Quantitative prey species detection in predator guts across multiple trophic levels by mapping unassembled shotgun reads.

Quantifying species trophic interaction strengths is crucial for understanding community dynamics and has significant implications for pest management and species conservation. DNA-based methods to identify species interactions have revolutionized these efforts, but a significant limitation is the poor ability to quantify the strength of trophic interactions, that is the biomass or number of prey consumed. We present an improved pipeline, called Lazaro, to map unassembled shotgun reads to a comprehensive arthropod mitogenome database and show that the number of prey reads detected is quantitatively predicted from the prey biomass consumed, even for indirect predation. Two feeding bioassays were performed: starved coccinellid larvae consuming different numbers of aphids (Prey Quantity bioassay), and starved coccinellid larvae consuming a chrysopid larvae that had consumed aphids (Direct and Indirect Predation bioassay). Prey taxonomic assignment against a mitochondrial genome database had high accuracy (99.8% positive predictive value) and the number of prey reads was directly related to the number of prey consumed and inversely related to the elapsed time since consumption with high significance (r  = .932, p = 4.92E-6). Aphids were detected up to 6 h after direct predation plus 3 h after indirect predation (9 h in total) and detection was related to the predator-specific decay rates. Lazaro enabled quantitative predictions of prey consumption across multiple trophic levels with high taxonomic resolution while eliminating all false positives, except for a few confirmed contaminants, and may be valuable for characterizing prey consumed by field-sampled predators. Moreover, Lazaro is readily applicable for species diversity determination from any degraded environmental DNA.