scigirl

The evolution news has returned. . .

There is an interesting news article in the April 18 issue of <a href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=119615 36&dopt=Abstract" target="_blank">Nature</a>, titled "The tale of the parasitic cuckoos" by AJ van Noordwijk.

There are 136 species of cuckoos. Around 50 of them are considered brood parasites, which means they lay their eggs in other birds' nests, and are therefore raised by the original nest-builder. The other species of cuckoos are non-parasites since they raise their own young, but do occasionally try to lay eggs in other birds' nests.

Selection has therefore created an 'arms race' between the cuckoos, who try to lay eggs that closely resemble the host's eggs, and the hosts, who evolved to recognize foreign eggs. Understanding how this relationship started in the first place has been a bit of a puzzle, but now researchers may have found some explanations, and more importantly, ways to test them.

Researchers Kruger and Davies hypothesized that the ancestors of the parasitic birds were non-migratory, and that they laid big eggs which they cared for themselves. So they analyzed traits related to ecology and life history for all of the cuckoo species to see which traits predated the evolution of parasitism and which ones evolved after.

First, they needed a reliable phlylogeny (a bifurcating tree showing ancestral relationships, basically an "evolution family tree." They ended up using two separate phyogenies previously described by other labs (one based on bone traits, the other based on DNA sequences). Note here that phylogenetic analysis can be tricky, and these two phylogenies were not identical.

Then they would take one trait, say egg size, and run it through the above trees.




The researchers carried out this analysis for 13 traits, and found that with the bone phylogeny, 34 contrasts could be made, and for the molecular tree, 19 contrasts could be made. Despite the differences with the two phlogenies,

They speculated that a biological reason for these traits is that a parasitic cuckoo can afford to expend less energy and inhabit a less productive habitat, because it does not have to feed its own young.

This type of study is important because it attaches probabilities to each transitional step in potential evolutionary sequences. What this means is, we can perhaps start to define causal events in evolution. Then we can make predictions and look at similar systems (say, insect parasitism), and see if the predictions hold up. If they do, than we can say with more confidence that, "X evolved because Y occured."

Another important finding from this study was the number of intermediate steps discovered between obligate parasites and birds that always raise their own young.

scigirl

[ May 10, 2002: Message edited by: scigirl ]

[ May 10, 2002: Message edited by: scigirl ]</p>

Dr.GH

Thanks for the information. What is the basis for the argument that each branch does not necessarily mark an evolutionary event? It would seem to my undertanding that each brance must represent at least one event, and perhaps more than one.

Also, how did the authors deal with the possibility of extinctions masking evolutionary events.

scigirl

Hello Dr. GH,

I don't think that's what the article was saying. They used the phylogenies (the trees whose branches obviously represent speciation) to test their theories whether certain traits (i.e. egg size) could explain causative patterns of speciation. An analogy could perhaps be when scientists look for a genetic disorder in a family tree: if they think that person X contained the original mutation, then people born before person X shouldn't have it, and neither should the branches that did not directly descend from person X. Does that make sense? Although I could have also misunderstood the article. But the cuckoo analysis is more complicated, since they were not only looking for several non-independent trait, but also looking for traits which could have caused the speciation in the first place. Of course, we can never "prove" that X caused Y (because we can't go back in time and observe cuckoo evolution), but this type of analysis is one way to address causation. Make some predictions, see if they hold up when you look at the phylogeny, then apply these same predictions to other similar systems. I hope this is the interpretation of the article anyway, because that's how I understood it to be!

Very good question. It didn't say. I think I will eventually read the original article (this was just a summary in Nature, I think the orig in is PNAS), since I find this topic a fascinating one. I'll keep you posted.

scigirl

lpetrich

I'd like to thank "scigirl" for her evolution update.

One of my interests has been evolutionary biology; the evolution part provides an organizing principle in what would otherwise be a chaos of features.

One interesting thing I discovered long ago, back around 1980, was a discussion of molecular biology which featured a bit of discussion of molecular evolution, notably a family tree of the molecule cytochrome c.

Most of the more recent branchings in that tree agreed with what was worked out with macroscopic features and the fossil record, giving confidence that earlier branches are reasonable deductions. And since the more recent branchings had happened at an approximately clocklike rate, one might be able to extrapolate back in time.

The earliest calibration point was insects and vertebrates at about 600 million years ago, a little below the base of the Cambrian.

But what was most interesting was that animals, plants, and fungi had diverged much earlier, at about 1.2 billion years ago, well into the Precambrian. One could look much farther back into time!

