More Recent Comments

Tuesday, February 25, 2014

My molecular evolution midterm test

My students wrote the midterm test today. Here are the questions. They had to answer the first question and 4 other questions (out of 5). How would you do?
  1. What’s the most important new thing about molecular evolution that you have learned in this course so far? Explain your answer by describing your "important new thing."
  2. If mutation rates are relatively constant then why does the molecular clock tick at different rates in different proteins?
  3. Many evolutionary biologists think that population genetics is the key concept in understanding evolution but biology students often complete several years of courses without ever learning about effective population sizes, mutation rates and the importance of random genetic drift. Why? Is it because population genetics is not a necessary key concept in evolution?
  4. Grad students at this university publish a journal called Hypothesis. A few years ago (2005) there was a student who wrote ...
    I am a grad student, and long hours at the bench have got me thinking of other things lately, including the idea of marriage. I came up with a few criteria to direct me on my quest for a wife, and near the top of the list was that she needs to know what a gene is. I thought that this would be a reasonable thing to ask for. I like learning about how we and the rest of life work, and knowing how, in a general sense, cells are programmed to do what they do is a pretty good indicator of similar interest. My friends, however, disagreed with me, and on several occasions, as I shared my list, I feared that things were going to get violent. They argued that I will never get married with such a short-sighted and elitist attitude.
    Imagine that you would only seek partners who knew what a gene was. What definition would you require and why?
  5. What’s the best evidence that a substantial amount of our genome is junk?
  6. Imagine that you are teaching a class and you ask students to calculate a mutation rate in humans based on what they know about biochemistry. What mistakes are they most likely to make and why?

37 comments :

Konrad said...

Some, er, *interesting* things in that test.

1. A question to which the correct answer depends primarily on what the student knew before starting the course. A student can answer "inside every cell there is a substance called DNA" and give you a description of its structure. This is certainly the most important thing to learn about molecular evolution if you didn't know it before (more important than anything you likely covered in the course), and the answer doesn't imply that the student failed to also learn the more advanced things you taught - would you deduct marks based on your perception of what the student was meant to know _beforehand_?

3. Really? A question about curriculum design and why some professors are bad at it?

4. "I'd accept any definition because I don't want to stay single" might well be the most correct answer for some students. Are you going to mark this wrong and, if so, why?

6. Really? You are asking which mistakes are common among students who don't understand the work? Did you teach this? (On the one hand I hope not, so as not to waste the time of the students who understood it the first time; on the other hand, how on earth are those students meant to know the answer to the question? I certainly don't.)

Is this course taught in the faculty of education or of science?

Larry Moran said...

Konrad demonstrates most of the things that are wrong with science education.

A student can answer "inside every cell there is a substance called DNA" and give you a description of its structure.

That's what I call a "kindergarten answer." Adults who are about to graduate with a biology/biochemistry B.Sc. in a few months don't get marks for answers like that.

Really? A question about curriculum design and why some professors are bad at it?

Yes, if that's how you want to answer the question.

"I'd accept any definition because I don't want to stay single" might well be the most correct answer for some students. Are you going to mark this wrong and, if so, why?

Yep. That would be wrong because this is a serious university course not a kindergarten course. It is not a childish game where students try to "beat" the professor by trickery.

... You are asking which mistakes are common among students who don't understand the work? Did you teach this? (On the one hand I hope not, so as not to waste the time of the students who understood it the first time ...

The number of students in my class who gave the correct answer at the beginning of the discussion is ..... zero. The number who gave the correct answer after one hour of debate is four out of twenty.

... how on earth are those students meant to know the answer to the question? I certainly don't.

I'm not surprised. Maybe you should take my course.

Anonymous said...

It seems to me that rather than have the students regurgitate facts you want the students to demonstrate what they've learned with questions that are a bit more open ended. So that facts are just the starting point for thinking thats a bit more creative and flexible. This is great but I have to say I sympathize with Konrad here. These questions ( except 2 and 5) seem rather subjective and arbitrary. To me a good quesion might be of the form: " a species of animal is discovered which has a haploid juvenile stage as part of its life cycle. How will this affect x y and z etc"
Come to think of it, you couse base an entire test on statements by Casey Luskin, Cornelius Hunter and J Wells and have the students explain why they're wrong

Jonathan Badger said...

