Genome assembly

Two rounds of read-merging were performed on the adapter-trimmed reads. The first round consisted of cascading Konnector runs (Vandervalk et al. 2015), where the 10 × Genomics and Illumina HiSeq reads were merged using k values ranging from 115–75 and 235–75, respectively, both with a step size of −10. Reads that were unable to be merged in the first round were subjected to a second round of merging with abyss-mergepairs (Jackman et al. 2017). The longer pseudo-reads and any remaining unmerged reads were then assembled using ABySS v2.2.5 (k = 96, 112, 128, 144, 160; kc = 3, 4). The best assemblies, as assessed by abyss-fac (Jackman et al. 2017), were passed to ntJoin v1.0.3 (Coombe et al. 2020) to perform iterative assembly-guided scaffolding runs, each with the following parameters: no_cut = True; k = 32; w = 250; reference_weights =’2’. Following scaffolding, a round of misassembly correction was performed on the resulting assembly using Tigmint v1.1.2 with span = 2. In addition to the post-ntJoin assembly, 10 × Genomics Chromium reads were also passed as input. This was followed by another round of scaffolding using ARCS v1.1.1 and LINKS v1.8.6 (c = 3; l = 3; a = 0.9; z = 3000; s = 90). Introduced gaps were filled with Sealer v2.2.3 (Paulino et al. 2015) (L = 150; P = 10 l; k = 75,85,95,105,115), yielding the final genome assembly.