Replication data for : DNA-based monitoring of bushmeat trade dynamics and hotspots across western and central Africa

We build on a four-gene mitochondrial DNA-typing approach to develop an unprecedented reference database framework for effectively tracing the bushmeat trade in tropical Africa. The dataset includes 8,673 mitochondrial sequences generated from 2,516 bushmeat samples collected over a 13-year period (2008–2020) across 10 tropical African countries and from two airport seizures in Europe. The dataset spans 19 vertebrate taxonomic orders and 171 putative species or higher-level taxa representative of the bushmeat trade, including primarily mammals (N = 153), as well as birds (N = 6) and reptiles (N = 11).

Samples underwent DNA extraction, PCR amplification and Sanger sequencing for four mitochondrial gene fragments following the DNA-typing protocol of Gaubert et al. (2015). The four markers encompass Cytochrome c oxidase I (COI; 658 bp), Cytochrome b (Cytb; 402 bp), and 12S (167 - 404 bp) and 16S (210 - 528 bp) ribosomal DNA.

Sequence quality and taxonomic assignment procedures were conducted on the newly produced dataset of sequences (N=4422). For COI and Cytb, we checked for the presence of potential pseudogenes (or NUMTS; Lopez et al., 1994), by screening for the disruption of their reading frame in Geneious prime v. 2023.3.1 (Kearse et al., 2012). For 12S and 16S, due to the absence of a reading frame, potential pseudogenes were screened for through long-branch and aberrant phylogenetic branching patterns (Triant & DeWoody, 2007). The sequences representative of 12S and 16S were first aligned with Muscle (Edgar, 2004) in Mega v.11 (Tamura et al., 2021) under default settings. Second, we used Gblock (phylogeny.lirmm.fr/phylo_cgi/one_task.cgi?task_type=gblocks), with both relaxed and default settings depending on the levels of divergence within orders, to remove poorly aligned positions, resulting in final alignments of 288-331 bp for 12S and 322-485 bp for 16S. Neighbour Joining (NJ) trees (Saitou & Nei, 1987) were generated with MEGA under the K2P (Kimura, 1980) model and the option “pairwise deletion”. Node robustness was assessed through 500 bootstrap (Felsenstein, 1985) replicates. The same NJ tree approach was used for COI and Cytb as a secondary assessment of sequence quality, as pseudogenes do not necessarily translate into reading frame disruption.

The data is organized as four distinct FASTA alignments of nucleotide sequences, representing different mitochondrial DNA markers: Cytochrome c oxidase I (COI; 658 bp), Cytochrome b (Cytb; 402 bp), and 12S (167 - 404 bp) and 16S (210 - 528 bp). Together with associated metadata, the FASTA files can be downloaded for local BLAST analyses.