unscientific understanding is correct, then a mutation of a dominant gene should spread more quickly than a mutation of a non-dominant one.

Yup, speciation is a messy, complex process which rarely follows the same path from one circumstance to the next.

 

[Kirk Roth]

All of the above is true except that eye color is more complex than described. The paler eyed persons can be blue, gray or green, and there is a definite difference in how dark brown a dark-eyed person is (hazel, brown, almost black, etc. ) Therefore, the best description of eye color is multifactorial.
Niels

When I was teaching biology about 20 years ago, the misconceptions about dominant traits somehow relating to evolution (meaning changes in gene frequencies) was common enough for us to huddle around strategies to explain. Rather than eye color, what seemed to click easiest was using polydactyly as an example. Having six fingers is a dominant trait and runs fairly close to simple Mendelian Punnett's Square style inheritance. However, there aren't six-fingered folks everywhere you look - its a rare trait that conveys no advantage to successfully producing offspring.

Eye color is far messier - a truthful story needs to encompass the fact that there is a gene for melanin and another for modulation, each with quirks of expression plus other genes, mutations, and environmental factors which affect eye color - as in my heterochromatic son. Then, work on explaining how particular eye colors might cause someone 10,000 years ago to have more grandchildren.... that is a can of worms that is not much fun to explain to a college freshman, or the general public for that matter!

But in a small scale way, the six-finger vs. eye-color discussion is useful in considering the difficulty of using genetic distance as a marker of speciation. Some genetic differences just don't matter, while some may matter anywhere from slightly to intensely to reproduction of those genes. If two populations have a "genetic distance" of 2-3%, but all of those are genes that don't matter to survival or reproduction or even aren't expressed as traits - what does it matter? But if two fly populations have a difference of only a single gene which happens to dictate what fruit they will reproduce on, or if populations of ground squirrels or goldenrods have either two or four sets of the same genes - those can be key differences.

 

[njlarsen]

When I was teaching biology about 20 years ago, the misconceptions about dominant traits somehow relating to evolution (meaning changes in gene frequencies) was common enough for us to huddle around strategies to explain.

I can tell you that misconception still is prevalent when the same students make it into medical school where I am teaching genetics. I may steal the example of polydactyly from you!
Niels

 

[Welsh Peregrine]

I remember using polydactyly as the same example in the same timeframe; interestingly I don’t recall the misconception coming up in the last 15 years or so, not because the students no longer make it, but because the specification has moved away from it being relevant to answering current exam questions!

 

[THE_FERN]

But if two fly populations have a difference of only a single gene which happens to dictate what fruit they will reproduce on, or if populations of ground squirrels or goldenrods have either two or four sets of the same genes - those can be key differences.

Of course it's rarely apparent or easy to determine whether a difference is "key" or not. So usually people default to whatever genetic or character distance information is at hand: "We don't [really] know how relevant it is but of the (few usually) genes we've looked at there's a sequence divergence of 5%; we've decided this means they're 2 different species".

Given that, as far we know, most speciation is quite gradual then we're bound to come across a spectrum of degrees of divergence. Imposing a categorical "this is/not a species" is always going to be arbitrary, therefore. This is true whichever type of characters you work with—molecular or phenetic.

 

[Jon.Bryant]

Rather than eye color, what seemed to click easiest was using polydactyly as an example.

Can any examples using humans really be applied to nature in general? We are a unique species in that we have created a society where physiology is not important to survival. We also are rather non selective in pairing - yes we fill our fashion magazines with stereotypes of beauty, but mostly there is someone out there for everyone - fat or thin, short or tall, five or six fingers, green, blue or brown eyes. Perhaps the Spartans would have eliminated polydactyl, as allegedly they put to death and infants with any sign of ‘weakness’.

Presumably it is not all about pure survival, but also sexually pairing and reproduction. I am sure that is we battled for a harem like an Elephant Seal beachmaster, our physique would have evolved to be different.

