Polymerase and primer biases in eukaryote ultra-long-read metabarcoding

Abstract

Ultra-long-read metabarcoding is a recently developed method for biodiversity monitoring that enables both high-accuracy species-level identification and reliable phylogenetic placement of taxa. Ultra-long-read analyses require high-processivity DNA polymerases for amplicons and primer pairs that fully cover the targeted taxonomic groups. So far, there is no published information on the relative suitability of commercial high-processivity polymerase kits and on the relative performance of primer pairs targeting a DNA fragment spanning much of the 18S, ITS and 28S rRNA gene regions in eukaryotes. Using PacBio sequencing, we assessed the performance of five commercial high-processivity DNA polymerases and three ultra-long marker primer pairs relative to a commonly used full-ITS primer pair on producing artefacts and recovering biodiversity in two mock communities and 10 soil samples. Our analyses revealed great differences in both polymerases and primer pairs in data quality and biodiversity. For testing evolutionary hypotheses, we recommend using the Solis or Fusion polymerases and the Euk575F-21Rngs primer pair in ultra-long-read metabarcoding analyses of soil eukaryotes. Compared with a previous study using the same DNA samples, the ultra-long amplicons revealed substantially lower biodiversity than medium-sized, full-ITS amplicons across taxonomic levels, suggesting that medium-sized amplicons may be better suited for routine soil eukaryote biodiversity monitoring than ultra-long amplicons.