Yep. That would be wrong because this is a serious university course not a kindergarten course. It is not a childish game where students try to "beat" the professor by trickery.

Then don't ask stupid questions on tests that can be beaten that way. Besides, even if you were trying to mock the student, this question could be interpreted as being sexist -- the stereotypical male scientist who despairs at finding a mere woman at his exhalted intellectual level.

Konrad said...

And of course your students know they won't get credit for answers that are honest and correct but don't match the "model" answer you have in mind. So you are forcing them to guess or memorize your model answers.

As for which misconceptions are common - the students know what their _own_ misconceptions were. They don't know (unless they've wasted their time learning things that are relavant only to pedagogy) what the most _common_ misconceptions were. And I bet even you don't know whether or to what extent the misconceptions you observe in your class are representative of those in other institutions - I wouldn't be surprised if there are systematic differences between the USA and Canada, for instance, and investigating this might make for a sensible study in Education. Not very relevant for Biology though.

Larry Moran said...

Then don't ask stupid questions on tests that can be beaten that way.

You really don't get it, do you?

Learning is a two-way process. Students have to take some responsibility for their own education and that means they have to give up on the idea that it's just a contest between students and the professor over how to ask precisely the right question that can't be challenged by a student lawyer looking for a way to get easy marks. I don't put up with that kind of childish behavior in my courses.

I do my best to treat students as responsible adults who are here to learn and I expect them to return the same level of respect.

this question could be interpreted as being sexist -

That's correct. There are a bunch of politically correct students and faculty at my university who are completely humorless. They delight in exposing examples of everthing they think is improper behavior. They are classic examples of the exact opposite of the kind of critical thinking I'm trying to teach.

Thanks for pointing that out.

Larry Moran said...

Come to think of it, you couse base an entire test on statements by Casey Luskin, Cornelius Hunter and J Wells and have the students explain why they're wrong.

That's the course I teach in the Fall semester.

Konrad said...

One way to treat students with respect is to set questions that test their understanding of the subject rather than their understanding of which answers are most likely to get credit from their professor. Once they have to start guessing what you expected them to know or not to know coming into the class, or which of the things they learned are considered most important by you, you're heading down the wrong path.

Larry Moran said...

The class was split up into six groups of four students each. At the beginning of the class they were asked to calculate the mutation rate in humans based on what they knew about biochemistry. After 10 minutes of discussion each group could either announce their final answer OR ask a question of fact (e.g. what is the error DNA replication). After five rounds all groups were prepared to submit an answer.

We then spent 15 minutes going over the answers and seeing why the groups gave different answers. In the rest of the class we discussed other ways of calculating mutation rates and why the answers differed by a factor of two or three.

The answer to the exam question will reveal whether the student understands the parameters that have to be taken into account when talking about mutation rates. It leads to an understanding of potential sources of error. The exercise is an experiment in critical thinking and I think it worked quit well. I'll probably use it again next year depending on how students answer the question.

Piotr Gąsiorowski said...

It is pointless... to continue this shit....

Pray, sir, feel free to discontinue your mental defecation.

Konrad said...

Quest: you are wrong - Larry generally makes a point of addressing the evidence (if you disagree, why do you bother reading his blog?) This is about knowledge and understanding, belief is irrelevant.

Larry: you've answered my question: you do teach which mistakes are common (among your own students, anyway). This is a sensible exercise not because you actually want the students to learn which mistakes are common, but because you want them to learn critical thinking. So (rather than asking a question that essentially tests whether they attended that session) I would test whether they have achieved the real goal by asking them a different question (one that wasn't discussed in class) that requires the same critical thinking skills.

Robert Byers said...

