Now I’ve discarded that “Spotlight on” jazz, I have to come up with slightly snappy titles, which is kind of annoying. Oh well, today we’re going to be talking about evolutionary trees/trees of life/phylogenetic trees (if you must, but a lot of them are still made without using genetics at all).
Here’s a great example:
The first tree like this was made in 1865. This dude, St George Jackson Mivart, gathered a heap of detail on different primates, and then made a tree showing which were most and least similar. It’s obvious that lemurs are pretty different to orangutans. But when you’re looking at two species of monkey that are both small and adorable, it’s kind of difficult to know which one was made into a pet by Justin Beiber, and which one… well, was lucky.
So St George Jackson Mivart made a tree separating primates on their spinal columns. Cool, a relationship between these species. He then made another tree, separating primates based on their limbs… and got a different tree.
Obviously this wasn’t an ideal situation, and things haven’t improved much.
Today, DNA is taken as the gold standard, because it shows changes that might not be physically seen. Different genes show different relationships, and while one study might use a particular gene to determine where on a tree an organism is, another gene might show something completely different.
As it turns out, another gene does show something completely different, basically all of the time (1).
Because of this, the dudes that did that study, Salichos and Rokas, developed an algorithm to check which bits of the “tree of life” (although in this case, more of a “tree of yeast”) were legit. This now allows really good (informative) genes to be selected with which to build these trees. (I kind of want to go on a rant about how this is a great example of singular value decomposition, but that’s really just exams talking.)
The thing is, sometimes people prune their trees really strangely. If you pick the wrong gene (or the wrong trait) to look at, then you’ll get really funky results. If you’re looking at the ability to open jars as a trait, we’re really closely related to octopuses (octopodes, if you will), but not so much to most mammals. But if you look at, say, having internal thermoregulation (“warm-blooded”), making milk for our young, giving birth to live young, being covered in hair, and having five digits at the end of each limb, then you come up with a tree that makes us most like other primates, similar to mammals, and not very much like octopodes.
The thing is, most of the time we don’t need all the information about random genes that everything has and there’s a little bit of divergence in. We’ve all got genes for ribosomes, and DNA polymerase, and that other really important junk your cells need to divide, but chance can make those genes in us look more like those genes in a cactus (pl: cacti), or a platypus (pl: platypus) than in a chimpanzee or bonobo.
It’s way better to look at things that actually inform on whether humans are more like cacti than capuchins, and that’s what these dudes have done, using some sweet computational algorithms and a large amount of yeast.
1. Salichos and Rokas (2013) “Inferring ancient divergences requires genes with strong phylogenetic signals.” Nature doi:10.1038/nature12130
2. Buckley TR, Attanayake D, Bradler S. (2009) Extreme convergence in stick insect evolution: phylogenetic placement of the Lord Howe Island tree lobster. Proceedings of the Royal Society, Biological Sciences. 276: 1055-1062