Promises and pitfalls of using high-throughput sequencing for diet analysis

The application of high-throughput sequencing-based approaches to DNA extracted from environmental samples such as gut contents and faeces has become a popular tool for studying dietary habits of animals. Due to the high resolution and prey detection capacity they provide, both metabarcoding and shotgun sequencing are increasingly used to address ecological questions grounded in dietary relationships. Despite their great promise in this context, recent research has unveiled how a wealth of biological (related to the study system) and technical (related to the methodology) factors can distort the signal of taxonomic composition and diversity. Here, we review these studies in the light of high-throughput sequencing-based assessment of trophic interactions. We address how the study design can account for distortion factors, and how acknowledging limitations and biases inherent to sequencing-based diet analyses are essential for obtaining reliable results, thus drawing appropriate conclusions. Furthermore, we suggest strategies to minimize the effect of distortion factors, measures to increase reproducibility, replicability and comparability of studies, and options to scale up DNA sequencing-based diet analyses. In doing so, we aim to aid end-users in designing reliable diet studies by informing them about the complexity and limitations of DNA sequencing-based diet analyses, and encourage researchers to create and improve tools that will eventually drive this field to its maturity.