RandomizationsX1500

To estimate the variability within a population from the available metagenomic data, random candidatus genomes (CG) were generated as follows using a series of custom perl scripts. First, we recruited reads to each CG requiring at least 95% identity and a coverage of 95% of the entire length of the read. Each read was non-redundantly assigned and aligned to a CG using default parameters in MUSCLE. For each CG population, we generated 100 random CG sequences using the metagenomic data that were recruited to consensus sequences, with each base having a probability of being assigned from its relative abundance in the underlying metagenomic sequence data. Here we show the result of 1500 randomizations.