Badgeromics 2008, er Metagenomics 2008

I'm attending the Metagenomics 2008 conference being held locally at UCSD. There's plenty of interesting talks (see Jonathan Eisen's coverage). However, today what struck me the most (egotist that I am) was how my work was brought up in three separate talks...

Alex Worden started out the day by discussing her work on picoeukaryotes (unicellular eukaryotes less than 2 microns in size) and there were a couple of slides dealing with the comparative work that Andy and I have done as part of her forthcoming Micromonas paper.

Then practically half of Shannon Williamson's talk about her metagenomics work on marine viruses was devoted to my phylogenetic pipeline, APIS (Automated Phylogenetic Inference System), which performs taxonomic binning of metagenomic data by the automated creation and interpretation of phylogenetic trees.

Then in the evening Jonathan Eisen brought up our 2005 paper as part of the justification for his new project to sequence hundreds of bacterial and archaeal genomes.

Thanks guys. Your checks are in the mail...

The Genome Sequence of Phaeodactylum tricornutum

What's this? Another eukaryotic genome paper with me on it? I blame Jonathan Eisen for putting me on that road to ruin -- I never touched the nucleated rascals before Tetrahymena. 

But seriously, single celled eukaryotes are cool, and not Eisen but rather Andy Allen (who doesn't have a web page or blog to link to, unfortunately) is the one to blame for getting me involved with his beloved diatoms.

P. tricornutum is the second diatom genome to be sequenced, and first of the pennate (elongated) lineage. We discovered that this genome was quite different from that of Thalassiosira pseudonana, the first diatom genome to be sequenced , which is a member of the centric (hatbox/cheesebox/petri dish) lineage.  We also identified about 600 genes that we believe to be of bacterial origin in this genome. As this was part of the analysis that I did, I'm perhaps biased in that I think this one of the most interesting results in the paper. This analysis wasn't just "BLASTology" (being just based on BLAST output), but rather based on phylogenetic trees for each predicted protein coding gene in the P. tricornutum genome.

There's been some press about the paper -- including this one in the German newspaper Die Welt -- whose sub heading "Diese Lebewesen entpuppen sich als eine kuriose Mischung aus Pflanze, Tier und Bakterium" implies that P. tricornutum is a curious mix of plant, animal, and bacterium -- not exactly true, but hints at the HGT analysis that we did. There's also this press release that mentions me and Andy by name, but so far no papers have used that part :-) 

Anyway, the paper (which is under a Creative Commons License, showing that even commercial publishers like Nature can be reasonable) is up on Nature's advance publication page as of yesterday. Check it out. And the supplemental, where most of the actual data is.

The Internet doesn't lie...

My immunologist friend Andreas Krueger has recently pointed out to me that according to a certain website the reading level of this blog is "Genius". I'm sure this is every bit as scientific as those quizzes that tell you which Star Wars character you are most like. Also, this ranking was made prior to this post -- it would be amusing if this post somehow weakened the rating...

Erdös numbers & other geekery

RPM over at Evolgen has recently posted on the (perhaps facetious) need for a version of an Erdös number in biology. If you don't know what an Erdös number is, it basically is the geekier version of "six degrees of Kevin Bacon". Paul Erdös (1913-1996) was a prolific mathematician who had many co-authors. If you co-authored a paper with Erdös, you have an Erdös number of 1, if you didn't collaborate with Erdös himself but instead with one of his collaborators, you have an Erdös number of 2, and so forth. It's a point of pride among mathematicians to have a low Erdös number.

Well, shouldn't biologists join in on the fun? RPM has suggested that Richard Gibbs should be our Erdös. Of course, he's honest enough to mention that he has a Gibbs number of one, but Gibbs isn't actually a bad choice, at least in genomics. I suggested Craig Venter or Claire Fraser (as I have a both a Venter and a Fraser number of 1), but RPM suggested that I wasn't really getting into the spirit of things. Still, (as you'll see), these connections still help me.

So first of all, having done a postdoc in a Computer Science department, I'd just like to state that I have a respectable Erdös number of 4. (Paul Erdös -> Charles Colbourn -> Derek Corneil -> Paul Kearney -> me).

My Gibbs number isn't so bad either, at 2. (Richard Gibbs -> Craig Venter -> me)

Someone on evolgen suggested Eric Lander. My Lander number is 3 (Eric Lander -> Michael Eisen -> Steven Salzberg -> me).