Friday, July 13, 2007

Metagenomics 2007 Update

Some brief notes on the Metagenomics 2007 meeting. For those who do not know - metagenomics is the simultaneous sequencing of the genomes of communities of microbes.

Day 1 : I missed the introductory talk by Larry Smarr (my flight was late). Then there was an opening poster session - a diverse collection of things. My favorite were those involving methods to sequence the genomes of single cells (e.g., one by Ramunas Stepanauskas). Such methods will be critical for getting reference genomes from organisms that are not abundant in a community but might nevertheless be important.

Then there was a dinner and a post dinner talk by Masahira Hattori. The talk was not overwhelmingly interesting ... sort of a good example of metagenomics as a fad. Hattori has extensive experience in genome sequencing - having been involved in sequencing the human genome as well as the genomes of other large charismatic furry organisms. As with many people who worked on the human genome, he moved into microbial genomic studies. He has done some interesting things in this arena, the most interesting to me being the sequencing of the genome of a symbiotic bacteria called Carsonella. He did talk about this and I learned something I missed in the paper - this genome was sequenced with the aid of genome amplification methods (see above).

Then Hattori talked about his recent work on human microbiome metagenomics. He has done a lot of work in this area including a big project in which 80,000+ sequence reads were generated for 13 healthy Japanese individuals (both adults and children). I confess much of his results/analyses were not convincing - they seemed to me to be something of the order of "I know I should be doing metagenomics - here is some." His talk did lead to one of the funnier moments of the meeting - someone asked a question that went something like this

"I know when I go to Japan my intestinal tract changes" leading to a groan from the audience and at least one "A little too much information" comment. The questioner was clearly going somewhere with then but before he got too far, Hattori interjected "It gets much better"

Day 2: Some notes.

Day 2 was a pleasant surprise. For many metagenomics related meetings you hear the same talks and the same data over and over and over again. That was not the case here.

The first talk was by Jed Furman. This was truly inspiring to me. He presented a tour de force of environmental microbiology where metagenomics was simply a component and he interwove the history of metagenomics too. Among his topics - use of ARISA in surveying microbes, studies of the functions present in marine archaea, species richness versus latitude, the scalar nature of sampling for microbes, changes in ocean communities over time, and functional redundancy. This to me is the ideal use of metagenomics - as a tool in other studies. In this sense I disagree with those who say metagenomics is a field. I view it more as a tool. A powerful tool, as Furman showed. But most powerful if used in the context of other studies. In particular he said "Remarkably few experiments other than sequencing have been done in environmental genomics and I hope that will change"

Other talks were by Dawn Field, David Relman, Jim Tiedje, Janet Jansson, Forest Rohwer, and Mitch Sogin. Just a few tidbits from these:
  • Janet Jannson who talked in part about a twin study of microbes associated with Crohn's disease lauded the idea of a Genomic Encyclopedia for Bacteria and Archaea (a project I am leading at the JGI ... if you want to know more stay tuned or ask). Also she referred to humans as "fermenters of microbes" a phrase I kind of like.
  • David Relman gave a good overview of his and others' work on the human microbiome. One thing he emphasized was the need to remember that rare organisms can be important and that most metagenomics projects do not sample the rare bugs well. And he too then talked about single-cell genomics. He also mentioned a paper I have been waiting for and did not know it had come out. This is a paper on microbial colonization of the human gut in infants by Chana Palmer et al. (with Pat Brown as senior author). In this paper they use a rRNA chip to track microbes colonization infants. This paper is spectacular. Really.
  • Forest Rohwer, who talked most about coral associated microbes and coral damage by people started off by saying "This is my humans suck talk"
There were also breakout sessions, one of which I chaired on "Computational Metagenomics." Our general conclusion from this - metagenomic informatics is really difficult, we need more interdisciplinary efforts to develop new approaches, and that more money is needed for computational efforts if we want to get the most out of metagenomic data (like we were going to say we need less money). All jokes aside, I do agree with this point - metagenomic sequence analysis is orders of magnitude more complex than analyzing any single-organisms genome. And we need a major ratheting up of computational efforts if we are to get the most out of these data sets that are being produced.


  1. 1) sorry I didn't catch up with you later on Friday - had an early flight to catch.

    2) re: your point on metaganomics as a field/tool. I often have similar discussions with those in the remote sensing field. Defining 'remote sensing' as a field is a bit of a misnomer - however, using r.s. to understand more about the biophysical characteristics of the subject is a way to contribute to one of many fields, such as physiology, biophysics, ecology, biometeorology, etc. This is analogous to using a microscope to view, produce data and learn more about the characteristics of a subject that can not be seen without megnification (ok, it is a simplistic example, but illustrates the point). By calling remote sensing, metagenomics, etc... a tool, it is not downplaying the importance of these tools, but a spade is a spade, and I think discussing in those terms avoids some confusuion when trying to explain the science and associated methods to those in other fields.

  2. The session led by Pat Schloss was great, because it was really about the limitations of metagenome data analysis that don't necessarily get much airtime, such as "we found the sequence here and not there, so the function is here and not there".

    And the Fuhrman work opens a new dimension in the field - with work started over five years ago!

  3. Y. Winogradsky6/16/2008 4:27 AM

    The EU-funded project MetaHIT (Metagenomics of the Human Intestinal Tract) has an official website:

  4. The first MetaHIT international conference on metagenomics will be organised in China in March 2010. More info on our website: