Making predictions for phylogenetic trees


Textbook readings


Notes, Comments and Updates

What's the point?

  1. To be able to make predictions about an organism based on it's phylogenetic placement


If a phylogenetic tree is to be useful as a model for evolutionary relationships, it needs to have predictive or explanatory value. In fact, trees allow you to make a wide range of predictions about an organism. If a new organism trees out as a member of a specific group, you would do well to predict that this new organism shares the properties that are generally found in that group. You would, however, be unable to predict properties that are variable in that group. For example, if you have an unknown organism and use it to create the following tree with representatives of all major phyla of Bacteria:

What can you learn from this? Well, your unknown trees out as a relative (although a distant one) of Bacillus, a Firmicute. But the next closest relative is, perhaps, Actinomyces, a member of a different phylum, the Actinobacteria. So, in other words, you've learned only that your organism is a member of the Firmicutes. Perhaps not a lot to go on, you might think, but actually you have learned a lot, because:

  1. Firmicutes are generally Gram-positive, i.e. they have a thick cell wall and lack an outer membrane
  2. Firmicutes are almost always heterotrophic
  3. Firmicutes are generally mesophilic
  4. Although there are lots of exceptions, a majority of Firmicutes are rod-shaped

And so you would predict that your unknown is likely a Gram-positive mesophilic heterotroph, and perhaps rod-shaped.

There are lots of perperties that vary a lot amongst Firmicutes, and so are properties you could not predict of your unknown:

  1. Some Firmicutes are pathogens but many are not
  2. Some Firmicutes are aerobic, some are anaerobic, others are facultative
  3. Endospore formation is found only in the Firmicutes, but many Firmicutes cannot form spores
  4. Most are hereotrophs, but what organic compunds they can use, and how they use them, varies widely

These, then, are properties you cannot predict of your organism.

So, what if you create a more specific tree of your unknown - based, perhaps, on the fact that a BLAST search showed that your sequence is similar to that of some Heliobacteria:

This tree is much more specific, and so you an learn a lot more. As you can see, it's closest relatives are members of the genus Heliobacterium. In fact, if you identify the common ancestor of all species of Heliobacteri, you can see that it is also an ancestor of your unknown - in other words, your unknown is (or should be) a species of Heliobacterium. As such, you would predict that it shares all of the properties that are common to other species of Heliobacterium, including:

  1. It is photoheterotrophic - getting it's energy from light and it's carbon from organics (cannot fix CO2) (Note that this is not what you predicted from the previous tree!)
  2. It used bacteriochlorophyl G and a type I photosystem
  3. It is anaerobic
  4. It can fix nitrogen
  5. It lives in soil
  6. It is rod-shaped, and has an exterior S-layer
  7. Although it lacks an out membrane, and so is formally Gram-positive, it will stain Gram-negative
  8. It can probably form endospores
  9. It is non-pathogenic

On the other hand, there are a variety of ways in which species of the genus Heliobacterium differ - these would not be predictable from your tree. For example:

  1. What kind of soil or root habitat does your unknown live in?
  2. What is required to induce endospore formation
  3. What accessory pigments does it use
  4. What carbon sources can it use (other than pyruvate, which all can use)?

If these kinds of predictions mystify you, think of it this way - what can you predict about the Tammar wallaby (which I presume you've never heard of) from THIS tree?:

Well, to start out with, it's a mammal - and so presumably has hair, breathes air, has an internal skeleton, is homothermic (warm-blooded), and probably nurses it's young. It seems to be related specifically to the representative marsupial (the Opposum, or "speed bump"), and so presumably has a pouch where it's young are kept. You cannot predict whether is eats meat or plants (or both), where it lives, how big it is, or anything else that varies amoungst mammals in general, or marsupials in particular. If you think of these trees as they would appear with familar organisms, how you interpret them and make use of them might be easier to visualize.