These questions get one toward a degree result? Marriage? Why students get things wrong?
Mutation rates is pure speculation unless its been witnessed(tested) over past amounts of time where mutations actually took place and perhpas produce results. otherwise its all poor sampling from presumptions.
The question to ask them if they are to strive for a higher intellectual investigation of molecular biology IS whether and what is the scientific molecular evidence that present DNA is a accurate model to make past molecular associations demonstrating common descent and this by evolution.
Now thats a question that would weed out non elect.

Jonathan Badger said...

Well, education is a two way street, but that includes the professor putting a bit of thought into the questions...

As for 4), I just don't see how the story serves any purpose other than confusing the issue and possibly offending some people. Asking students for their definition of a gene and their justification for it is all good and well -- why not leave it at that?

Konrad said...

I can just picture it:

"7. Whether and what is the scientific molecular evidence that present DNA is a accurate model to make past molecular associations demonstrating common descent and this by evolution."

Konrad said...

The story definitely changes the question (unless we imagine a student who would only select a partner who can repeat their preferred definition verbatim - and if I were marking it I would fail such students as a matter of principle :-) ).

Instead of giving a single definition, the students are being asked what would make a definition acceptable or unacceptable, i.e. to specify the entire set of potential definitions that, in their opinion, would convey a sufficient understanding of what a gene is. This presents some problems:

1) Their choice of "sufficient" can be completely arbitrary except that some choices won't pass muster, so they have to guess what would please the examiner.

2) It is not clear that such a question can be sensibly answered under realistic time and ink constraints.

3) It is even less clear that the question tests anything worth testing.

Anonymous said...

I read those questions and wondered if I'm glad I never took a class from Larry Moran, or sorry I didn't! Assuming that he provided enough background, as apparently he did for #6, those questions would be thought-provoking. They suggest the whole class would be an interesting challenge.

I'd hate to grade such questions, though. With open-ended questions like these it's surprisingly easy to tell who knows the material (and may make some errors) and who doesn't get it (though relevant correct vocabulary might be thrown at the answer). It's harder for me to decide how to grade the failures that have correct bits in them.

Larry Moran said...

I'm committed to trying this style of teaching since I strongly believe it's the right way to go. We have got to stop teaching the "memorize and regurgitate" style of course.

But you've put your finger on the tough part. It's hard to assign grades to the middle part of the class. Students also have to submit an essay on a topic of their choice. Fortunately, the essay grades and the test scores show a strong correlation so it looks like subjective grading is pretty fair.

Larry Moran said...

BTW, we did the same thing for the question "What is a gene?" After the discussion was over I gave out references to several papers and posts on the issue so the students could see that there were several different legitimate answers. I described my preferred definition and explained to the students that whenever I use the word "gene" that's what I mean.

I forgot to mention that two weeks before the test I handed out a list of 30 questions, including the six ones you see above, and I told the students that all the questions on the test would be taken from the list.

Pedro A B Pereira said...

Mutation rates is pure speculation unless its been witnessed(tested) over past amounts of time where mutations actually took place and perhpas produce results. otherwise its all poor sampling from presumptions.

Fortunately, we have been doing precisely that for quite a long time. For example, we can trace the geographical spread of a bacterial or viral disease by analizing mutations starting from an outbreak. It's routine work in molecular epidemiology. It's also easy to set up in a microbiology lab. An there are many indirect ways to get extimations too. Nailing down exact mutation rates, specially for complex eukaryotes is more difficult, but it's not like we're dealing with estimation errors of orders of magnitude.

Unknown said...

Question 2 is one we give in our first-year biology course at Western. Molecular evolution in first year that's how to teach evolution...no more finches beaks!

Uncivilized Elk said...

"And of course your students know they won't get credit for answers that are honest and correct"

I don't think the word honesty means what you think it means. If you know what your instructor is trying to get at (which one should if they attended classes), yet you purposefully answer the question by adhering strictly to the way it was written while ignoring all context and the contingency of the course you took, you're not being honest. You're being a dick who will probably fail their exam.

Uncivilized Elk said...

