Header image alt text

Evolution, History and Society

Perspectives from a Third Career

Introduction – December 2017

Posted by admin on November 15, 2017
Posted in Uncategorized  | No Comments yet, please leave one

Having let this site sit moribund for a couple of years, I’m hoping to be a little more active, now that I am entering my fourth career, that of professor emeritus. Posts, when they occur, will probably a bit more general in nature, but I will attempt to stick to the themes of genetics, evolution and history. There are actually 40 days left until that career commences, so don’t expect much until after Christmas, when I will be spending more time working on this, as well as my more technical site Population Genetics: Applying Theory to Data.

Crayfish!!

Posted by admin on February 11, 2018
Posted in Uncategorized  | No Comments yet, please leave one

So one of the big stories in evolutionary biology last week was this paper, which described the genome of the “marbled crayfish” (shown below), and which supported three important hypotheses:

  1. They are triploid organisms, most likely originating from the fertilization of a rare diploid egg of the slough crayfish (Procambarus fallax), a species found naturally in Georgia and Florida but also propagated in aquaria.
  2. They reproduce parthenogenetically, and thus their population is clonal.
  3. They have rapidly in Northern Europe and are now doing so in Madagascar.

From *Nature*

There are other questions that come up, such as whether the original event occurred in an aquarium situation and how parthenogenicity developed (A consequence of triploidy? Some other fortuitous mutation?), but what has fascinated me has been the science news coverage of it. The original paper (well written and well worth reading) was published in Nature Ecology and Evolution, a new open access journal that is part of the Nature empire, so perhaps first we should look at what they had to say about it:

Geneticists unravel secrets of super-invasive crayfish

Notice that the title is pretty neutral – while “unraveling secrets” and “super-invasive” may be a bit dramatic, the story is told in a fairly straightforward manner. Then there is Science:

An aquarium accident may have given this crayfish the DNA to take over the world

Now we are getting a bit more dramatic. The headline focuses on what is, at best, a speculative conclusion. Now let's turn to a national newspaper in the United States that overall has sound science coverage, The New York Times

This Mutant Crayfish Clones Itself, and It’s Taking Over Europe

Now this is actually a good article, written by Carl Zimmer, who is an excellent writer, especially about evolution (I taught from the textbook he coauthored for many years). But “clones itself”? “Taking over Europe?” Please. The article is overall sound, but the title is (to put in mildly) over the top.

Finally, all of the above (including the original paper) treat the marbled crayfish as a new species (Procambarus virginalis – don't you love the species name?). So it is worth it to read Jerry Coyne's take on this:

A “parthenogenetic” crayfish reproduces without sex: is it a new species?

Jerry is, of course, the coauthor of the best book on the subject of speciation, and he and his students have made tremendous contributions to the field over the years. In addition, he is a believer in the importance of skepticism in science. So it is not surprising that he has doubts about species status for this creature. I don't necessarily agree with him on this – his argument is that since they reproduce clonally, the Biological Species Concept can't be applied. So it is, of course, for most of the microbial world. But that gets to the heart of the question of what constitutes a “species”, and it's too late at night to pursue that in depth.

In any event, this has been a fun story to follow, and I'm glad to see lots of good science writing about it. The headlines may be hyperbolic, but the coverage has been sound.

Explanation of the Previous Post

Posted by admin on February 7, 2018
Posted in Uncategorized  | No Comments yet, please leave one

OK, so usually I use this site to post nontechnical (and nonpolitical stuff), but since my professional time has for many years been spent living and breathing R and RStudio, I wanted to explore the ability of directly uploading blog posts to a wordpress site. Since this is, at least in part, a “sandbox” site, I figured that it would be a good place to try directly compiling (or knitting) and posting an “Rmarkdown” document to WordPress. For those who wish to emulate it, see This blog post for brief instructions. One confusing point (at least for me) – when setting the options for your site

  1. where it says c(user=“password:), 'user' is your username (not in quotes), and your password should be in quotes.
  2. For the url, adding /xmlrpc.php to the end of your standard one is essential.

Otherwise, it should work. This post was generated by that means, as very likely will be posts in the future.

