Noise and the accurate determination of microbial diversity from 454 pyrosequencing data

To investigate the effect of pyrosequencing noise on OTU diversity estimates we pyrosequenced two DNA samples comprising mixtures of known 16S rRNA V5/V6 bacterial sequences. The first sample contained 23 sequences at least 7% divergent to enable us to calibrate noise on the GSFLX machine. The second contained 90 sequences some very similar at very different concentrations. This sample was used to test different OTU construction algorithms. Including a new algorithm we have developed for removing noise from pyrosequenced amplicons. Finally we pyrosequenced an environmental DNA sample of 16S V5/V6 rRNA sequences from Priest Pot lake. http://people.civil.gla.ac.uk/~quince/PyroNoise.html