Synteny and rearrangement detection

To assess variation in chromosome-scale synteny, we compared the assemblies of each Morpho to the assembly of M. jurtina, the closest relative of Morpho with a karyotype of 29 chromosomes and for which a high-quality chromosome-level assembly (based on N50 values and BUSCO score, accession ID GCF_905333055.1) is available [29]. We used MUMmer 3.23 [41] to align the masked assembled genomes of M. helenor, M. achilles, and M. deidamia to the M. jurtina genome. The output produced by MUMmer is an ASCII delta file that was then filtered and parsed using the utility programs delta-filter and show-coords from MUMmer. Synteny was visualized with the MUMmer results in R with the packages circlize v 0.4.12 [42] and Paletteer (https://github.com/EmilHvitfeldt/paletteer) using the Rscript from [43] described here: https://github.com/bioinfowheat/Polygonia_calbum_genomics/blob/7c75aac624157faa3ab229e3fc1e0e315302194d/synteny/circlePlot_nucmerOutput.R, removing short contigs, short alignments (less than 200 bp), and low-identity alignments (less than 90% identity).

In order to detect potential genome rearrangements between Morpho and closely related species, we estimated the whole-genome collinearity between the Morpho assemblies and 5 closely related Nymphalidae species whose genomes exhibit good-quality assemblies in the NCBI genome database: M. jurtina (GCA_905333055.1), P. aegeria (GCA_905333055.1), Erebia ligea (GCA_923060345.2), Melanargia galathea (GCA_920104075.1), and Lasiommata megera (GCA_928268935.1) using D-GENIES [44]. Paired alignments between a Morpho species and 1 Nymphalidae species were performed using the minimap2 aligner [45] in D-GENIES, treating each Morpho species genome as the query and the Nymphalidae species genome as the target reference. We also used D-GENIES to pair-compare the genomes of the 3 Morpho species. As D-GENIES revealed differences between Morpho species in the contig corresponding to the Z chromosome (see Results), we used SyRI [46] to study in detail the rearrangements in the sequences of this contig between the 3 species. We generated paired alignments of the Z contig with minimap2 and ran SyRI with the option -c on.sam files. SyRI requires that the 2 compared genomes represent the same strand, and in the case of M. achilles, the orientation of the sequence produced by Hifiasm was complementary to the sequences of M. helenor and M. deidamia. We then reverse-complemented this sequence in order to make the alignments. All the genomic structures predicted by SyRI were plotted using plotsr [47].