Genome Evolution Course 2011

www.yanaiweb.com/genome

Itai Yanai, Technion – Israel Institute of Technology

Michal Levin, Teaching assistant

Problem Set #5 assigned May 23^rd, 2011

To be submitted by email as a PDF or WORD document in English or Hebrew by midnight June 7^th, 2011 to Michal Levin mlevin@tx.technion.ac.il.

Evolution by Horizontal Gene Transfer. Solve the following five problems.

Problem 1: Gene loss vs. horizontal gene transfer. The tree below is based upon a multiple alignment of a concatenation of the ribosomal proteins from various organisms. Assume that it is the “correct tree of life”. The profile on the right indicates which kingdom each genome comes from (bacteria, yeast, archaea). For each of the three given phylogenetic distributions (a,b, and c) each representing a distinct gene family calculate:

· What is the minimal number of horizontal transfers that need to be postulated? Indicate also the (one) origin of the gene (for which this number is minimized).

· What is the minimal number of gene losses that need to be postulated? Indicate also the (one) origin of the gene (for which this number is minimized).

· Assuming that a gene loss is 4 times as common as a horizontal transfer, which of the previous two solutions is the most parsimonious scenario?

Problem 2: Anomalous DNA composition. Why does the method of anomalous DNA composition fail to identify ancient transfers?

Problem 3: Evolutionary scenarios. Explain how in light of horizontal transfer the following two sentences are incorrect:

• Any two homologs within the same genome are paralogs.
• If a gene is universal, i.e. exists in all genomes like a ribosomal gene, it is expected to yield a tree free from horizontal transfers because a transfer of a redundant gene would not be useful.

Introduction to problems 4 and 5: We will be using the Unicellular version of the COG database. First: Choose a COG that appears to be a good candidate for horizontal transfers. What we will be mainly trying to detect are inter-kingdom transfers (the most impressive of all transfers!). Furthermore, it will be most useful to zoom in on transfers between the bacterial and archaea kingdoms. I would recommend selecting a COG with a phylogenetic pattern containing more than one but not full representative in archaeagenomes and more than one but not full representation in the bacterial genomes. Representation is marked as “|” and absence as “-“ in the profiles:

The archaea genomes are highlighted in yellow on the left, followed by the eukaryotes in purple, and then various kinds of bacteria in many colors. You can get a better look at the COG by clicking on its names. I chose COG0114 - fumarase. From the fumarase phylogenetic profile, we find a scattered phyletic distribution. Representatives of the gene family are present in 5 of the 13 archaea genomes and in two of the three eukaryotes. In bacteria, all actino-, alpha-, proteo-, and most gamma- bacteria have a fumarase gene. However, gram-positive and other (less classified) bacteria do not have the gene. What are the two possible explanations for this profile?

Problem 4: Do the archaea and bacterial genomes cluster separately in your COG? Hint: From the fumarase tree, it is observed that VNG1356G (an archaea gene) does not cluster with the other archael genes (see below).

Problem 5: Unexpected ranking of sequence similarity among homologs. Can more evidence be gathered for horizontal transfer for the misplaced gene on the tree? Compare the sequence similarities of the gene you have chosen as a possible horizontal transfer. Hint: This can be easily achieved by clicking on the gene of interest (in the gene list below). The output is shown sorted by the degree of similarity. VNG1356G, the gene I have chosen to examine, is most similar to fumC, an E. coli gene, followed by many other bacterial genes. The gene MA1001 from the Mac archaea genome which was expected to be more similar to VNG1356G than the bacteria, is much further down in sequence similarity in this list. This strongly suggests that a horizontal transfer occurred here. (Note that further analyses are required in order to determine the directionality of the horizontal transfer which you are not required to do here).