Bushmeat genetics: setting up a reference framework for the DNA typing of African forest bushmeat

The bushmeat trade in tropical Africa represents illegal, unsustainable off-takes of millions of tons of wild game - mostly mammals - per year. We sequenced four mitochondrial gene fragments (cyt b, COI, 12S, 16S) in >300 bushmeat items representing nine mammalian orders and 59 morphological species from five western and central African countries (Guinea, Ghana, Nigeria, Cameroon and Equatorial Guinea). Our objectives were to assess the efficiency of cross-species PCR amplification and to evaluate the usefulness of our multilocus approach for reliable bushmeat species identification. We provide a straightforward amplification protocol using a single 'universal' primer pair per gene that generally yielded >90% PCR success rates across orders and was robust to different types of meat preprocessing and DNA extraction protocols. For taxonomic identification, we set up a decision pipeline combining similarity- and tree-based approaches with an assessment of taxonomic expertise and coverage of the GENBANK database. Our multilocus approach permitted us to: (i) adjust for existing taxonomic gaps in GENBANK databases, (ii) assign to the species level 67% of the morphological species hypotheses and (iii) successfully identify samples with uncertain taxonomic attribution (preprocessed carcasses and cryptic lineages). High levels of genetic polymorphism across genes and taxa, together with the excellent resolution observed among species-level clusters (neighbour-joining trees and Klee diagrams) advocate the usefulness of our markers for bushmeat DNA typing. We formalize our DNA typing decision pipeline through an expert-curated query database - DNA BUSHMEAT - that shall permit the automated identification of African forest bushmeat items.