Assembling de novo Aedes genomes from single specimens

Aedes mosquitos are important arbovirus vectors, and as such, much time and energy has been spent on developing resources elucidating their biology. The genomic resources for two species in this genus are among the best in the world, allowing for novel control methods to be developed and interesting biology to be elucidated. However, these genomic resources are composed of pools of individuals to meet DNA input requirements. Recent advances in sequencing technology by Pacific Biosciences obviate these requirements, as sequencing libraries can be constructed from single wild-caught mosquitoes, yielding large quantities of highly accurate, circular-consensus long reads (HiFi). In response, numerous new genome assemblers have been written specifically to handle these new datasets but share a common issue in that they artefactually duplicate heterozygous regions of the genome. To mitigate this issue, other programs have been written to purge artefactually duplicated sequences. Here, we evaluate the performance of four HiFi assemblers—HiCanu, HifiASM, flye, and IPA—using HiFi reads from Ae. aegypti aegypti and Ae. aegypti formosus. We then evaluate the performance of two purging programs—PurgeDups and PurgeHaplotigs. We evaluate the performance of these programs using N50, assembly size, and BUSCO scores, and compare them to the Ae. Aegypti L5 reference genome. Our findings show no single combination of assembler are purging program outperforms other combinations on all metrics, but numerous combinations produced assemblies comparable to the reference.