The identification of such extensive contamination of human sequence across databases and sequence types warrants caution among the sequencing community in future sequencing efforts, such as human re-sequencing.
http://plosone.org/article/info:doi/10.1371/journal.pone.0016410#article1.front1.article-meta1.abstract1.p1

NCBI has responded to this article at http://www.ncbi.nlm.nih.gov/About/news/18feb2011.html.

NCBI writes: "While we understand that there is some contamination in sequences submitted to the public databases, we do not believe that the problem is as widespread as reported in the paper. Most of the examples provided by the authors were of preliminary data submitted to the archives under rapid data release policies. Database users must understand that preliminary data are just that, preliminary, because they have not been screened and vetted the way finished data have."