{"paper_id":"08d3d35e-b17a-489e-a050-477313ee2245","body_text":"A haplotype-resolved \nbluethroat (Luscinia s. \nsvecica) genome assembly \nuncovers the complex MHC \nregion \n \nAuthors:  \nMarius A. Strand  1, Emily L.G. Enevoldsen  1, Ole K. Tørresen 1, Morten Skage 1, Giada Ferrari \n1, Ave Tooming-Klunderud 1, Erica H. Leder  \n2,3, Jan T. Lifjeld 2, Arild Johnsen  2 , Kjetill S. \nJakobsen 1* \nAddresses: \n1) Dept of Biosciences, Centre for Ecological and Evolutionary Synthesis (CEES), University of \nOslo, Oslo, Norway  \n2) Natural History Museum, University of Oslo, P .O. Box 1172 Blindern, 0318 Oslo, Norway \n3) Department of Marine Sciences, University of Gothenburg, Gothenburg, Sweden \n*corresponding author \nAbstract \nWe describe a chromosome-level, haplotype-resolved genome assembly from a female \nbluethroat (Luscinia s. svecica). The assembly comprises two pseudo-haplotypes of 1461 Mb \nand 1171 Mb, with 77.4% and 88.4% scaffolded into 40 autosomal chromosomes and the W \nand Z sex chromosomes (haplotype one). Assembly completeness is high (BUSCO 99.2% and \n94.9%), with 22,462 and 18,769 annotated protein-coding genes for haplotypes one and \ntwo, respectively. The use of Oxford Nanopore Technologies sequencing enables resolution \nof genomic regions that are often fragmented in genome assemblies, including the \nhypervariable Major Histocompatibility Complex (MHC). We find that MHC loci include both \nthe canonical organization of tandemly duplicated MHCIIβ genes with a single MHCIIA, and a \ndistinct arrangement in which MHCI and MHCIIβ loci are interspersed in intermixed arrays, \nand that substantial structural differences between haplotypes are directly resolved in the \nassembly. \n \nKeywords: Earth Biogenome Project Norway, MHCI, MHCII, Long-read, Oxford Nanopore\n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 30, 2026. ; https://doi.org/10.64898/2026.03.26.714473doi: bioRxiv preprint \n\nMain Text \nIntroduction \n \nThe bluethroat (Luscinia svecica svecica) is an ecological model species that has been \nextensively studied during the last three decades, particularly within the areas of sexual \nselection and sperm competition (Johnsen, Andersson, et al., 1998; Johnsen, Lifjeld, et al., \n1998; Johnsen & Lifjeld, 2003). This charismatic Eurasian passerine shows strong subspecies \ndifferentiation, with several morphologically distinct subspecies across its range, differing \nmainly in the overall size of both sexes and the colour of the male throat ornament (Cramp, \n1988; Johnsen et al., 2006). Few bird species have been more thoroughly studied with \nrespect to multiple matings and their consequences for male and female fitness (Fossøy et \nal., 2007; Johnsen et al., 2000; Krokene et al., 1996; Questiau et al., 1999; Rekdal et al., \n2019). This has revealed selective advantages to both sexes of engaging in extra-pair \nmatings, thus giving rise to offspring with higher cell-mediated immune response (Fossøy et \nal., 2007; Johnsen et al., 2000) that are closer to an intermediate optimum in major \nhistocompatibility complex (MHC) allele diversity than their within-pair half siblings (Rekdal \net al., 2019). \nThe bluethroat has been subjected to a wide range of molecular analyses since the advent of \nmolecular ecology, including paternity analyses using multilocus DNA fingerprinting (Krokene \net al., 1996) and microsatellites (Johnsen, Lifjeld, et al., 1998), MHC-associated mate choice \nbased on amplicon sequencing of multi-copy class II alleles (Rekdal et al., 2019), \nphylogeographic structure using Sanger sequencing of mtDNA and single-copy nuclear genes \n(Hogner et al., 2013) and SNP-analyses based on short-read whole genome sequencing \n(unpublished data). However, access to a high-quality reference genome is needed to gain a \ndeeper understanding at the molecular level