More Recent Comments

Friday, May 03, 2013

The Molecular Evolution Exam

The students in my Molecular Evolution course have written the final exam. I finished grading the exams yesterday and the final course marks have been submitted. Now I don't have to think about teaching until August.

Just for fun, I'm posting the final exam questions. How do you think you'd do? I gave the students a list of all possible questions on the last day of class. The only one they hadn't seen was question #2. They knew that the first question would definitely be on the exam.

Students had to answer questions 1 & 2 and choose any three of the other questions.
  1. Choose a subtopic from your essay and explain it better than you did in your essay and/or rebut the comments and criticisms made by the marker/grader.
  2. Imagine that identical female twins were born to a woman in 1000 AD. Imagine that you could find a direct descendant of each twin in 2013. If you sequence the complete genomes of the descendants, approximately how many differences would you expect to find? How do these compare to the differences between any two randomly selected individuals from the same part of the world? Explain your reasoning and describe any assumptions you make. Think carefully before you answer. The second question is the most important one. (Human mutation rate = 130 mutations per generation. Haploid genome size = 3.2 × 109 bp.)
  3. There are hardly any pseudogenes in bacterial genomes. Why haven’t pseudogenes been eliminated from our genome?
  4. Explain the two-fold cost of sex. Why is this a problem in evolutionary theory?
  5. Richard Lenski’s group at Michigan State University has been following the evolution of 12 cultures of E. coli for over 50,000 generations. All 12 cultures are grown under exactly the same experimental conditions and the mutation rate is high enough that every culture has been exposed to multiple mutations at every base pair in the genome. Explain why only one of the cultures has evolved the ability to use citrate as a carbon source.
  6. Why is the Three Domain Hypothesis being challenged by some molecular evolutionary biologists? What are the alternatives?
  7. What do you think of Kirschner and Gerhart’s “Theory of Facilitated Variation.” Is this something that has to be incorporated into a new extended version of evolutionary theory? What, if any, are the limitations of the theory and what, if any, new insights into evolution does it provide?


20 comments :

Rosie Redfield said...

Q2. Humans reproduce sexually so after about 40 generations (assuming reproduction at age 25 on average) only about 1/(2^40) of the DNA would come from the original twin, if the mating partners were always unrelated. And this tiny fraction would have accumulated some mutations. But the population of mating partners is limited, so the 40-generation descendants of the twins would be no more similar than any two randomly chosen individuals.

Larry Moran said...

I was surprised that only one-third of my students got the right answer. I think the distractors were too potent.

Larry Moran said...

BTW, it was fun watching you on national television the other night even though I disagreed with most of what they said about MOOCs.

John Harshman said...

I'm not familiar with the term "two-fold cost of sex". What does it mean? I can think of at least three potential costs of sex, so are you leaving one or more out or does it mean something else entirely? (1. loss of half the reproductive capacity; 2. removal of descendants from any adaptive peak you may be on; 3. any cost associated with locating, attracting, and dealing with a mate.)

Robert Byers said...

#4
The question only is a issue if one assumes that sex selection etc had time to affect biology.
This is only true if there was time enough. Time is a issue of geology and not biology in these issues and so the right answer would be that since sex selection issues are only speculative and this is a coarse in science then no answer is the right answer.
What would be my mark?
by the way these questions being the origin for degrees is says a lot about these university getting degrees that are to get them ahead of others.
The workplace demands a sharper intellect then mere memorizing of these points.
its just my observation that university kids these days seem not as sharp as kids who go to lesser education or none.
Perhaps too many kids get into university. its not the highe elite of the old days anymore.

AllanMiller said...

Really? You gave me a debate over it at talk.origins back in the day! :)

Personally, I think it is a widespread mistake [awaits the booming voice of Biology offering a corrective for my presumption!]. The twofold cost relates to the fact that asexual females can produce offspring (rather: grandchildren) at twice the rate of their sexual cousins - the 'cost of males'. There is also a 'gene-centric' version: any gene gets into only half of gametes, but can get into all offspring in an asexual - the 'cost of meiosis'.

