Just found this article on 454 pyrosequencing - haven't read it yet.

If Eid et al. are able to deliver what they promise in this article, in five years we will be able to rapidly sequence ~100,000 fosmid sized chunks of DNA from a microbiome. A lot of the frustrating limitations that exist in dealing with the short 50 - 400 bp reads from 454 pyrosequencing - i.e. inability to capture the neighborhood that a gene fragment is in, difficulty in identifying mobile, rapidly evolving elements in a metagenome (i.e. CRISPRs), difficulty in assigning structural and functional predictions to protein gene fragments, will be removed.

I'm going to start posting papers that I've found to be helpful in getting a handle on looking at the Metagenomic datasets that we had sequenced from the 2nd and 3rd round of Magneto-FISH bead-capture experiments.