Portal de Periódicos da CAPES

Riding giants

2007; Wiley; Volume: 9; Issue: 1 Linguagem: Inglês

10.1111/j.1462-2920.2006.01222_4.x

1462-2920

Philip Hugenholtz,

Protist diversity and phylogeny

We've all seen the graphs showing the exponential rise in sequence data in the public databases. This of course is the result of great strides forward in high-throughput sequencing and, incredibly, new technologies are on the verge of putting us into sequence overdrive. A traditional Sanger sequencer produces per run just ∼70 kbp, whereas a 454 pyrosequencer produces 30 Mbp, and Solexa is promising up to 1 Gbp. And other technologies are in the works. So it doesn't take a crystal ball to see that the bottleneck will rapidly become computational. In fact, Darren Platt, head of informatics at JGI, only half jokingly predicts that just storage of the data alone will become limiting and that the storage medium of choice in the future may be the DNA itself, i.e. it will be cheaper to resequence the DNA than store the information electronically. Hopefully things won't get to that extreme and we will step up to the computational grand challenge. What better target to aim this elephant gun at than microorganisms. After all, they constitute the bulk of the biomass and evolutionary and metabolic diversity on the planet. Ironically, we may run out of characterized microbial isolates to sequence in the not too distant future. Plans are afoot to sequence all ∼6000 described species which should take ∼200 Gbp; this amounts to 5 years of dedicated work at a production facility like JGI based on current capacity, but likely much less time as the new technologies come on line (200 Solexa runs?).1 The natural microbial world on the other hand represents a limitless source of sequencing targets, with the added benefit of no cultivation bias. Initial forays into microbial community sequencing (metagenomics) have been very promising, but reinforce our suspicions that we have barely scratched the surface of the microbial world. Indeed, if a recent survey of the deep sea using 16S pyrotags (Sogin et al., 2006) is anything to go by, we have barely brushed the surface. And this goes 10-fold for the virosphere. What will environmental -omics look like in the future? Here are a few predictions: snapshot samples will be replaced by time series and fine-scale spatial sampling, and via this, we will no longer have to try to understand the plot of the film by looking at the corner of one frame, we will see the film in motion. Collection of sequence-associated data, or metadata, will become more standardized and detailed facilitating meaningful correlations between communities and their ecological settings. Viral, bacterial, archaeal and eukaryotic fractions will be routinely sampled and sequenced in parallel so that different trophic levels can be analysed in conjunction. Expressed mRNAs and proteins will be routinely obtained and analysed from the same environmental samples to get a window on community function instead of just metabolic potential. Fractionating individual populations and cells from the community for independent sequencing will become commonplace and greatly facilitate dissection and interpretation of the community data. But moreover, population genomics will mature in its own right and sampling (sequence coverage) of naturally occurring populations will go much, much deeper bringing the evolutionary processes that drive and shape populations into sharp focus. Population geneticists will be drawn to the field in droves. Also, the structure and dynamics of microbial populations will be placed convincingly into their many and varied ecological contexts. The whole process will be much faster, and the data made publicly available much sooner in a fully integrated format. In order to make sense of these massive data sets, modelling will assume a central role in microbial ecology. As a result, it will transition from a mainly qualitative descriptive discipline to a quantitative predictive one. However, I think that ecosystem predictability will be noisy, more like predicting the stock market than gravitational orbits, and general principles will be hard fought and won. As Tom Curtis is likely to point out in these pages, we would do well to use macro-ecological theory as a guide in this endeavour. The wild card in the deck is synthetic biology. The idea of treating cells as chasses, genes as parts, and designing your own organism to specification is alien to most ecologists, particularly if the designer organism is assembled from parts that are partitioned as discrete functional units in a community. But I believe both fields stand to benefit enormously from one another. By reinventing life, synthetic biology will fail repeatedly, but in so doing will accelerate our understanding of metabolism and regulation, and improve ecological modelling efforts. And by observing nature in detail at the molecular level, microbial ecology will provide many design constraints to synthetic biologists that they will not have to uncover by trial and error. I thank Héctor García Martín, Victor Kunin, Gene Tyson and Trina McMahon for feedback on my naval gazing.