This leads to the assumption that sex must have a twofold benefit, in order to offset this assumed cost. It's an illusion.

1) The organism 'cost' is a cost of gender, not of sex, and gender (differential gamete size) is a feature of multicellular organisms (it is distinct from mating types in single-celled organisms). There may remain a mystery as to why gendered sexual multicellular forms have not been wiped out by asexual ones, but a number of factors - differential evolutionary lifetimes, rates of cladogenesis, the ecological consequences of limited variation - will impinge on the relative distribution of asexual forms about the eukaryote tree. Asexual eukaryotes always arise in a milieu where the sexual form has tenure.

2) the assumed gene cost misses something vital about an 'asexual gene' - any that can eliminate the gamete cost immediately stop being 'genes' (evolutionary alleles, not protein-coding regions), and the whole genome becomes the evolutionary allele. No stretch of DNA can increase its representation in 'the population', as by becoming asexual, it leaves it, forming a set of individuals almost indistinguishable from 'the population'. Asexuals can be treated in population genetic analyses too, but applying those analyses across the sexual/asexual divide can be misleading. Most genes can't do anything about it anyway: to a gene, meiosis is simply resumption of the ancestral state (haploidy), not a 'halving'. There is some bickering, but mostly it's orderly.

John Harshman said...

Ah. I apparently misinterpreted "twofold". I thought it meant "two separate costs", while you tell me it means "a factor of two".

There are single-celled organisms that still produce gametes of different sizes, you know, often small, mobile ones vs. big, nutrient-packed, stationary ones.

Larry Moran said...

The two-fold cost of sex applies to organisms that can only reproduce either sexually or asexually. It mostly applies to animals. That's because the sort of biologists who used to think about these problems apparently didn't know about any other kinds of species. :-)

Imagine a female lizard or snail that can reproduce by parthenogenesis. She will pass on 100% of her genome to her daughter. Now imagine a female lizard or snail that can only reproduce sexually. She will only pass on 50% of her genome to her daughter. She has to produce two offspring in order to contribute the same amount of DNA to the next generation.

Joe Felsenstein is an expert. He can correct any misconceptions.

The bottom line is that it is not at all clear what advantage (if any) sexual reproduction confers that overcomes this disadvantage. One could assume that sexual reproduction is not advantageous but in this case that doesn't seem likely given the widespread occurrence of sex.

Larry Moran said...

What would be my mark?

Zero. What did you expect?

There are many ways to answer the question but if you start out assuming that the Earth is only 6000 years old and evolution didn't happen then you are lying about facts and that's not allowed.

The workplace demands a sharper intellect then mere memorizing of these points.

The fact that you think these questions involve memorization says a lot about why you don't understand science and don't like education.

John Harshman said...

That was my option 1. But it applies to just about any organism, not just to animals. Taking a gene-centric view, why should any allele consent to a lottery in which it has a 50% chance of not being picked? Just about any form of sex can be characterized in this way, with the exception of the bacterial sex I'm familiar with, in which the donor gives away a copy but the donee loses none of its genome.

Faizal Ali said...

I have to say, I'm very impressed by the questions themselves. In my undergrad science courses, evaluation was almost entirely thru multiple choice exams, where it was possible to attain a very good grade without actually understanding the subject all that well. Perhaps you have to show you're interested in the subject beyond its merely being a premed requirement before they let you get the good stuff?

Larry Moran said...

It does apply to all organisms but it gets complicated for single-cell organisms since they routinely reproduce by mitosis and only engage in sexual reproduction on rare occasions.

Mikkel Rumraket Rasmussen said...

Taking a shot at #5, if I remember correctly, the cit+ mutant actually required first a neutral potentiating mutation to get fixed by drift. Only when drift had fixed this neutral mutation, could the cit+ mutant metabolise citrate. Further, drift also had the effect that despite the cit+ mutant arising multiple times once the neutral mutation had fixed, only in 1 our of ever 200.000 cases did natural selection manage to fix it the actual cit+ mutation.

