« September 2009 | Main | November 2009 »

October 23, 2009

Experimental evolution of sex (revised)

"I show that a similar cost of sex exists when asexual mutants arise... but not when the species is a self-fertile hermaphrodite.... Although individual fitness (expected reproductive success) is assumed to be equal for sexual and asexual females, the heritability of fitness is... twice as high in asexual females" -- Richard Michod, Darwinian Dynamics

I should be working on my book, but a paper that just came out in Nature got me thinking about sex. A population with half males and half females will grow only half as fast as one consisting only of females that self-fertilize or clone themselves. So, many people have asked why sex evolved.

That's an interesting question, but I'm not sure about the rationale. As noted by Michod, a population of self-fertilizing hermaphrodites doesn't have any intrinsic growth advantage over a population of hermaphrodites that mostly cross-fertilizes. So is the problem sex, or males?

Evolutionary changes in gene frequency over generations depend on whether individuals with a given gene survive and reproduce more than other members of their population, not on the consequences for overall population growth. (Individuals can move between populations.) So we really have two related questions:
1) why do genes for producing male offspring persist? and
2) why do genes for cross-fertilization persist in species that can self-fertilize?

From an individual perspective, it's not apparent that producing male offspring is always a bad idea. Do couples with two sons have fewer descendants than those with two daughters? It can depend on the sex ratio in the population. If a human couple produces one offspring of whichever sex is in the minority, their offspring may have an easier time finding a mate.

But what about cross-fertilization? If a female cloned herself, her offspring would have all of her genes, rather than just half of them. So the frequency of genes for self-fertilization would tend to increase, unless individuals resulting from cross-fertilization were more likely to survive and reproduce. An offspring with half as many of one's genes, but a 2.1-fold better chance of survival (maybe because a sexual partner contributes different disease-resistance genes) gives a greater increase in fitness. So, one key to understanding the evolution of sex (cross-fertilization) is to measure the survival of individuals with one parent versus two, under conditions that plausibly occurred at critical points in a species ancestry.

This week's paper, "Mutation load and rapid adaptation favour outcrossing over self-fertilization", set out to "recapitulate the evolutionary process under the specific conditions predicted to favour either selfing or outcrossing." Levi Morran, Michelle Parmenter, and Patrick Phillips used the nematode, C. elegans, which consists of males and hermaphrodites. (This mix, and the lack of pure females, suggests there can be individual benefits to maleness, whatever the consequences for the population as a whole.) They used genetic manipulation to make populations that only self-fertilized or never self-fertilized, exposed them to high mutation rates or to a bacterial pathogen, and let them evolve.

Their hypothesis was that cross-fertilization would limit the accumulation of bad mutations: if two individuals with two different bad mutations mate, some of their offspring will get both mutations and die, but some will get neither, whereas self-fertilizing populations may not have any mutation free individuals. Sure enough, mutations accumulated in the self-fertilizing population, resulting in decreased fitness. Cross-fertilization was also beneficial to the populations exposed to the pathogen: the population made to cross-fertilize evolved resistance, perhaps because they could combine good mutations from different parents.

As discussed above, however, just because some trait benefits the population as a whole doesn't guarantee that it will evolve. So a more interesting results used populations where the amount of cross-fertilization was allowed to evolve. An increase in mutation rate caused two different strains to evolve more cross-fertilization. Pathogen exposure seemed to have a similar effect, although there was a lot of variability.

Previously, it was thought that even the low rate of natural cross-fertilization in this species was enough to provide most of the benefits, but they saw improvements with additional cross-fertilization. This was achieved by increasing the percentage of males, which they suggested would provide additional benefits via sexual selection. If males fight over females and the healthiest males win, or if females choose the healthiest males, maybe we aren't so useless to our populations after all.


October 16, 2009

Evolving resistance to cheaters

This week's paper, "Cheater-resistance is not futile" was published in Nature. It describes experimental evolution of the social amoeba Dictyostelium, whose propensity to cheat other members of its species was discussed by Will Ratcliff in a recent guest post titled "Sneaky slime molds."

When two Dictyostelium strains are mixed in a reproductive structure, cheaters contribute fewer cells to the stalk that holds up the reproductive spores. Could the presence of such cheaters select for cheater-resistance genes, just as the presence of owls or hawks selects for mouse genes that make their coats match the soil color?

To find out, the researchers started with a genetically diverse population (the raw material for natural selection) and added a cheater. By definition, the cheater would tend to increase in frequency over cycles of reproduction, but they prevented that by using a cheater they could kill after it had had its effect on the relative reproduction of the other strains.

