Orchestia grillus isolate:DR04-Leg-1-Zymo_S33 Genome sequencing and assembly. Orchestia grillus isolate:DR04-Leg-1-Zymo_S33

The O. grillus genome was assembled de novo using short reads and a de Bruijn graph approach implemented by Ye, et al. (2012). Reads were quality controlled (QC) prior to assembly using the QC tools implemented in SGA (complexity and homopolymer filter; Simpson and Durbin 2010) and BBmap (Bushnell 2016). We assessed the impact of our filtering using the preqc function of SGA. Optimal k-mer size for graph construction was determined by iterative trials at various k values. Best results were obtained at k=51. As both invertebrates and members of a natural population, O. grillus individuals are expected to harbor high levels of genetic variation. These high levels of variation increase the complexity of the assembly and may result in many heterozygous regions of the genome to be assembled as redundant contigs (i.e., heterozygote haplotypes assembled as different contigs). We controlled for heterozygotic contigs using the algorithm implemented by Pryszcz and Gabaldon (2016). This algorithm uses information contained in the contigs and the pair-end reads to collapse heterozygous contigs into homozygous scaffolds.