Premium
Clustering of fungal community internal transcribed spacer sequence data obscures taxonomic diversity
Author(s) -
Yamamoto Naomichi,
Bibby Kyle
Publication year - 2014
Publication title -
environmental microbiology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.954
H-Index - 188
eISSN - 1462-2920
pISSN - 1462-2912
DOI - 10.1111/1462-2920.12390
Subject(s) - biology , internal transcribed spacer , cluster analysis , in silico , taxonomic rank , sequence (biology) , dna sequencing , computational biology , evolutionary biology , identification (biology) , phylogenetic tree , genetics , ecology , dna , taxon , statistics , gene , mathematics
Summary Next‐generation DNA sequencing has enabled a rapid expansion in the size of molecular fungal ecology studies employing the nuclear internal transcribed spacer ( ITS ) region. Many sequence‐processing pipelines and protocols require sequence clustering to generate operational taxonomic units ( OTUs ) based on sequence similarity as a step to reduce total data quantity and complexity prior to taxonomic assignment. However, the consequences of ITS sequence clustering in regard to sample taxonomic coverage have not been carefully examined. Here we demonstrate that typically used clustering thresholds for fungal ITS sequences result in statistically significant losses in taxonomic coverage. Analyses using environmentally derived fungal sequences indicated an average of 3.1% of species went undetected ( P < 0.05) if the sequences were denoised and clustered at a 97% threshold prior to taxonomic assignment. Additionally, an in silico analysis using a reference fungal ITS database suggested that approximately 25% of species went undetected if the sequences were clustered prior to taxonomic assignment. Finally, analysis of sequences derived from pure‐cultured fungal isolates of known identity indicated sequence denoising and clustering were not critical in improving identification accuracy.