So, it's largely due to the stochastic nature of drift, which both managed to fix the neutral mutation requied, but also worked against the fixation of the beneficial mutation for a time.

AllanMiller said...

John Harshman Taking a gene-centric view, why should any allele consent to a lottery in which it has a 50% chance of not being picked?

I tried to address this in my point 2) above. Genes have very little mechanistic opportunity. They mostly lie inert on chromosomes during meiosis. The expression level is not visible. There are mechanisms to disable or disadvantage your allele, for so long as you inhabit the same cell, but not many, and they typically start to meet copies of themselves in pretty short order, dissipating the drive. Turning off meiosis altogether is simply not an option for any 'selfish gene' - gene-centric selfishness relies on recombination. Those genes that succeed in stopping meiotic reduction are no longer genes in that evolutionary sense. That's the true paradox.

There is really no need for 'consent', even borrowing Dawkins's figurative language. Genes are blissfully unaware that anything untoward is going on. Haploids come together, maybe do a few mitoses together, then part. Haploidy is probably simply the ancestral state, and there is no reason to preserve the diploid beyond its period of usefulness (which may yet cross numerous iterations of mitosis - see: aphids, multicellular germlines, and other organisms that return to the haploid only rarely).

AllanMiller said...

LM One could assume that sexual reproduction is not advantageous but in this case that doesn't seem likely given the widespread occurrence of sex.

Yes, a differential must be responsible for the patterning of the distribution, but that doesn't make it necessary to assume that the 'inherent' differential is twofold. The patterning applies across isogamous and anisogamous branches; the twofold cost of males is restricted to anisogamous groups only.

Homologues of genes involved in meiosis are found throughout the eukaryote tree, both sexual and asexual. The entire clade is presumably descended from a sexual ancestor, and all asexuality is secondary reversion. Among asexuals, the retention of identifiable homologues must indicate (if they are not cryptically sexual) that they are of recent evolutionary origin. The patterning may actually indicate not that sex is necessarily strongly advantageous in direct competition, but that in a world already well-stocked by sexuals, asexuals tend not to persist indefinitely even if they do manage to extinguish their parent population (difficult against a standing population of variant residents). If they manage to win that competition, they have all the variation and evolutionary response of their predators, prey or other competitors to contend with, and they tend to be sexual, for the same reasons - differential persistence.

Anonymous said...

Taking a gene-centric view, why should any allele consent to a lottery in which it has a 50% chance of not being picked?

Maybe you are taking the wrong gene's view. A gene promoting recombination, which provides the advantage of mixing and matching, would be able to leave many more copies of itself each time it recombined with advantageous other-gene combinations. Organisms without this gene might be less "evolvable," therefore would leave less successful descendants. At least recombination in experimental evolution speeds up the finding of successful variants several fold over experiments without recombination ...

Larry Moran said...

The potentiating mutation isn't known but it may be a mutation in arcB. That mutation may have been slightly detrimental making it even more unlikely.

The actualizing mutation required a complex rearrangement of the citT locus and this event is much rarer than a single point mutation.

AllanMiller said...

NE Maybe you are taking the wrong gene's view. A gene promoting recombination ...

Recombination is actually a separate issue from segregation. The '50%' issue relates to the fact that homologous chromosomes are hauled into separate gametes, only one of which can provide a replacement for the parent in steady state. The fact that reciprocal recombination occurs before final partition does not change the odds for any gene, including a recombiner.

The issue is whether a 'permanent segregation blocker' gene could spread around a population. It couldn't - it could create a near-identical clonal population, but it could not gain access to a locus within the sexual population and spread as an allele. People often treat the issue as if it could.

Gordon Anderson said...

1000AD! tut tut. Surely 1000CE for a good atheist

Unknown said...

The three domain hypothesis has been challenged by some scientists as they are curmudgeons. The alternative is to eat one's hat if proved wrong, if I recall.