Shotgun metagenomics

Shotgun metagenomic sequencing was conducted following the procedure described in the Illumina TrueSeq DNA sample preparation protocol. The triplicated samples for each successional stage collected in July 2012 were subjected to shotgun metagenomic profiling (n = 15). Seasonal variations in the abundance of marker genes were later interrogated by primer-specific quantitative PCR assays (see below). In brief, aliquots of each DNA sample were mechanically sheared before entering the Illumina library generation protocol. Libraries were size-selected to 170–180 bp using an agarose gel. Sequencing was carried out in a paired-end (2 × 100 bp) Illumina HiSeq2000 run at the Argonne National Laboratory in the Next Generation Sequencing Core (NGS). Raw, unassembled Illumina reads were paired, dereplicated and quality filtered in MG-RAST (Meyer et al., 2008). Putative open reading frames on the quality-controlled sequences were called using FragGeneScan (Rho et al., 2010). Metagenomes were functionally annotated using BLASTX searches against the KEGG (Kyoto Encyclopedia of Genes and Genomes) Orthology (KO) identifiers (Kanehisa et al., 2008). For the taxonomic assignments of selected KOs, sequences of specific KOs were retrieved and annotated against the M5nr (an MD5 nonredundant database) (Wilke et al., 2012) using the best-hit organismal classification method. Functional and taxonomic annotations of sequences were carried out with a maximum e-value cutoff of 10⁻⁵, a minimum percent identity cutoff of 60% and a minimum alignment length cutoff of 15. All sequence data have been deposited in the MG-RAST database. The reference IDs of the metagenomes are provided in Supplementary Table 2.