Ergaster

I think here they are talking about the specific evolutionary event of the origin of brood parasitism, and not the general "event" of branching. In the diagram, species A and B are sister species, so if they both share the trait, and C+D do not, then the trait arose once just before the node (or branch point). In these types of analyses, it is generally assumed that if all the members of a clade share the same trait, then that trait was present in the common ancestor of the clade (represented at the node). If species A and species C share the trait, but are members of different clades, then the trait arose twice--once on the branch leading to A and once on the branch leading to C (there are other alternatives to explain each of these patterns, but they are less parsimonious--i.e. require more evolutionary events to explain them). From what I gather, the researchers examined the distribution of a number of traits across both a molecular and a morphological phylogeny, and then looked for the ones that covaried.

That approach is called historical ecology, and it is pretty interesting. It can also be used to correlate the origin of traits to biogeography and ecological distributions.

The question of extinction is a good one. If it turned out that (using their diagram again) that there was an extinct sister species to A and B (say A') which was not a parasite, then at the very least you'd still have two evolutionary events: an origination before the node, and then a loss in the extinct species, OR independent originations in A and B, acquired after the node, which would mean that the common ancestor of A+A'+B was not a parasite (and you still have two events, so on the basis of that trait alone you would not be able to say which might have occurred). So yeah--the inclusion of extinct or fossil species in phylogenies has the potential to radically alter the explanations of the pattern of trait distribution.

Peez

I am a little rusty, but if you really want to get into it the standard reference is (IIRC) The Comparative Method in Evolutionary Biology by P. H. Harvey and M. D. Pagel (Oxford University Press, 1991). The basic concept is that looking for correlations among traits in different species runs into the problem that traits may not have evolved independently in related species. For example, if brood parasitism evolved in one particular species of cuckoo which happened to be migratory, then species that descend from that parasitic, migratory species will tend to be parasitic and migratory. The result could be a bunch of species that are both parasitic and migratory. This can create a correlation between the propensity to parasitism and the propensity for migration, even if there is no relationship between the two other than common descent.

To avoid this problem, one can use a phylogenetic tree (if available) and look at contrasts between taxa rather than traits of individual species. For example, take two closely-related species. One is migratory and parasitic while the other is neither. Here we can say that migration and parasitism are together not just because of common ancestry, because the closely- related species inherited neither. Of course, the two traits may have evolved together by chance, so we must look at many contrasts and conduct suitable statistics.

The most obvious contrasts are between extant species, but we can infer the traits of recent common ancestors (branch points) and calculate contrasts among these. When two species have a different trait, there has been one "evolutionary event" (the trait has changed from the common ancestor), though it is not always obvious which species (or whether both) has changed. Extinct species do not influence the analysis any more than any other datum that is not being used, as the critical data are contrasts in different traits that might be correlated.

I hope that I have got it right, it has been a while.

Peez

Dr.GH

Thanks for the comments. I have much more confidence in molecular phylogenies based on non-coding sequences (such as:

<a href="http://www.pnas.org/cgi/content/full/96/18/10254#Top" target="_blank">here</a>

If the trait tree then maps onto the molecular tree you have a stronger result.

Coragyps

And another neat review of stuff that this old non-biologist knew nothing of:

"Planetary Biology--Paleontological, Geological, and Molecular Histories of Life"

Steven A. Benner, M. Daniel Caraco, J. Michael Thomson, and Eric A. Gaucher
Science May 3 2002: 864-868.

The abstract:
It has some fairly astounding examples, like that DNA/enzyme differences in modern yeasts point to the origin of their ability to ferment sugar to alcohol only about 80 million years ago. They suggest that this is no coincidence, as sugar-bearing fruits arose or became widespread about this time, I suppose as the first significant source of sugar out in the open for exploitation. And fruit flies, too, arose around this time, and apparently feed on the yeast that grows on the sugar. (The bio grad students feed on the alcohol, I guess.)

Another example is reconstruction of ruminant ancestry from divergence of their digestive lysozymes: they use these to digest the cell walls of the bacteria that break down cellulose in their rumens. The molecular data points to the Oligocene, when the fossil record shows climate cooled off, hard-to-digest grasses displaced tropical plants, and the big non-ruminants died out. There have even been experiments to "recreate" the "ancestral" enzymes.
Neat stuff.

lpetrich

Actually, fruit-fly larvae (maggots, grubs) eat the fruit itself; the yeast are a bit small to sort out. But the emergence of fruit flies did happen at an appropriate time