What exactly is bad about "subjective" questions? They usually prompt much more explanation and thought. I also see nothing arbitrary other than maybe including that quote from the journal in the gene question.

Uncivilized Elk said...

For fuck's sake, I have never taken Larry's course and I can tell what he is trying to get at. It's not fucking difficult; these people are supposed to be in college.

And while I think the following statement is complete and utter horseshit (yes, I'm sure people are super duper delighted to be calling out white guys perpetuating the same harmful tropes they have been for fucking ages instead of annoyed and exhausted):
"That's correct. There are a bunch of politically correct students and faculty at my university who are completely humorless. They delight in exposing examples of everthing they think is improper behavior. They are classic examples of the exact opposite of the kind of critical thinking I'm trying to teach."

if that statement actually WAS written in a journal, it's that group of douchebags being sexist in how they phrase "I want an intelligent partner", Larry isn't endorsing it or anything. Though like I said, I find that comment I quoted goddamn ridiculous. "Oh, haha, you feel oppressed and hurt by the constant blindness of the privileged in society. How uncritical of you!"

Uncivilized Elk said...

" Their choice of "sufficient" can be completely arbitrary except that some choices won't pass muster, so they have to guess what would please the examiner."

Jesus Christ, it's like you've never taken a written test before. I'm flabbergasted at how strung up over semantics you get. I agree that the question is needlessly wordy, but come on, most questions have an unwritten but implied component of "you must explain or describe ____ sufficiently".

It's like you get a written test on history and are told to give a comprehensive description of a certain event, and you whine to the teacher how "comprehensive" is arbitrary and would also take too much ink. For crying out loud, what's wrong with you?

John Harshman said...

I would have thought #6 was a strictly empirical question. How do you get a mutation rate from your knowledge of biochemistry? Also, I'm interested in the answer. It would be handy to know just what mistakes a student is likely to make.

John Harshman said...

You're saying that morphology is unimportant to evolution? Is it really necessary to leave out the finch beaks just because you also bring in molecular evolution?

Larry Moran said...

Read: Mutation to see all my post on mutation and Estimating the Human Mutation Rate: Biochemical Method to see the specific example of calculating mutation rate from a knowledge of biochemistry.

The most common errors are: (1) not taking into account the error rate of DNA repair, (2) not knowing the correct size of the haploid genome, (3) not knowing the number of cell divisions between zygote and formation of gametes, and (4) not considering the separate contributions of sperm and eggs.

TheBrummell said...

"How would you do?"
1. I haven't taken the class, and I indeed I've never taken a class titled "Molecular Evolution" or something similar. I have, however, taken classes titled "Molecular Biology" and "Evolutionary Biology", though that was a while ago. I remember being surprised by some facts that were presented in those classes, so I'll answer with one of those: The most important new thing I learned was the range of potential sources of error one has to take into account when building phylogenies using datasets of mixed molecular (e.g. gene sequence, protein sequence) and morphological data. Beyond the obvious quesions regarding how to compare "traits" as diverse as an A->G mutation in a third codon position and a doubling of some morphological ratio (e.g. beak length to beak thickness), there are layers of concerns and pitfalls around issues such as molecular clock calibration against fossils, population size changes through time and the effects on neutral evolution, and uncertainties about phylogenetic trees, and how those uncertainties affect overall estimates of things like time-since-divergence based on which nodes are well supported and which are not. And then there are the statistical problems, such as appropriate analytical methods for building phylogenies (e.g. neighbour joining alone seems to have very situation dependent pros and cons) and how to interpret the methods section of a paper that presents a new phylogeny of some group of organisms. Much of this discussion was new to me when I encountered it in Molecular Biology and in Evolutionary Biology, and I found it very interesting.