While I am not a devotee of all aspects of the so-called “tidyverse”, I am taking to ggplot, one reason (aside from the attractiveness of the outcome) is the ability to add layers to graphics. That permits one from starting with the simple and moving to the complex. Here's an example. First we will generate some randomly distributed data and plot its density

dat <-data.frame(Data=rnorm(1000,0,1)) # Create some normally distributed data
pl <-ggplot(dat,aes(Data)) +
  geom_density() 

pl

plot of chunk unnamed-chunk-2

Now, suppose we want to superimpose a histogram of the actual points( and will make them red and somewhat transparent). We can add it to the plot rather simply:

pl2 <-pl +
  geom_histogram(aes(y=..density..),fill="red", alpha=.5,bins=40)
pl2

plot of chunk unnamed-chunk-3

And we could do more. But it turns out that this is not the only package to use such an approach – the coala package (available on CRAN) one to build coalescent models for simulation via ms (and a couple of other simulators) in a similar fashion. A coala model has 3 elements:

  1. The basic model, consisting of the sample size and number of genes to simulate.
  2. features, such as scaled mutation rate, migration rate, divergence model, etc.
  3. summary statistics – things like nucleotide diversity, segregating sites (returned as a matrix), Tajima's D, and trees in Newick format. So let's do a simple simulation of a 10 copies of single gene with the default length of 1000 and a scaled mutation rate (θ) of 4, and ask for a tree to be generated.
activate_ms(priority=400)
mod1 <-coal_model(sample_size = 10,loci_number = 1, loci_length = 1000) +
  feat_mutation(rate=4) +
  sumstat_trees()
sim1 <-simulate (mod1,seed = 5)

Now we can use ggtree to make a plot of it (and layers can be added as above for ggplot). First, however, we have to convert the Newick tree to a phylo object using the read.tree function from ape

tr <-read.tree(text=sim1$trees[[1]])
ggtree(tr) +
  geom_tiplab()

plot of chunk unnamed-chunk-5

Now suppose we want to learn more about the model – perhaps get nucleotide diversity and Tajima's D. We can add those summary statistics to the model and rerun it:

mod2 <-mod1 +
  sumstat_nucleotide_div() +
  sumstat_tajimas_d() + 
  sumstat_sfs()
sim2 <-simulate(mod2,seed = 5)

And we can see the output as follows:

paste0("Nucleotide Diversity = ", sim2$pi)
## [1] "Nucleotide Diversity = 6.84444444444444"
paste0("Tajima's D = ", sim2$tajimas_d)
## [1] "Tajima's D = 0.089447094573295"
ggplot(data.frame(n=1:length(sim2$sfs),sfs=sim2$sfs),aes(x=n, y=sfs)) +
  geom_col(fill="red", alpha=.5) + 
  scale_x_continuous(breaks=1:10)

plot of chunk unnamed-chunk-7

  labs(x="Number of Derived Alleles",y="Count")
## $x
## [1] "Number of Derived Alleles"
## 
## $y
## [1] "Count"
## 
## attr(,"class")
## [1] "labels"

And there is a lot more, but this is just a taste.

There are a lot of possible answers to this question, and to most people they (rather rightly) start as cost. For most of my career, the real cost of higher eduction to students and their families has outstripped the rate of inflation by a substantial margin. We can argue about the reasons for that – declining public support and the costs of athletics and other noneducational amenities come to mind – but the ultimate result has been the rise of two trends. First, careerist thinking. It is not surprising that, given the financial pressures they face, students and their parents want tangible outcomes. And that leads to an unfortunate response on the part of colleges and universities – the idea of students as customers or consumers. But there is something else that has become the norm that, in my view, is at once both a consequence of, and more insidious than, these trends. That is the unsigned student evaluation of teaching.

Wait a minute, you say. Shouldn’t students have the right to critique their professors? After all, it is absolutely true that there are bad ones out there, and that failure to learn can often be the result of failure to teach. So no doubt – student input has merit in evaluation of teaching. But let me scream the critical word – UNSIGNED. No university to my knowledge requires that student evaluations of teaching be signed. At the same time, of course, no graduate or professional school will accept a faculty evaluation of a student that is not signed. So, when I write a letter that can affect the professional future of a student, I have to put my personal credibility on the line (which I am more than willing to do so). Conversely, students can, to be blunt, be no more thoughtful than internet trolls, and in so doing have major effects, often negative, on the professional futures of faculty. There’s a disconnect here – why has it gone largely uncommented?

