Using Individual Scale de novo Assembly to Identify Mutations Causing a Phenotypic Trait

The Mediterranean fruit fly (medfly) is an important agricultural pest of many fruit and vegetable species. To protect the mainland United States from this pest, the sterile insect technique (SIT) is employed, involving release of tens of millions of sterile male medfly into the Los Angeles basin of California weekly. These flies have several mutations making them amenable to mass release. Using a chromosomal translocation between the 5th chromosome and the male Y chromosome, females are homozygous recessive for both a temperature sensitive lethal mutation and a white pupal mutation, allowing straightforward separation of male and female flies and generation of male only release strains. Males are maintained heterozygous for these alleles through the chromosomal translocation, linking wild-type phenotype with the sex chromosome. While the relative position of these mutations is known (5th chromosome), the genes and specific mutation causing the traits are not known. To address this, we developed a crossing scheme to isolate these mutations from the SIT line in the background of an inbred lab line (see Sim et al submission). This cross allowed QTL analysis to identify regions of the genome that are strongly linked to these traits. To identify causative mutations, we performed whole genome sequencing of individual F4 flies resulting from this cross including individuals that had the observed mutated phenotypes and those that did not. Utilizing 2 X 250 bp paired-end sequencing on Illumina HiSeq 2500 and subsequent assembly and analysis utilizing DISCOVAR/DISCOVAR-denovo, we were able to generate individual scale assemblies and compare these assembly graph structures surrounding the QTL locations to identify specific loci and mutations in the genome that are consistent with the phenotypes observed in the individuals sequenced. This list represents potential causative mutations for the traits in this SIT line, and currently confirmation of the causative mutations is be determined using targeted gene editing approaches with CRISPR/CAS9, to recreate these mutations in wild-type lab lines. This experiment demonstrates the utility of comparing individual-scale genomic assemblies in non-model organisms to reveal direct structural variations between these assemblies, in contrast to relying on reference based mapping approaches (e.g. GATK) to identify variants between individuals.