DNA metabarcoding diet analysis in ruminants is quantitative and integrates feeding over several weeks

Dietary DNA metabarcoding is an established method, especially useful for resolving the diverse diets of large mammalian herbivores (LMH). However, despite longstanding research interest on the topic, we still lack unequivocal evidence on the potential of DNA metabarcoding to reflect proportions of ingested dietary plants in LMH. One major aspect to consider is the time window during which ingested diet remains detectable in faecal samples. This parameter is currently unknown for LMH, thus potentially hindering the scope of ecological conclusions. Another unknown factor is quantitative performance, i.e. the extent to which the amount of ingested biomass can be assessed based on sequence read abundances. We assessed DNA metabarcoding quantitative performance and DNA half-life detectability for plants with different digestibilities in a controlled feeding experiment with three female Eurasian tundra reindeer (Rangifer tarandus tarandus). Reindeer were fed birch twigs (Betula pubescens) and increasing biomass of lichen (mainly Cladonia stellaris). Relative read abundance positively correlated with ingested lichen biomass, suggesting potential for deriving dietary proportions in free-ranging reindeer on natural pastures. Dietary DNA was consistently detected within a few hours upon ingestion, with a mean half-life detectability of 30 and 14 hours for birch and lichen, respectively. However, dietary DNA remained detectable in faeces for at least 26 days post-feeding, indicating that a single faecal sample can provide an unsuspectedly integrative estimate of diet in ruminants. Together, our findings provide novel empirical validation of DNA metabarcoding as a tool for diet analysis in LMH.