Shi, Y., Dick, C.M., Karpan, K., Baetscher, D., Henderson, M.J., Sethi, S.A., McPhee, M.V., Larson, W.A. Towards absolute abundance for conservation applications: estimating the number of contributors via microhaplotype genotyping of mixed-DNA samples. submitted to Molecular Ecology Resources.
Abstract
Molecular methods including metabarcoding and qPCR have shown promises for estimating species abundance by quantifying the concentration of genetic material in field samples. However, the relationship between specimen abundance and detectable concentrations of genetic material is often variable in practice. DNA mixture analysis represents an alternative approach to quantify specimen abundance based on the identity of unique alleles in a sample. The DNA mixture approach provides novel opportunities to inform ecology and conservation by estimating the absolute abundance of target taxa through molecular methods; yet, challenges with genotyping many highly variable markers in mixed-DNA samples have prevented its widespread use. To advance molecular approaches for abundance estimation we explored the utility of microhaplotypes for DNA mixture analysis by applying a 125-marker panel to 1,179 Chinook salmon (Oncorhynchus tshawytscha) smolts from the Sacramento-San Joaquin Delta. We assessed the accuracy of DNA mixture analysis through a combination of mock mixtures containing DNA from up to 20 smolts and a trophic ecological application enumerating smolts in predator diets. Mock DNA mixtures of up to 10 smolts could reliably be resolved using microhaplotypes and increasing the panel size would likely facilitate identification of more individuals. However, while analysis of predator gastrointestinal tract contents indicated DNA mixture analysis could discern the presence of multiple prey items, poor DNA quality prevented accurate genotyping and abundance estimation. Our results indicate that DNA mixture analysis can perform well with high-quality DNA, but methodological improvements in genotyping degraded DNA are necessary before this approach can be used on marginal quality samples.