Discussion
This study underscores the importance of preliminary studies to assess primer combinations and database choice for the mycobiome sample of interest and raises questions regarding the validity of fungal abundance estimates.
Methods
Here, we compared the accuracy of taxonomic identifications and abundances from mycobiome analyses which vary among three commonly selected target gene regions (18S, ITS1, or ITS2) and the reference database (UNITE - ITS1, ITS2 and SILVA - 18S). We analyze multiple communities including individual fungal isolates, a mixed mock community created from five common fungal isolates found in weanling piglet feces, a purchased commercial fungal mock community, and piglet fecal samples. In addition, we calculated gene copy numbers for the 18S, ITS1, and ITS2 regions of each of the five isolates from the piglet fecal mock community to determine whether copy number affects abundance estimates. Finally, we determined the abundance of taxa from several iterations of our in-house fecal community to assess the effects of community composition on taxon abundance.
Results
Overall, no marker-database combination consistently outperformed the others. Internal transcribed space markers were slightly superior to 18S in the identification of species in tested communities, but Lichtheimia corymbifera, a common member of piglet gut communities, was not amplified by ITS1 and ITS2 primers. Thus, ITS based abundance estimates of taxa in piglet mock communities were skewed while 18S marker profiles were more accurate. Kazachstania slooffiae displayed the most stable copy numbers (83-85) while L. corymbifera displayed significant variability (90-144) across gene regions.