Questions About Clusters
by Deborah Jacobs-Sera
Hi all,
Here are questions to consider.
- Why did Dr. Hatfull use words like arbitrary, convenience, and lie when describing the clusters?
- Do phages belong in clusters (of any sort)? Why do we put them there?
- How does the question get answered?
Do you have any questions about Dr. Hatfull's presentation? Add a separate post if you do. Thanks!
1. We do not know if phages really do belong in clusters since there are only 60 phages that were put into clusters and there is a much larger population of phages -- to generalize about all phages from just those 60 phages wouldn't be right. Putting them in clusters though can be used for convenience (see question 2). 2. I think phages do belong in the clusters of genomic similarity - the clusters Dr Hatfull talked about in the last meeting. You can also group phages in terms of morphology, type of bacteria they infect, or other characteristics but I'll just talk about clusters of genomic similarity. We put them there because it helps to identify groups of phages that may have similar morphologies, behaviors, and other characteristics. When finding out a usage/characteristic of one of the phages, you can see if this usage/characteristic is present in all phages of that group. Putting phages in clusters of genomic similarity can also give insight as to the evolution of phages - the more similar the two genomes, the more likely they evolved close together in the phage evolutionary tree. 3. An important thing Dr Hatfull brought up was that only 60 phages were clustered and thus there are tons of phages that haven't been put into clusters. Will they follow the pattern and fit into those clusters, make new clusters, etc.. we do not know. This question can be answered if we sequence more phages to be able to better generalize about the phage population.
One thing I was curious about (that Dr. Hatfull didn't really talk much about) was what phages being in the same cluster meant for their evolutionary relationships. Most likely, phages in highly similar clusters have more recently diverged, but in cases like he brought up with cluster H, it may be a little cloudier. For example, two phages might share sequence similarity over a large percentage of their genome, but might have a low degree of identity in that sequence similarity. Two other phages might only be similar over 50% of their genomes, but have those 50% be almost identical. I wonder what that says about who's more related to whom. Maybe--given the "orgy of recombination" (Dr. Hatfull's published words, not mine) that takes place amongst phage, and the mosaic nature of phage genomes--it's just not really possible to say too much about evolutionary relationships between phages from DNA alone. Please feel free to enlighten me!
Dan, Your questions make me just look at the data and wonder if it demonstrates an example of time in the process. I guess what I mean is can we see a set of changes among phages that suggest this must be first, this must be second, ... Also when calling genes, there are lots of tiny predicted genes that may very well not be proteins. They are tiny with very limited identifiable coding potential. However, when that sequence is represented in many genomes, doesn't that suggest conserved sequence and doesn't conserved sequence suggest necessity? And related to your comment again, if that conserved sequence is similar in genomes of one cluster does that have a different value than if you saw that same conserved sequence in genomes of different clusters?
I think that in this context, question 1 almost answers question 2. We make the clusters as an arbitrary way of organizing the genomic data that is convenient for use when analyzing the data itself. Other than this, there are not many more methods of classification of phage other than maybe host range and virion morphology. It's just another way of keeping track of things and tracking evolutionary relationships and phenotypes. I am curious to see what happens as we sequence more and more genomes using higher throughput methods. Previously, the genomes were done one by one, and the ratios of the clusters changed rather slowly. Now that we are sequencing 12, or even 48 at a time, I'm curious as to how the cluster distribution will stand after the next 454 run.
How about some of the newer phagehunters answer??
Yeah. What Drew said.