In silico DNA barcoding surpasses whole genome sequencing for species identification from vector surveillance pools

Mosquito-borne diseases are responsible for over 600,000 deaths annually, mainly in sub-Saharan countries. Mosquito surveillance is a crucial element of vector control programmes to assure they remain effective. This study focused on optimizing the use of the MinION sequencer for interrogating mosquito pools from surveillance programmes, with simplified bioinformatic workflows for use in African laboratories nearer to the field. Mosquito pools were created using different human disease vector species, mimicking possible field trap contents. Some pools were spiked with DNA from the malaria parasite Plasmodium falciparum and filarial worm Brugia malayi. In the first instance, three pipelines were used to map reads to full reference genomes and their accuracy was compared. Subsequently, mapping reads to full assembled genomes was compared to mapping to a concatenation of diagnostic barcoding sequences for evaluation of relative abundances of mosquito vector species and pathogens. The results show that a combination of Minimap2 with samtools is the most accurate pipeline and a targeted approach preferable to whole genome in estimating species abundance. The MinION device was shown to be effective for interrogating mosquito pools, with moderate training requirements for data analysis. This provides a practical solution to vector surveillance challenges in sub-Saharan Africa.