*5.9. How Quantitative Is DNA Metabarcoding?*

Finally, there is continued debate over whether DNA metabarcoding may characterise pollen samples in a quantitative manner, with mixed results across studies [74,85,86,154,155]. Quantification has been found to be affected by a combination of marker and primer used, pollen type, mixture characteristics, and PCR conditions [88,156–158]. It is likely that relationships between the proportion of DNA reads and pollen counts are more likely for the most abundant taxa within a sample [83,159]. Similar to microscopy, rare taxa are difficult to detect using pollen metabarcoding [87]. Whilst this is a limitation, studies examining insect floral resource use often place greater focus on those plants detected at higher abundance. For this reason, along with the potential biases which can occur, DNA metabarcoding should be considered as semi-quantitative and relative read abundance used for downstream analysis [160]. We do not recommend the use of presence/absence approaches due to rare taxa being overstated and abundant taxa devalued [160].