Equally if women were as selective as female Flame Bowerbirds, many of us less perfect examples of the men would simply never pair.

In highly selective species, presumably a variant in plumage may be success or disaster - if green eyes increased the chances of pairing 4 to 1, then surely it would make a difference to the balance of genes controlling human eye colour.

I can imagine that colour mutations in highly selective species could be quickly incorporated across a small island population. But surely this would just be a ‘morph’ and not a new species.

Going back to my original point, can we use genetic difference to define a species?, Can we accurately estimate the time taken for this genetic difference to have accumulated? and most importantly, can this timeframe be align with our understanding of geological timeframes - i.e do out theories pass the sniff test?

From reading the thread above, I suspect that difference in DNA cannot be used as a definition of a species, and that time frames for accumulation of genetic differences will vary depending on the genes involved, so are at best a crude measure of how long ago divergence occurred.

 

[Frank van Drogen]

so are at best a crude measure of how long ago divergence occurred.

Generally, non-coding chromatin, without presumed function or selective pressure, is used to obtain temporal information. Not perfect, but relatively reliable.

 

[Jon.Bryant]

Generally, non-coding chromatin, without presumed function or selective pressure, is used to obtain temporal information. Not perfect, but relatively reliable.

In Oriental Bird Club Taxonomic Summaries, they reference years of separation, which presumably are stated in the original paper arguing for species status.

Do you know if author’s of Scientific papers can be ‘relied upon’ to use non-coding chromatin, without presumed function or selective pressure? Or are calculations likely to be less accurate?

The OBC region includes many island endemics and it would be interesting to look at geological time frames for island formation in south east Asia, as a comparison.

 

[Frank van Drogen]

In Oriental Bird Club Taxonomic Summaries, they reference years of separation, which presumably are stated in the original paper arguing for species status.

Do you know if author’s of Scientific papers can be ‘relied upon’ to use non-coding chromatin, without presumed function or selective pressure? Or are calculations likely to be less accurate?

The OBC region includes many island endemics and it would be interesting to look at geological time frames for island formation in south east Asia, as a comparison.

Would you have a link to a paper?

(Many different methods actually. With different advantages and disadvantages. Wouldn't know what they use).

 

[njlarsen]

This should be detailed in methods section of each paper. The oldest version of all of this used a specific section of the Mitochondrial DNA which is where a 2% rule arose from, and was shot down many times since. In nuclear genes, it is often introns that are used, with the assumption that there is very little function hidden in such areas. However, it is clear that the rate of change differs from different types of markers and that the precision with which one can calculate a time from last common ancestor differs between different markers and different time periods. The uncertainties that should be apparent from the above is why you sometimes see statements about having used fossils to calibrate the molecular clock -- and if I was not in deep water explaining the above, explaining how to know for sure the age of a fossil is even further beyond me.
Niels

 

[THE_FERN]

This should be detailed in methods section of each paper. The oldest version of all of this used a specific section of the Mitochondrial DNA which is where a 2% rule arose from, and was shot down many times since. In nuclear genes, it is often introns that are used, with the assumption that there is very little function hidden in such areas. However, it is clear that the rate of change differs from different types of markers and that the precision with which one can calculate a time from last common ancestor differs between different markers and different time periods. The uncertainties that should be apparent from the above is why you sometimes see statements about having used fossils to calibrate the molecular clock -- and if I was not in deep water explaining the above, explaining how to know for sure the age of a fossil is even further beyond me.
Niels

If one's interested in the molecular clock (2% per million years) I'd urge you to read the nature review I linked. Afaik, this rough figure---which will obviously vary somewhat depending on circumstances---is still generally used as a yardstick

 

[jurek]

Generally, using the genetic difference, for any species definition and level of difference used, assumes that the selection which produced the difference acts with the same strength on different bird populations and genes. Which, after a second thought, is obviously wrong.

The only exception is a genetic difference of populations co-existing and able to hybridize - if they don't, there must be an obvious barrier to the gene flow.