Tree responsibly: Part 2

In part 2 of this series, I want to talk about how we can draw trees that promote tree thinking instead of undermining it. But first let’s consider whether we can or should worm our way out of this possibly time-consuming task. You might say, well, I just draw the trees – it’s the viewer’s job to interpret them correctly.* However, it is well established that the many members of any audience (whether students or practicing scientists) are not trained in tree-thinking and will come to incorrect conclusions about what a tree is meant to convey. 

The other option is to warn the audience that the way the tree is drawn is likely to evoke some misconceptions. Indeed, this is essentially the approach taken on the evogeneao website, which has an interactive tree of life and other teaching resources. This page has a long section about distortions and limitations that explains that their tree is drawn to tell the human’s story and thus makes it look like evolution is goal-oriented (with the goal being humans). I'm glad they included this section and it may provide its own learning goal (leading the reader to say, how does this tree make me think that? How could it be drawn not to make me think that?). But I remain nervous that readers won’t scroll and dig into what may be viewed as details, details. So my opinion is that if we want readers to interpret trees correctly, we should actually draw them in ways that dispel misconceptions and encourage correct interpretation.  Here are some of my suggestions and related thoughts:

First, those of us that spend a lot of time wandering the phylogenetics literature know that trees more often look like what’s on the left than what’s on the right.

Tree on the left adapted from here.  

And what’s on the right looks a lot more like what’s in textbooks, review articles, and popular science.  So guess what’s in the middle?  People making choices about which taxa to show, how to draw the tree, and what story they want to tell with the tree.  Given that, let’s talk choices.

Ladder-y versus ultrametric

Both of the trees above have the same topology (implying the same set of relationships), but the one on the left is much more likely to trigger ladder-thinking because some of the tips are higher and others are lower.  The easy fix is to make the tree ultrametric, as on the right, helping to communicate that all taxa are extant (as they are in this pretend example) and that all have continued to evolve towards the present. 

We also choose which taxa to include, and by simply choosing those which make our trees a little more balanced, we can make it harder to view any tip as ‘basal’. Berkeley's Understanding Evolution site has a quite a few other nice tips for making trees that are less prone to misinterpretation.

Maximally unbalanced versus more balanced (oh and I rotated it at one node)

So now a suggestion that is possibly more painful, which is that those of us that estimate phylogenies take on more of the burden of creating education-friendly, misconception-fighting, biology-celebrating tree diagrams. Work on flatworms? Awesome! Make a lively worm-y phylogeny for your homepage that can be freely downloaded and shared. Put it on your lab t-shirt! Put it in the corner of your lab posters! Take it to outreach events! I know that most of us phylogeneticists are not graphic designers, but I think the dividends to come from one glorious, engaging figure are worth the time (and/or monetary investment). And we have such a long way to go to outnumber the 'march of progress' type images that dominate the representation of evolution.

And my last thought is just that I wish that some kind programmer would make a truly easy piece of software to draw trees for educational purposes. Because if you are wondering what I do to make tree figures for teaching (and blogging), I create a pretend dataset in Mesquite and manipulate the tree to fit my purposes. Which sometimes require tinkering in Illustrator. Most teachers don’t have the time to learn Mesquite solely for this purpose, much less the funding for Illustrator licenses. Please feel free to reply to this blog if there are programs you think meet this need.**

In closing, thanks (I guess) to Matt Hahn for trolling me with tree figures to provoke me into writing about this.

____________________________________________________________

*And a few years ago that was more or less my view. We can’t ‘protect’ everyone from seeing trees that foster misconceptions, so it’s better that we teach them how to read trees. We certainly do need to do the latter, but it’s not enough. People who are experts in this area also think so.

**I am aware of FigTree and TreeEdit, but I still think these are probably too fiddly as they require a user to start with a tree file. I’m picturing more like the user says, "I need X tips" and then they move the branches, annotate branches with words, change formats, etc. Mesquite can do most of theses things, but having taught with it, I don’t think it’s reasonable to expect a teacher to wander through the menus just to make a vertebrate phylogeny that’s less ladder-y than what is in the textbook. I also don't think it would be fair to put all the burden on Mesquite, which already serves many needs in the systematics community and does so beautifully. Maybe someone who is good at Java could make a MesquiteLite to pull out and adapt the drawing tools. Even better if it could be an interactive website that students could use! But I'm getting ahead of myself...