Monday, June 04, 2012

Jonathan Eisen's #Genomics #Microbiology #Evolution Online Office Hours

OK.  I love that people ask me lots of questions about evolution, genomics, microbes, a mix of the two, etc.  But I just cannot keep answering single one-off emails about these topics.  So I am starting online office hours.  If you have any general questions about phylogenetics, evolution, genomics, microbes, or any of the work done in my lab, please post your questions here.  And I will try to answer them.



  1. I wonder what you think about SSU rRNA genes as measures of community profiles, which are now really common in all of environmental micro and microbial ecology. I also am curious your thoughts on the utility of measures of phylogenetic clustering (like David Ackerley's tool phylocom, but there are many) to infer evolutionary pressures on communities. At first I thought of these tools as breathing new life into SSU ribosomal RNA data, but maybe they distracting us from asking questions about what shapes microbial communities in better, or more sophisticated ways?

    1. Not 100% sure what you are asking but some comments here.

      1. I think SSU rRNA analyses have been and will continue to be critically important for microbial ecology studies, despite being imperfect

      2. With the increasing use of UNIFRAC and QIIME many researchers are in essence doing something equivalent to PHYLOCOM. PD (phylogenetic diversity) is being used extensively to both measure single communities and also to compare and contrast communities.

      3. I am not sure these approaches have breathed new life into microbial ecology studies per se. PD and phylogenetic ecology measures are used in microbial studies mostly simply to calculate distance between communities (e.g., with UNIFRAC). There has been much less in the "phylogenetic ecology" area especially in comparison to plants and animals, although see papers like Kembel et al. 2011 (on which I am a coauthor)

  2. Hi Dr. Eisen--I'm the National Geographic researcher that's been trying to get in touch with you (Kelsey Nowakowski). I understand that you're getting an overwhelming amount of inquires and I would like to ask mine here, but I'm sure that would breach some type of confidentiality policy I've agreed to follow. That being said, is there anyway you could briefly talk on the phone with me? I just want to check the plausibility of a couple ideas. I assure you it won't take too long. I watched your talk from the Compass Summit. Perhaps this project could help get people interested in a Field Guide to Microbes. I'd buy that app.

    1. Kelsey - I will track you down today or tomorrow ...

  3. Hello Dr. Eisen,

    We have discovered a very unusual enzyme that is involved in the biosynthesis of an antibiotic, which attaches ribose-5-phosphate to the polyaromatic aglycone. We also have the crystal structure of the enzyme available, but that has not been published yet.

    The enzyme family is very poorly characterized and all of the homologous sequences (~755) deposited into databases are annotated as "pseudouridine glycosidases" that are involved in primary metabolism in degradation of RNA. My suspicion is that some of these sequences are not annotated correctly, but they might be doing something novel/more interesting. The problem we are facing is finding out which ones are the non-pseudouridine glycosidases.

    To my mind, we have a situation where have in a single phylogenetic tree many orthologous sequences with few (interesting) paralogous sequences. I have been trying to convince myself that the way to identify the paralogous sequences would be comparative phylogenetic analysis, where we construct another tree (from for instance 16S rDNA sequences) that conveys the evolutionary relationships of the various organisms. The "real" psudouridine glycosidases should follow this normal evolutionary distribution (unless there is horizontal gene transfer) and any outliers we may find would be the interesting ones. But it seems to be very annoying for me to do this so - is my logic flawed in some way? Is there a better way to do this?

    I hope you have time to answer these questions!


    1. Well, the first thing I would do is build a tree of ALL the homologs and overlay onto that true known experimentally determined functions (rather than annotations). That way you can see if know functions group into any part of the tree.

      This is basically what I suggested doing in My Original 'phylogenomics' papers


Most recent post

Another day to think, to pause, to ponder.

Panorama of Sycamore Park and the memorial to Karim   A bit over 10 years ago I wrote a blog post that I repost all the time. Entitled "...