2. As a quick stab at this one: different proteins experience different amounts of purifying or stabilizing selection because different proportions of their total amino-acid sequence will be under relaxed or stringent selection. To clarify, imagine a short protein with an active site (we'll assume this protein is an enzyme) and some other sequence that's mostly just spacer between critically-important residues. Mutations in that sequence that affect the active site are strongly selected against because the function of that protein is changed, but most mutations in the spacer amino acids are neutral or near-neutral and can accumulate through time in different lineages carrying that protein. Another protein might be very long and have relatively few amino acid positions that are under strong selection, and great chunks of that sequence may be free to accumulate many neutral mutations.
In addition, amino acids are encoded by different numbers of codons - some by only one, others by as many as six different 3-base RNA codons. This means some amino acid substitutions will be caused by a single DNA mutation, while others will be silent as regards the amino acid composition (with the caveat that tRNA abundances and ratios are probably important for determining the maximum rate of protein production in at least some cells).
I'm going to leave alone the idea that some proteins (the whole molecule) are under more intense purifying selection than others and that many proteins may actually be evolving entirely by drift in many lineages.
Short version: proteins have many differences that mean a constant genome-wide mutation rate will lead to different rates of fixation of mutations within the sequences of individual genes, due to differences in selection intensity across the length of each protein's sequence.


...Google won't allow really long comments. Apologies for the length, continued in next comment...

TheBrummell said...

Part the second:
3. Part of this apparent mismatch in priorities - working biologists believe evolutionary biology is fundamental to all aspects of biology, yet core concepts in evolutionary biology are taught relatively late in many biology cirricula - may stem from historical inertia (course structures take a long time to change in many university departments) and may stem from a perceived heirarchy of concepts.
It may be that some concepts, such as the main tenets of cell biology (cells come from cells, living organisms are made of cells, all the cells in a multicellular individual have the same genome*) or of ecology (food chains, food webs, nutrient cycles, succession, responses to disturbance, keystone species) are easier to teach to individuals with little or no former trainining in biology. However, I am not aware of thorough tests of this idea that some important biological concepts are inherently more difficult for people to grasp, or that some "background" information is necessary before it becomes worthwhile to discuss ideas such as selection vs. drift. Many of the key concepts of population genetics are simply applied statistics, such that the only necessary prerequisite might be a course in probability and statistics. On the other hand, I personally found it an "aha!" moment when I finally realized *why* most mutations (in coding sequences) are negative - and the answer has to do with allele dominance, protein synthesis and function, and development in multicellular eukaryotes, concepts that may be divided among introductory biology courses in genetics, physiology, and developmental biology**.
I think population genetics is a necessary key concept (or set of concepts) in evolutionary biology, and its relegation to late (and often optional) courses in undergraduate biology training may reflect historical forces in academic departments, a path of building knowledge upon earlier foundations through a 3- or 4-year degree program, and on acceptance of a compromise position among professors that may wish to devote their teaching time to some courses or subdisciplines at the expense of others.
* with modifications in some specialized tissues, such as the highly duplicated chromosomes of Drosophilid salivary gland cells
** Come to think of it, I suspect developmental biology is almost always a 3rd or 4th year course in biology departments.

4. Again, because I have not taken Dr. Moran's Molecular Evolution course, I might be working with a different definition of the word "gene". Rather than tackle this question (we get a choice to omit one, right?), I"ll be facetious and say "get new friends, anybody who threatens violence during a discussion of one's preferences or strategies for finding long-term romantic love is clearly a deranged individual and best avoided."

TheBrummell said...

Part the third:
5. Hmmm... "best" implies a single, most-correct answer that was probably discussed at some length in class. I could talk about the abundance of inactive transposable elements in the human genome (without looking up the relevant posts here at Sandwalk, they're something like 40% of the genome? I think?). Or I suppose I could talk about Dr. Gregory's "Onion Test" that challenges any argument for majority function in the human genome with the question (paraphrasing, with apologies for errors and poor writing) "why do onions require so much DNA in their cells, if the proposed function for human non-coding DNA is true? Are onions that much more complex / protected by dilution from mutations / evolutionary potential mumblemumblemumble (I don't understand the "evolutionary potential" arguments)".
Where was I? Oh, right, a test. Um... I'll just go with: We have looked and looked, and there has never been any evidence of function of great swathes of the human genome, such that a fair description of at least 80% of the DNA in each human cell is "junk", as in, having no known function. Evidence of absence is available in this case, because many predictions about function for bulk DNA or for particular segments / sequences / motifs have been tested and failed.

