Laboratory protocol is important to improve the correlation between target copies and metabarcoding read numbers of seed DNA in ground beetle regurgitates

DNA metabarcoding is increasingly important for studying feeding interactions, yet it remains unresolved whether reporting read counts or occurrences is to be preferred. To address this issue for gut content samples, basic experimental data on the relationship between read numbers and initial prey DNA amounts and how both change over digestion time is needed. Using regurgitates of the carabid Pseudoophonus rufipes the digestion of Taraxacum officinale seeds was documented for 128 h post-feeding to determine how the number of (1) seed DNA copies and (2) metabarcoding reads change over digestion time, and (3) how they correlate to each other. Additionally, we tested (4) whether PCR cycle-numbers during library preparation affect this correlation. The number of copies and reads both decreased with digestion time, but variation between samples was high. Read and copy numbers correlated when using a library preparation protocol with 35 cycles (R(2) = 42.0%), yet a reduction to 30 cycles might significantly improve this correlation, as indicated by additional PCR testing. Our findings show that protocol optimization is important to reduce technical distortions of read numbers in metabarcoding analysis. However, high inter-sample variation, likely due to variable digestive efficiency of individual consumers, can blur the relationship between the amount of food consumed and metabarcoding read numbers.