homework07.md

Homework 7

Problem 1

You know what's awesome? Aquatic mammals. Seals, whales, manatees, they have all the cool stuff that sets us mammals apart: mammary glands, rockin' three-bone inner ear design, and a powerful four-chambered heart, among other helpful innovations. But these guys got tired of the grind of climbing trees, dealing with the sun, being thirsty, and walking all the time. Legs aren't cool. You know what's cool? Flippers. Mammals need oxygen, can't breathe underwater, you say? No bro, you can't breathe underwater because you lack imagination. Aquatic mammals think outside the box. They say, jump in, the water's just fine.

What we're going to do here is figure out how the aquatic mammals fit in with the rest of the mammals. Did the aquatic mammals never come out of the water in the first place, evolving from an aquatic ancestor? Did they evolve from a single land mammal that decided to go back into the water? Or, perhaps, did several distinct land mammals independently decide they preferred to live at sea?

Let's do a little background on the history of mammals. The extant mammals comprise three monophyletic groups: monotremes, marsupials, and eutherians. There aren't that many monotremes, and my favorite would be the duck-billed platypus. Marsupials you know and love. Eutherians, that's the rest of us lot, the ones that have a placenta. The most agreed-upon picture of the relationships between these three groups is like so:

That is, the data suggest that the eutherians and the marsupials are more closely related to one another than to the monotremes.

Then, there is quite a bit of diversity within the eutherians. Lots of both bark and bite. Here's a clouds-level picture of the relationships between the eutherians:

Ever wondered whether a dog is more closely related to a pig or a horse? Or how the shrew and the elephant are related to the elephant shrew? Boom.

I've hidden some stuff from this eutherian tree above. In particular, there's no mention of our aquatic mammal friends. You're going to use protein sequence data from the cytochrome b gene residing on the mitochondrion to figure out where they go, and in the process, try to answer the questions we had above about the history of the aquatic mammals.

1. Make a FASTA-format file named hw08_lastname.fasta that contains a number of cytochrome b protein sequences, as follows: grab one representative of each of the 12 terminal groups on the boreoeutherian clade. For instance, for "Carnivora" you might use a sequence from a grizzly bear or, being true to your school, a panther. For the atlantogenatan clade, just grab elephant, aardvark and armadillo. Then grab sequences for three aquatic mammals: the blue whale, the harbor seal, and the manatee. Then grab duck-billed platypus to use as an outgroup. There should thus be a total of 12 + 3 + 3 + 1 = 19 sequences in your file. You'll probably want to modify the headers in the file to have just the organism's common name, in order to make the rest of the assignment easier. For example, this guy:

   >gi|5835120|ref|NP_007172.1|CYTB_10543 cytochrome b [Equus caballus]

   you'll probably want to replace with

   >Horse

   As a sanity check, the protein sequences should all be of roughly the same length. Use the web to help mapping an organism's common name into its latin binomial name.

   While you can certainly just search "this_mammal cytochrome b" in NCBI or Ensembl to do this, there's a faster way. One way to speed things up is to use blastp. Grab one cytochrome b sequence, e.g. that panther's, and blast it against "refseq_protein", restricting to "Mammalia", and then tell blast to return 5000 target sequences. When the results finally come back, click on "Taxonomy Report" to arrange the results by taxonomy. That should give you a screen full of hits that makes it easy to get cytochrome b sequences. Just click one of the links on the right of the page to get to the corresponding sequence's reference page.

   Include this FASTA file in your submission, please.

2. Align the sequences you gathered, and make a phylogenetic tree out of them. I'd recommend using EBI's MUSCLE server, because you can just use it online and it makes it easy to get a phylogenetic tree. Include the tree image in your writeup.

3. What do you find are the the most similar sequences, respectively, to your three aquatic mammals?

4. Based on your tree, do you think that aquatic mammals descend from a single land-based mammalian ancestor that went back into the water, or do you think that several distinct land-based mammals went back into the water independently? Explain.

5. Is your answer above definitive? If so, explain why. If not, explain what additional evidence you would seek in order to be more definitive (with respect to single/multiple origins of aquatic mammals).

6. Why is cytochrome b often used by molecular phylogeneticists? There are plenty of other genes, and there are plenty of other mitochondrial genes. Why do researchers use this one?

Problem 2

Consider the following distance matrix:

	A	B	C	D	E
B	5
C	4	7
D	7	10	7
E	6	9	6	5
F	8	11	8	9	8

You're going to build a phylogenetic tree based on these distances using two different methods: UPGMA and Neighbor-Joining (NJ). In addition to the course slides, you might find these slides helpful.

You'll use the program neighbor from the PHYLIP package to build the trees. Head to this website, which will allow you to run neighbor online. (You can also download it easily if you want to run it locally.)

Once you're there, drill down to phylogeny > distance > neighbor to get to the interface for the program. Paste in the distance data as formatted in the included distances.txt file. Tell the site not to email you, and click open the 'advanced options' button. The only option you want to mess with under 'advanced options' is "Distance method", where you can switch between NJ and UPGMA.

1. Build an NJ tree. Either by hand or on a computer, redraw the tree that the program finds, including the branch lengths. Be sure to correctly represent in your drawing whether the tree is rooted or not.

2. Build a UPGMA tree. Either by hand or on a computer, redraw the tree that the program finds, including the branch lengths. Be sure to correctly represent in your drawing whether the tree is rooted or not.

3. The two trees should be different, both in terms of topology and branch lengths. Which tree would you say better represents the distance data, and why?

4. What went wrong with the method that made the worse tree?