6. Earlier comments describe a class exercise and discussion about these issues. Not having attended that discussion, I'll bow out of this one and accept a lower grade on this test.


How'd I do?

Jonathan Badger said...

We have got to stop teaching the "memorize and regurgitate" style of course.
Maybe. But the other extreme which is just as bad if not worse is the humanities-style free-style response where there are no objectively right or wrong answers -- just answers the professor/TA likes or not. One of the reasons I became I scientist is that I was deeply unimpressed with the rigor and lack of objectivity in the humanities, and I wouldn't want to see science education go that way.

Larry Moran said...

Jonathan,

What's the point of your comment? Have you ever seen any evidence that science education is going in the direction you suggest? Are you implying that my course is like that because that's what it seems like to me?

Jonathan Badger said...

Well, I don't know what you actually teach in the course, and it may be more rigorous than it looks from this test, but your choice of questions seems rather fuzzier and less mathematical than I would except in a molecular evolution course. Do you cover things like substitution models for maximum likelihood phylogeny later on? And you apparently have an *essay* for a molecular evolution course, which seems rather odd outside the humanities or maybe the "physics for poets" style science courses for non-majors.

Robert Byers said...

Fine. If its been witnessed its just witnessing a very common mutationism in something in biology. however evolutionism is about mutation monsters creating the glory of biological reality.
Its NOT been witnessed and is completely speculation.
Even if true it would still be lucky speculation.
you make the case that creationists demand about scientific evidence before claims of conclusions are said to be from science.
evolutionism has mutated logic to its aims .

Larry Moran said...

Now I understand where you are coming from. You expect me to teach a technology course and not fundamental principles and concepts of evolution at the molecular level. Undergraduate students do not need to memorize substitution models for maximum liklihood algoriithms. That's a good illustration of what's wrong with undergraduate education.

Essays give students an opportunity to demonstrate their critical thinking skills and show that they can learn on their own. They also allow me to access their ability to communicate and that's an important part of education. The students choose their own topics, but I have to approve.

Some of the topics are: junk DNA, the role of recombination in evolution, whether the rate of human evolution has changed recently, whether epigenetics is important in evolution, what molecular evolution tells us about the Cambrian explosion, and whether the molecular clock ticks at a constant rate in most lineages.

I don't think it's the least bit "odd" for undergraduates to be writing essays on such topics but I'm not surprised that you find this kind of teaching very strange. I think my course is much more "rigorous" than the kind of course you would teach. You seem to think that "rigorous" requires memorizing mathematical formulae whereas I think that "rigorous" means teaching students how to think for themselves.

Tom Mueller said...

@The Brummell

Impressive!

re: Question #5 - I suspect that Larry discussed at great length Ford Doolittle's c-value paradox blasting ENCODE ... as a matter of fact, I could not imagine Larry failing to blast ENCODE in class, making this a "gimme" question for his students.

To follow up on question #5, I am still scratching my head over this post
http://sandwalk.blogspot.ca/2014/01/can-some-genomes-evolve-more-slowly.html

How is it that molecular clocks across lineages (even considering identical proteins) click at slower rates depending on which lineage is examined?

I can think of several possible answers, but hesitate to embarass myself in public.

Tom Mueller said...

@ Larry

I would love to have been one of your students!

Clearly you are moving up Bloom's Taxonomy and see yourself less as "sage-on-the-stage" but more as "guide-on-the-side"

What is also clear, is that it is very difficult to evaluate your evaluation without knowing what happened in your class not to mention examining your marking rubric.

I would have phrased some of your questions differently - but that is a matter of style. As students accustom to your style and understand where you are coming from, the thrust of these questions would become more obvious to them than to those who do not know you so well.

As the mists of time erase the memories of university experiences, your students will surely remember you and your classes.