Improving accuracy of DNA diet estimates using food tissue control materials and an evaluation of proxies for digestion bias

• Published in: Molecular Ecology
• Main Authors: Thomas, Austen C., Jarman, Simon N., Haman, Katherine H., Trites, Andrew W., Deagle, Bruce E.
• Other Authors: Pacific Salmon Foundation
• Format: Article in Journal/Newspaper
• Language: English
• Published: Wiley 2013
• Subjects: Phoca vitulina
• Online Access: http://dx.doi.org/10.1111/mec.12523; https://api.wiley.com/onlinelibrary/tdm/v1/articles/10.1111%2Fmec.12523; https://onlinelibrary.wiley.com/doi/pdf/10.1111/mec.12523

Abstract Ecologists are increasingly interested in quantifying consumer diets based on food DNA in dietary samples and high‐throughput sequencing of marker genes. It is tempting to assume that food DNA sequence proportions recovered from diet samples are representative of consumer's diet proportions, despite the fact that captive feeding studies do not support that assumption. Here, we examine the idea of sequencing control materials of known composition along with dietary samples in order to correct for technical biases introduced during amplicon sequencing and biological biases such as variable gene copy number. Using the Ion Torrent PGM © , we sequenced prey DNA amplified from scats of captive harbour seals ( Phoca vitulina ) fed a constant diet including three fish species in known proportions. Alongside, we sequenced a prey tissue mix matching the seals’ diet to generate tissue correction factors ( TCF s). TCF s improved the diet estimates (based on sequence proportions) for all species and reduced the average estimate error from 28 ± 15% (uncorrected) to 14 ± 9% ( TCF ‐corrected). The experimental design also allowed us to infer the magnitude of prey‐specific digestion biases and calculate digestion correction factors ( DCF s). The DCF s were compared with possible proxies for differential digestion (e.g. fish protein%, fish lipid%) revealing a strong relationship between the DCF s and percent lipid of the fish prey, suggesting prey‐specific corrections based on lipid content would produce accurate diet estimates in this study system. These findings demonstrate the value of parallel sequencing of food tissue mixtures in diet studies and offer new directions for future research in quantitative DNA diet analysis.