I don’t have all the answers, but I do have a couple of suggestions. First, it is absolutely true that, prior to the mid-60’s, students had no voice in the functioning of universities and colleges, and when the absurdity of that situation became evident, giving opportunity for students to evaluate their professors seemed like at least a partial fix. But why not signed evaluations? Here the argument is usually that, given the power relationship between faculty and students, anonymity protects the latter from retaliation. And then second, as costs increased and pressures grew to justify them, administrators found evaluations to be a convenient way of (insincerely I think) showing their concern for student input.

Regardless of all of the intricacies of how the process has developed, there is absolutely no evidence of which I’m aware that says that anonymous student evaluations, as collected and aggregated, have done one whit to improve teaching or learning. Rather, they have led to the professor as salesman (especially at lower tier institutions), putting on performances that make students comfortable and assigning grades that generate satisfaction (I’d love to have the data, but my strong prediction is that there is a high correlation between the level of a student evaluation in a course and her or his expectation with respect to final grade.

So what do we do? Student evaluations of teaching aren’t going away. So in a Swiftian sense, I will make some modest proposals:

  1. Student evaluations of teaching instruments must be designed carefully. I’m not an expert in survey design, but I would suggest that Likert scale responses to “The professor knows his field” are NOT appropriate. It is the role of the faculty to judge expertise, not the student.

  2. Evaluation must be mandatory. All students must submit a complete evaluation if they expect to get a grade in the class. That grade will be the one assigned by the professor in the absence of knowledge about the nature of evaluations.

  3. Here’s the key one ( I bet you can’t guess) – evaluations must be signed. The faculty member should not see the signature, however the department chair and dean should. That way, the student will know that she or he must be accountable for what she writes, but that her identity will not be known to the subject of her critique. In other words, a system parallel to that used by professional and graduate schools in evaluating candiates.

SO – carefully designed instruments and mandatory signed response. Students are moving into a world where personal responsibility matters more than ever – I’m putting these suggestions forward as ways that we can help them in that process and at the same time get more informed and objective student evaluation of teaching that can guide us in our quest to improve teaching and learning.

Migrating away from Dropbox

Posted by admin on November 17, 2017
Posted in Uncategorized  | No Comments yet, please leave one

Step one of this process is complete. I’ve managed to move my mapped photos to github. I have a love-hate relationship with that platform, but at least personal sites at github.io are free. The result can be seen here

Cross post

Posted by admin on July 31, 2016
Posted in Uncategorized  | No Comments yet, please leave one

“Dr. Science Teacher” nails it regarding belief vs. acceptance.

The best Twitter thread ever

Posted by admin on June 12, 2015
Posted in Uncategorized  | No Comments yet, please leave one


FRED and R

Posted by admin on July 18, 2012
Posted in Uncategorized  | No Comments yet, please leave one

Also nothing to do with Evolution, but one of my projects this summer has been to learn the R programming language. This morning I learned how to access data from the Federal Reserve (via FRED) and play with it in R. The result is a somewhat political take on recent economic numbers – I tried to publish an HTML file I generated here, but it didn’t work. However, if you want to see it, it is available here

Ok, so I’m following Evolution 2012 vicariously via Twitter, and clearly the big event of the night has been Rosemary Redfield’s talk, which happened (by coincidence?) to coincide with the release of the paper from her lab (as well as a companion paper from another lab) in Science refuting the “arsenic life” story, which got such a big media splash a year or so ago. When Carl Zimmer posted that the embargo on the paper had been lifted, I immediately went to Science, to find (of course) that the paper, clearly one of interest to the general public, seemed to be behind a paywall. I tweeted my disappointment, and within seconds, Carl responded that it’s available at Arxiv, and much to my delight, it was.

I may be wrong (and I hope not), but I think this is one more nail in the coffin of the Nature/Science/Elsevier model for scientific publishing. Look at it rationally. Here was a study, funded by NASA (i. e. by the public), proposing what is by any standard a revolutionary claim, and which, by the way, is now freely available online. And two studies that directly refute it are hidden behind a publisher’s paywall? Come on. I’m not one to rant about my rights as a taxpayer, but in this case I’m tempted to. What we have is a situation in which the public has funded questionable science, but our ability to judge it is being blocked, first by an “embargo”, and second by the financial interests of a private party. Enough! I know that there are a lot of financial implications of the Open Access model to work through, but the time has past when publicly funded science can be hidden from public scrutiny.