One adult Erebia cassioides female was collected by Kay Lucek on the 10 August 2022 by hand-netting near Grindelwald, Switzerland (46.654582°N, 8.025222°E). A female, the heterogametic sex in Lepidoptera, was sampled to also assemble the W chromosome. Samples were kept alive until flash-freezing and storage at -80°C. Species identification was conducted visually in the field.

The wings of the specimen were deposited at the Muséum d'Histoire Naturelle de Neuchâtel, Neuchâtel, Switzerland (voucher ID MHNN-65-8969). Wing pictures of another female from the same population are presented in Fig. 1.

The genome sequence of Erebia cassioides was produced under the European Reference Genome Atlas Switzerland (ERGA-CH) framework. The umbrella BioProject for this species is PRJEB98373. The primary genome assembly is available on the European Nucleotide Atlas (ENA) under accession number GCA_976986335 and the alternate assembly under accession GCA_976984905. The metadata and associated raw sequencing reads for the samples used for genome assembly and scaffolding (assigned Tree of Life ID ilEreCass2) are available under BioSample accession SAMEA120234452. The code used to produce the assembly is available on GitHub (https://github.com/camille-cornet/ErebiaGenomeAssembly) and archived on Zenodo (Cornet and Lucek 2025).

The genome of Erebia cassioides is diploid with 10 chromosome pairs as estimated by karyotyping studies (2n = 20; De Lesse (1960), Robinson (1971)). Before sequencing, genome size estimates were approximately 500 Mb, based on two pre-existing chromosome-level Erebia assemblies (K. Lohse et al. 2022, O. Lohse et al. 2022) and a contig level assembly of another E. cassioides female of the same population (Augustijnen and Lucek 2024). The sex chromosome system was expected to be ZZ/ZW (female heterogamy) as for most butterflies (Traut et al. 2007).

For genome assembly, high-molecular-weight (HMW) DNA was extracted from half a thorax using the Qiagen MagAttract HMW DNA Kit (Qiagen, Hombrechtikon, Switzerland), following the manufacturer’s instructions.

For assembly scaffolding, Hi-C data was generated using the head of the same individual that was used for PacBio sequencing. After grinding in liquid nitrogen, prior to library preparation, crosslinking was performed using the Arima High Coverage HiC kit (Arima Genomics, San Diego, CA, USA), following the manufacturer’s instructions.

For genome assembly, PacBio sequencing was performed by the Genomics Technologies Facility (GTF, Lausanne, Switzerland) on the PacBio Sequel IIe, generating 1.63 million HiFi reads or 16.4 Gb of data. Average HiFi read length was 10 kb.

For assembly scaffolding, Hi-C library preparation following crosslinking was performed using the High Coverage HiC kit (Arima Genomics, San Diego, CA, USA), following the manufacturer’s instructions. Sequencing on an Illumina NovaSeq 6000 (300 cycles, 150 bp long paired-end reads) was performed by the Next Generation Sequencing Platform (NGS Platform, Bern, Switzerland), resulting in 155 million read pairs.

After quality control of the HiFi reads using NanoPlot v.1.32.1 (De Coster and Rademakers 2023), reads were assembled into contigs using hifiasm v.0.16.0-r369 (Cheng et al. 2021) in primary assembly mode, with the settings -D 10 -N 200 to improve assembly of repetitive regions. Haplotypic duplication was then removed from the primary assembly and an alternate assembly was produced, using purgedups v.1.2.5 (Guan et al. 2020) with default settings.

For scaffolding of the draft primary assembly into chromosomes, the quality of the Hi-C reads was checked using FastQC v.0.12.1 (https://github.com/s-andrews/FastQC) and five bases were trimmed off the 5’ end of each read to reach a better mapping rate. Poly-G tails were trimmed while also applying a minimal length filter of 120 bp and a PHRED quality filter of 30 using fastp v.0.23.4 (Chen 2023). The reads were mapped to the primary assembly using the Arima mapping pipeline v.03 (https://github.com/ArimaGenomics/mapping_pipeline) with a minimum mapping quality of 1. The assembly was scaffolded using YaHS v.1.2 (Zhou et al. 2023) with default settings and manual curation was performed using Juicebox v.1.11.08 (Durand et al. 2016). The scaffolded assembly was then decontaminated using BlobToolKit v.4.3 (Challis et al. 2020) and tiara v.1.0.3 (Karlicki et al. 2022) to remove any bacterial or mitochondrial scaffold. The mitogenome was assembled from the raw HiFi reads using MitoHiFi v.3.0.0 (Uliano-Silva et al. 2023). The HiFi reads were then mapped back to the scaffolded assembly to close gaps with tgsgapcloser v.1.1.1 (Xu et al. 2020). Scaffolds corresponding to haplotypic duplication or unplaced repetitive regions were then manually removed. The quality of the final primary assembly was assessed to confirm that it reaches the Earth Biogenome Project (EBP) standards (Rhie et al. 2021), using BUSCO v.5.7.1 (Manni et al. 2021) for functional completeness with the Lepidoptera OrthoDB v.10 dataset (Waterhouse et al. 2013), gfastats v.1.3.6 (Formenti et al. 2022) for contiguity and Merqury v.1.3 (Rhie et al. 2020) for k-mer completeness, base pair quality and false duplication rates.

The Z and W sex chromosomes were identified based on the coverage of the HiFi reads (generated from a female individual) and Illumina reads from a male individual from the same population (Augustijnen et al. 2024) mapped back to the primary assembly using minimap2 v.2.21 (Li 2018) and bwa mem v.0.7.17 (Li 2013), respectively, followed by samtools coverage v.1.14 (Danecek et al. 2021) with a maximum depth of 100 and a minimum read and mapping quality of 30. The correspondance between the chromosomes of E. cassiodes and Merian elements, i.e. ancestral linkage groups in Lepidoptera (Wright et al. 2024), was assessed using the lep_busco_painter tool v.1.0.0 (https://github.com/charlottewright/lep_busco_painter), based on the BUSCO output.

The primary assembly of Erebia cassioides has a total length of 546,119,514 bp, distributed in nine autosomes, one Z chromosome, one W chromosome and the mitogenome (Table 1, Figs 2, 3). The GC content is 36.8%. The draft assembly, comprising 23 contigs, had a contig N50 of 45.1 Mb and L50 of 5. The primary assembly has a scaffold N50 of 55.7 Mb, L50 of five and twelve gaps (total gap length: 1,200 bp). The gene content completeness analysis of the assembly resulted in a BUSCO completeness score of 98.8% (98.2% single-copy and 0.6% duplicated, Fig. 2). The alternate assembly has a total length of 405,746,844 bp distributed in 2046 scaffolds, with a contig N50 of 515 kb. Its BUSCO completeness score is 79.2% (78.4% single-copy and 0.8% duplicated). Combining the primary and alternate assemblies, the k-mer completeness score is 97.9%, false duplication rate 1.7% and the assembly has a base quality value of 60.1 (Fig. 4). The chromosome-level scaffolds, confirmed by Hi-C data (Fig. 3), are named according to size (Table 1). Chromosome painting with Merian elements illustrates the distribution of orthologues along chromosomes and highlights patterns of chromosomal evolution relative to Lepidopteran ancestral linkage groups (Table 1, Fig. 5).