Artificial and natural duplicates in pyrosequencing reads of metagenomic data

Our method is available from http://cd-hit.org as a downloadable program and a web server. It is important not only to identify the duplicates from metagenomic datasets but also to distinguish whether they are artificial or natural duplicates. We provide a tool to estimate the number of natural duplicates according to user-defined sample types, so users can decide whether to retain or remove duplicates in their projects.