They did this several times, and each time one or two cheater-resistant genotypes took over the selected population. For example, starting with a population that made less than 40% of the spores when mixed 50:50 with the cheater, they evolved a population that made about 50%. (The cheater-resistant strain didn't push its advantage, apparently!)

In at least some cases, they identified the specific mutation that let their mutant hold its own against the cheater. The gene "has no annotated homologues in other organisms", so it's probably not a universal anticheater gene. In fact, it didn't even work against all Dictyostelium cheaters. Still, it would be interesting to know how it works at the molecular level.

Meanwhile, cooperators take heart. It's possible to keep cheaters under control, without becoming a cheater yourself.

October 12, 2009

Darwin at the Smithsonian

I recently had two or three hours to spend at the Smithsonian, en route to the airport. I hadn't been to the natural history museum for awhile, and was interested to see how they were celebrating Darwin's anniversaries this year. Pretty well, it turns out. Banners outside advertised a Darwin exhibit and "Plants and butterflies: partners in evolution." Inside, there was apparently an organized "Evolution Trail", which I didn't have time to follow.

The Darwin exhibit is off the entrance hall with the elephant and has a mix of biographical and scientific exhibits. My main criticism was their definition of "co-evolution" as being limited to evolution for mutual benefit. Evolutionary arms races (e.g., between hosts and parasites) are also coevolution. The entrance hall on the other side, where I came in, has two display cases of Darwiniana.

The butterfly exhibit was dominated by a live butterfly room inside a larger room with displays on the coevolution of plants and butterflies, with fossils labeled "examine the evidence." I was happy to pay $6 admission to the butterfly room since I wanted to make a donation anyway and enjoyed having a frittilary land on my nose.

Near the Oceans exhibit was a display of Burgess Shale fossils I hadn't seen before, including Pikaia, a tiny 500-million-year-old chordate. We chordates have evolved a lot since then. Nearby were some fossil stomatolites.

The mammal room was great, focusing on adaptations in everything from bats to giraffes (splaying front legs to drink, with an explanation of adaptations to limit blood flow to head) to pangolins with termite mounds. Right in the middle of the floor was a window down to fossil hominid footprints.

I wish I could have stayed longer. One problem with a quick visit to the Smithsonian is that post9/11 hysteria has closed most of the bag-check rooms. You can't bring your luggage into the museum and if you leave it somewhere, they'll try to detonate it. (Luggage made of sapient pearwood can defend itself, but I wouldn't recommend bringing it to Washington!) But here's a secret tip for my regular readers only: the 4th St. entrance to the National Gallery still has a check room, complete with x-ray machine. Don't tell too many people, or they'll probably close it.

Coming up in March: the Hall of Human Ancestors!

October 2, 2009

Local TV new blows Ardipithecus story

If you don't believe in evolution, you might not want to listen to this next story. Scientists reported this week on a new fossil, possibly a human ancestor, older than Lucy. The good news is, we're not descended from chimps after all. The bad news is, chimps and humans are descended from the same ancestor.

That's a paraphrase of how our local TV news covered Ardipithecus ramidus, the fossil hominid discussed in a series of papers in this week's issue of Science. Read all about it on Carl Zimmer's blog. The TV anchor didn't say which scientists claimed we are descended from chimps, perhaps because no scientist has made that claim. Chimps have evolved over the six million years or so since our last common ancestor, including their split with bonobos. Can we expect the story below on the TV soon?

Startling breakthrough in human genetics! You aren't descended from brother after all, or even from your cousin. You and your brother still have the same parents, and you and your cousin have the same grandparents, though. I hope that doesn't upset you too much.

We don't know for sure that present-day humans are descended from Ardipithecus . It's a reasonable hypothesis, but any hypothesis is, by definition, subject to possible disproof. For example, if we found another fossil that was clearly much more similar to modern humans, dating from the same time or earlier, then we'd conclude that Ardipithecus probably has no surviving descendants.

But this species probably isn't too far from the direct line of descent between our common ancestor with chimps and modern humans. Suppose you wanted to know what your great grandmother looked like, but there was no surviving picture of her. If you had pictures of her sister or her daughter, that would give you some idea, even if all her living descendants are descended from a son.

Of course, much of what we know about our ancestors now comes from analyzing DNA in humans and closely related species. We can figure out when vitamin C synthesis was lost or adult lactose tolerance gained, for example. But we don't yet understand development enough to predict height, foot shape, etc., from inferred DNA sequences of ancestral species. So keep those fossils coming!