Just got back from the "First International Conference on Metagenomics" which was held in San Diego. Despite that this is clearly NOT the first international conference on metagenomics it was not bad.
For those who do not know, metagenomics is the term used when people do DNA sequencing directly from environmental samples without isolating organisms in the first place. This term was coined by Jo Handelsman et al. in an article in 1998, where they referred to all the DNA and its coding potential in soil as the soil "metagenome."
The meeting was hosted by UCSD/CalIT2 which are trying to move into the metagenomics field in a large part due to the large grant they have from the Moore foundation to build a metagenomics database with the Venter Institute. The database is called CAMERA and it is planning to have its first release shortly.
To be honest, even though I am involved in CAMERA, the UCSD/CAMERA folks would be better off not trying to make it seem like they are the only people organizing meetings in this area. Nevertheless, the meeting was pretty good.
There were talks by people focusing on different aspects of metagenomics, including data collection, databasing, and data analysis as well as some interesting biology. My favorites were one by Jeff Gordon, from Wash. University in St. Louis. He is doing some of the most spectacular stuff in studies of the human microbiome and he discussed a few of the studies from his group. Most importantly, he emphasized the use of germ free animals as a model system. Basically, they raise animals in completely sterile conditions and have produced mice and fish and other species that have no microbes associated with them. This allows them to do experimental manipulations to ask controlled questions about host microbe interactions. My other favorite talk was by Ford Doolittle, who even though I disagreed with some of the things he said, he always challenges the audience to rethink their assumptions. In this case, he talked about the species concept in microbes and why he thinks it does not have much us.
Overall, I got the feeling that people were being a little too worried about the difficulties in metagenomics. Yes, analyzing sequence data from environmental samples is complicated. Yes, all the bioinformatics is harder because you are dealing with a mixed sample of DNA fragments and you do not know which fragment comes from which organism in the sample. And yes, the databasing and data analysis can be very complicated because the amount of raw data and metadata can be huge. But in the end, metagenomics has the potential to be an incredibly powerful tool in studies of microorganisms in nature. And the fact that it is somewhat harder than standard genome sequencing does not mean that we are not already learning a lot from it. What we need to keep in mind is that it is simply a tool - and to try and turn it into a field (which is what it seemed like some of the players would like) is a mistake.
If you are interested in the meeting itself, the talks and discussion sessions are available here.