On the unreliability of published DNA sequences

Abstract

Summary• Here, the reliability of published fungal nucleic acid sequences is tested by the critical re-evaluation of 206 named sequences obtained from public-access databases.• Sequences from the ribosomal RNA (rRNA) gene cluster were examined as these are commonly used to establish fungal phylogeny and evolution, and are also increasingly employed in the identification of fungi from nonculture based studies.• Fifty-one rRNA internal transcribed spacer (ITS) sequences were obtained for species of Amanita , 55 ITS sequences were obtained for species of Phoma and 100 rRNA small subunit sequences were obtained from representative genera of the order Helotiales. In each case, the fungal group was selected partly on the basis of sequences deposited by three or more laboratories in order to avoid sample bias. The results suggest that up to 20% of the sequences available for each group may be unreliable, and this proportion is supported by additional informal observations.

DNA phylogeny reveals polyphyly ofPhomasectionPeyronellaeaand multiple taxonomic novelties

“…Phoma identification is problematic and gives rise to many misidentifications (Bridge et al 2003), but most strains studied here have been classified properly. Strains that could not be identified upon collection due to overlapping species characters now can be delimited and defined with molecular characterization tools.…”

A common new species ofInocybein the Pacific Northwest with a diagnostic PDAB reaction

“…Inaccurate taxonomic annotation of DNA sequences of Inocybe on GenBank.-Several studies have pointed to the alarming problem of taxonomically misidentified DNA sequences in public data repositories such as GenBank (Bridge et al 2003, Vilgalys 2003, Nilsson et al 2006, Bidartondo et al 2008. Indeed, about 20% of taxonomic entries are probably misidentified to the species level (Nilsson et al 2006).…”

Wood production, wood technology, and biotechnological impacts

“…Nonetheless, the NCBI GenBank or the EMBL-Bank, as public databases open to sequence submission by anyone, have been shown to contain wrongly assigned sequences 8Naumann et al 2007). The amount of mistakes is actually very significant and it was estimated that upwards of 20% of the named sequences may be misidentified in some fungi (Bridge et al 2003). To demonstrate this problem, we performed a BLAST search at the NCBI GenBank with the sequence of the ITS region of Trametes versicolor ATCC strain 32745.…”