On Microbial Sequencing

This is the 9th year the NSF & USDA  hold  a workshop for their microbial sequencing program awardees. (Full disclosure: I am not one of them).  Most of the talks are by the awardees themselves, and there were some great talks. For me an interesting angle was it to see how software is being developed as an integral part of the research. The workshop was organized by Lita Proctor from the NSF and Ann Lichens-Park from the US Department of Agriculture,  and it was well worth getting up for at the ungodly hour of 6:45 am on two sunny San Diego weekend mornings. (Saturday and Sunday, 10-11/ January).

The Schmidt Lab is interested in upland soils are the Earth’s largest methane sink. In agricultural soil this capacity decreases: less methanotrophs. They estimate it takes 75 years for a microbial community to recover once agricultural soil stops being tended.They have shown a clear decrease in methanotrophs in agricultural soil, but it is not clear why the decrease occur.

Another interesting find in his lab was a technological one; they found an artifact in 454 sequencing: artificially replicated sequences. He estimated that about 10-30% of sequence replications found in his 454 runs are due to this artifact. The Schmidt lab actually wrote software to correct this, and they made it publicly usable via a web site.

Shannon Williamson from JCVI talked about marine viruses. Her group isolated bacteriophages (viruses that infect bacteria) using a combined filtering and bacteriocidal approach. When they analyzed the resulting  viral metagenome, they found genes that are homologous to known bacterial genes, yet cluster together in a “viral” branch of the tree. The most interesting explanation would be that some of these genes have been appropriated by the viruses to maintain the host while viral replication takes place. It is hard to rule out simple horizontal gene transfer though, since those viral genes were not compared with bacterial metagenomes from where the original samples were taken, but with known genomes only. Then again, how do you get a “bacterial only” metagenomic samples with no viral contaminants (including lysogenic phages)?

Eric Allen from the Scripps Institute of Oceanography talked about the metagenomics of extreme halophiles taken from Lake Tyler in Australia. Extreme halophiles live in very saline waters,. An interesting argument he made was for the tractability of such a system, due to the small number of species. He showed a cool slide of  a stick-figure standing at the bottom of a staircase, with the stairs themselves getting progressively higher, representing different microbial habitats ordered by their complexity.  The bottom stair was very low, and that was the acid mine drainage system. Acid mine drainage kills almost everything, and there are less than 10 microbial species that have been identified in this habitat. The next stair, slightly higher, was that of extreme halophiles, more complex than acid mine drainage, yet still species-poor. The top stair, twice as high as the stick figure itself was soil metagenomics: >10,000 species and quite insurmountable! One cool thing about extreme haliophiles is their shapes: many have regular geometric shapes: diamond shaped, or squares.

Square Microbe

Square microbe! from Bolhuis et al. BMC Genomics 2006 7:169 doi:10.1186/1471-2164-7-169

The second day was opened by James Collins,  assistant director for directorate of biological sciences at the NSF. He talked about  NSF’s need to keep its finger on the pulse of biology: the field is in flux both due to new technologies that open up new questions, and changes in the biosphere itself due to global climate change: we are going through the largest mass extinction since the KT (dinosaur) one.

One cool figure that he had illustrated “trans-disciplinary” versus “interdisciplinary” vs. “multidisciplinary” vs. “separate disciplines”; reproduced here from memory. I guess that taking such complex human activities and interactions and placing them into Venn diagrams holds an appeal to my need to simplify and categorize.


Cross-talk between scientific disciplines.

Then the floor was opened for discussion.  Seemed like almost everyone agreed that the more genomes the better, and they would like to see more genomes and metagenomes coming up. A heated discussion then opened as to how funds should be allocated within this rather nebulous goal. To me it is clear that all levels of genomic annotation are currently insufficient. Some estimate gene calling error rate in whole genomes to be > 30%! I won’t even try to think about how high the error rate might be in metagenomes.  Computational Function Prediction is problematic too (I should know, I have been organizing a function prediction meeting for the past 4 years).  Perhaps we need more money for developing high throughput experimental methods as we continue with the bioinformatics. As much as I hate to say it, the curve that represents advances in bioinformatics annotation methods seems to be not as steep as it was five years ago.

The last talk I attended was by Nancy Moran from Arizona University. Nancy is a co-author of a paper I really like in PLoS Biology that talks about insect symbionts.The Glassy Winged Sharpshooter is an insect that feeds on tree sap. It harbors two bacterial symbionts: one supplies the vitamins to the host, whereas the other produces the essential amino acids. their genomes are very much in accord with those functions, each actually lacking the pathways the other provides, and feeding off each other’s products as well as feeding the host.

Another cool insect symbiont is the aptly named Hamiltonella defensa. This microbe protects pea aphids from parasitic wasp larva. The wasp lays its  eggs in the live aphid, and in an aphid that does not have H. defensa that larva would hatch and eat the aphid alive from the inside-out. But good-ol’ defensa kills the larva when it hatches in the aphid. (Or bad-ol’ defensa if you view aphids as a pest).  H. defensa is transmitted between insects (horizontally) as well to the next generations (vertically). Interestingly enough, some of these toxins are not encoded by the genome of H. defensa, but by a phage, APSE, that infects the bacteria. Aphids that carry the bacteria survive better than aphids that don’t, and those that carry the bacteria which, in turn, carry the virus survive even better.

I will let Jonathan Swift finish this one for me:

"So nat'ralists observe, a flea
Hath smaller fleas that on him prey,
And these have smaller fleas that bite 'em,
And so proceed ad infinitum."

Now excuse me while I shampoo my kids for head lice and boil wash all their clothes & beddings.

Share and Enjoy:
  • Fark
  • Digg
  • Technorati
  • del.icio.us
  • StumbleUpon
  • Facebook
  • Reddit
  • Twitter
  • FriendFeed
  • PDF
  • email
  • Print
  • Google Bookmarks

Comments are closed.