Massively parallel sequencing (MPSS) of bacterial 16S rDNA has been widely used to characterize the microbial makeup of the human and mouse gastrointestinal tract. However, techniques for fungal microbiota (mycobiota) profiling remain relatively under-developed. Compared to 16S profiling, the size and sequence context of the fungal Internal Transcribed Spacer 1 (ITS1), the most common target for mycobiota profiling, are highly variable. Using representative gastrointestinal tract fungi to build a known "mock" library, we examine how this sequence variability affects data quality derived from Illumina Miseq and Ion Torrent PGM sequencing pipelines. Also, while analysis of bacterial 16S profiles is facilitated by the presence of high-quality well-accepted databases of bacterial 16S sequences, such an accepted database has not yet emerged to facilitate fungal ITS sequence characterization, and we observe that redundant and inconsistent ITS1 sequence representation in publically available fungal reference databases affect quantitation and annotation of species in the gut. To address this problem, we have constructed a manually curated reference database optimized for annotation of gastrointestinal fungi. This targeted host-associated fungi (THF) database contains 1817 ITS1 sequences representing sequence diversity in genera previously identified in human and mouse gut. We observe that this database consistently outperforms three common ITS database alternatives on comprehensiveness, taxonomy assignment accuracy and computational efficiency in analyzing sequencing data from the mouse gastrointestinal tract.
Keywords: Amplicon sequencing; Database; Fungal ITS; Gut; Internal transcribed spacer; Microbiome.
Copyright © 2015 Elsevier B.V. All rights reserved.