of several of the recent findings in this species. \nMore specifically, a haplotype-resolved genome assembly will allow investigating the \nstructural genomic organization of MHC, which is known to be highly diverse and duplicated \nin passerine birds (Minias et al., 2019). Furthermore, it should enable phasing of the allelic \ndiversity of MHC at the individual level, improving on the pool of unphased alleles obtained \nfrom the amplicon sequencing method (Rekdal et al., 2018). Here, we present a \nhaplotype-resolved assembly of a bluethroat (Luscinia s. svecica) genome generated using \nONT long-read and Hi-C sequencing data. The public availability of this reference genome \nwill facilitate further genomic research on population structure, subspecies differentiation \nand MHC-based mate choice in this species.  \nMaterial and Methods \nSample acquisition and DNA extraction \nBlood samples were taken by brachial venipuncture from a second calendar year female \nbluethroat caught by mist netting at the Øvre Heimdalen field station, Øystre Slidre, \nInnlandet, Norway (61.419N, 8.893E) on the 31st May 2022, under permission from the \nNorwegian Food Safety Authority (FOTS ID 29575) and the Norwegian Environment Agency \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 30, 2026. ; https://doi.org/10.64898/2026.03.26.714473doi: bioRxiv preprint \n\n(ringing licence 680). Accession number in the DNA bank of the Natural History Museum, \nUniversity of Oslo: NHMO-BI-107700. \n \nDNA isolation for Oxford Nanopore Technologies (ONT) and PacBio long read sequencing \nstarted from 25µl frozen blood which were split over two Circulomics Nanobind CBB BIG \nDNA kit reactions (disks), following the manufacturer’s recommendations. Quality check of \nthe amount, purity and integrity of the isolated DNA was performed using a combination of \nQubit BR DNA quantification assay kit (Thermo Fisher), Nanodrop (Thermo Fisher), and \nFragment Analyser (DNA HS 50kb large fragment kit, Agilent Tech.). \nLibrary preparation and sequencing for de novo assembly \nBefore library preparation, a dilution of the concentrated DNA stock was purified an \nadditional time using AMPure PB beads (1:1 ratio). Approximately 7.5 µg of purified HMW \nDNA was sheared into an average fragment size of 30-35 kb large fragments for ONT with \nspeed code setting of 30+31, using the Megaruptor3 (Diagenode). The same method and \namount was used to prepare DNA for PacBio library preparation, although speed code \nsetting was increased to 32+33 to obtain shorter fragments with an average length of approx \n17-20 kb. Two ONT libraries starting from 3 µg sheared DNA each were prepped following \nthe 30 kb Human variation sequencing cell line protocol using Ligation Sequencing Kit V14 \n(SQK-LSK114) according to manufacturer guidelines. Sequencing was achieved with two \nR10.4.1 PromethION Flow Cells (FLO-PRO114M) on the PromethION P2i system (ONT) with \nthree rounds of loading per cell (300 ng/loading). Basecalling was done using the Dorado \nSUP v.5.0.0 model. For PacBio library preparation, 5 µg of fragmented DNA was used \nfollowing the PacBio protocol for HiFi library preparation using the SMRTbell® express \ntemplate kit 2.0. The final HiFi library was size-selected with a 10 kb cut-off using a \nBluePippin instrument (Sage Biosciences) and sequenced on one 8M SMRT cells using the \nSequel II Binding kit 2.2 and Sequencing chemistry v2.0 on the Pacbio Sequel II instrument. \nBoth ONT and PacBio sequencing was performed by the Norwegian Sequencing Centre, a \nnational sequencing core at the University of Oslo, Norway.  \n \nStarting with 10-15 µl frozen nucleated blood, a Hi-C library was prepared using the Arima \nHigh Coverage HiC kit (Arima Genomics Inc.), following the manufacturer’s \nrecommendations (document part number A160162v01). Final library quality was assayed as \nabove in addition to qPCR using the Kapa Library quantification kit for Illumina (Roche Inc.). \nThe library was sequenced with other libraries on the Illumina NovaSeq SP flowcell with \n2*150 bp paired end mode at the Norwegian Sequencing Centre. \nGenome assembly and curation, annotation and evaluation \nA full list of relevant software tools and versions is presented in Table 1. KMC (Kokot et al., \n2017) was used to count k-mers of size 32 in the ONT reads, excluding k-mers occurring \nmore than 10,000 times. GenomeScope (Ranallo-Benavidez et al., 2020) was run on the \nk-mer histogram output from KMC to estimate genome size, heterozygosity and \nrepetitiveness while ploidy level was calculated using Smudgeplot (Ranallo-Benavidez et al., \n2020). The ONT reads were assembled using hifiasm (Cheng et al., 2021) with Hi-C \nintegration, producing two pseudo-haplotype-resolved assemblies: pseudo-haplotype one \n(hap1) and pseudo-haplotype two (hap2). These are hereafter referred to as haplotype 1 \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 30, 2026. ; https://doi.org/10.64898/2026.03.26.714473doi: bioRxiv preprint \n\nand haplotype 2, or hap1 and hap2. Unique k-mers in each assembly/pseudo-haplotype \nwere identified using meryl (Rhie et al., 2020) and used to create two sets of Hi-C reads, one \nwithout any k-mers occurring uniquely in hap1 and the other without k-mers occurring \nuniquely in hap2. K-mer filtered Hi-C reads were aligned to each scaffolded assembly using \nBWA-MEM (Li, 2013) with -5SPM options. The alignments were sorted based on name using \nsamtools (Li et al., 2009) before applying samtools fixmate to remove unmapped reads and \nsecondary alignments and to add mate score, and samtools markdup to remove duplicates. \nThe resulting BAM files were used to scaffold the two assemblies using YaHS (Zhou et al., \n2023) with default options. FCS-GX (Astashyn et al., 2023) was used to search for \ncontamination. Contaminated sequences were removed. Merqury (Rhie et al., 2020) was \nused to assess the completeness and quality of the genome assemblies by comparing to the \nk-mer content of the Hi-C reads. The assemblies were manually curated in PretextView using \nRapid Curation 2.0. To generate Hi-C contact maps, Hi-C reads were mapped to the \nassemblies using BWA-MEM (Li, 2013) with the same parameters as used for scaffolding, and \ncontact maps were generated with PretextMap and visualized in PretextSnapshot. \nChromosomes (including sex chromosomes) were identified by alignment to Luscinia luscinia \n(GCA_034336685) and Luscinia megarhynchos (GCA_034336665), supported by inspection \nof Hi-C contact maps. A second curation pass focused on microchromosomes was carried \nout by selecting scaffolds ≤20 Mb. MicroFinder (Mathers et al., 2025) was used to map \nconserved microchromosome genes to scaffolds, and candidate microchromosomes were \nidentified by generating a gene density BED track for visualization in PretextView. Curated \nscaffolds were reintegrated into the assembly, followed by a final curation pass in \nPretextView. Gfastats (Formenti et al., 2022) was used to output different assembly statistics \nof the assemblies. Assembly statistics were visualized using BlobToolKit and BlobTools2 \n(Laetsch & Blaxter, 2017), along with blobtk (see Supplementary Fig. 2). BUSCO (Manni et al., \n2021) was used to assess the completeness of the genome assemblies by comparing against \nthe expected gene content in the aves lineage. MitoHiFi (Uliano-Silva et al., 2023) was used \nto search for mitochondrial DNA in the final assembly and reads. \n \nTable 1. Software tools: versions and sources. *Annotation \nSoftware tool Version Source \nBlobToolKit 4.3.2 https://github.com/blobtoolkit/blobtoolkit \nblobtk 0.5.8 https://github.com/blobtoolkit/blobtk \nBUSCO 5.8.2; 6.0.0* https://gitlab.com/ezlab/busco \nhifiasm  0.25.0 https://github.com/chhylp123/hifiasm \nDorado v.0.9.1 https://github.com/nanoporetech/dorado \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 30, 2026. ; https://doi.org/10.64898/2026.03.26.714473doi: bioRxiv preprint \n\nKMC 3.2.4 https://github.com/refresh-bio/KMC  \nGenomeScope 2.0.1 https://github.com/tbenavi1/genomescope2.0  \nHiFiAdapterFilt 2.0.1 https://github.com/sheinasim/HiFiAdapterFilt  \nPretextView 1.0.1 https://github.com/wtsi-hpag/PretextView \nPretextMap 0.1.9 https://github.com/wtsi-hpag/PretextMap \nPretextSnapshot 0.0.4 https://github.com/wtsi-hpag/PretextSnapshot \nmeryl 1.3.0 https://github.com/marbl/meryl  \nBWA-MEM 0.7.18 https://github.com/lh3/bwa  \nsamtools 1.21 https://github.com/samtools/samtools  \nYaHS 1.2.2 https://github.com/c-zhou/yahs \nFCS-GX 0.4.0 https://github.com/ncbi/fcs \nfcsadaptor 0.2.2 https://github.com/ncbi/fcs \nMerqury 1.3 https://github.com/marbl/merqury \nAGAT 1.5.1 https://github.com/NBISweden/AGAT \nMitoHiFi  3.0.0 https://github.com/marcelauliano/MitoHiFi \nminiprot 0.18 https://github.com/lh3/miniprot \nGALBA  1.0.11 https://github.com/Gaius-Augustus/GALBA \nRED 2018.09.10  http://toolsmith.ens.utulsa.edu/ \nFunannotate 1.8.17 https://github.com/nextgenusfs/funannotate \nHMMR3 v3.2.1 https://github.com/EddyRivasLab/hmmer \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 30, 2026. ; https://doi.org/10.64898/2026.03.26.714473doi: bioRxiv preprint \n\nEvidenceModeler 1.1.1 https://github.com/EVidenceModeler/EVidenceM\nodeler \nHelixer 0.3.6 https://github.com/usadellab/Helixer \nDIAMOND 2.1.13 https://github.com/bbuchfink/diamond \nInterProScan 5.62-94.0-foss-202\n2a \nhttps://www.ebi.ac.uk/interpro/search/sequence/ \nEMBLmyGFF3 2.4 https://github.com/NBISweden/EMBLmyGFF3 \nRapid Curation 2.0 964d17e997e00c6\n9f25940cf96d3658\nbda631147 \nhttps://github.com/Nadolina/Rapid-curation-2.0, \nMicrofinder 0.2 https://github.com/sanger-tol/MicroFinder \n \nWe annotated the genome assemblies using a pre-release version of the EBP-Nor genome \nannotation pipeline (https://github.com/ebp-nor/GenomeAnnotation ). First, AGAT \n(https://zenodo.org/record/7255559) agat_sp_keep_longest_isoform.pl and \nagat_sp_extract_sequences.pl were used on the Gallus gallus (bGalGal1.mat.broiler.GRCg7b) \ngenome assembly and annotation to generate one protein (the longest isoform) per gene. \nMiniprot (Li, 2023) was used to align the proteins to the curated assemblies. \nUniProtKB/Swiss-Prot (UniProt Consortium, 2023) release 2025_3 in addition to the aves \npart of OrthoDB v12 (Kuznetsov et al., 2023) were also aligned separately to the assemblies. \nRed (Girgis, 2015) was run via redmask (https://github.com/nextgenusfs/redmask ) on the \nassemblies to mask repetitive areas. GALBA (Brůna et al., 2023; Buchfink et al., 2015; Hoff & \nStanke, 2019; Li, 2023; Stanke et al., 2006) was run with the chicken proteins using the \nminiprot mode on the masked assemblies. Helixer (Holst et al., 2025) was run using the \nvertebrate-specific model (vertebrate_v0.3_m_0080). The funannotate-runEVM.py script \nfrom Funannotate was used to run EvidenceModeler (Haas et al., 2008) on the alignments of \nchicken proteins, UniProtKB/Swiss-Prot proteins, aves proteins and the predicted genes from \nGALBA and Helixer. The resulting predicted proteins were compared to the protein repeats \nthat Funannotate distributes using DIAMOND blastp and the predicted genes were filtered \nbased on this comparison using AGAT. The filtered proteins were compared to the \nUniProtKB/Swiss-Prot release 2025_3 using DIAMOND (Buchfink et al., 2015) blastp to find \ngene names and InterProScan was used to discover functional domains. AGATs \nagat_sp_manage_functional_annotation.pl was used to attach the gene names and \nfunctional annotations to the predicted genes. EMBLmyGFF3 (Norling et al., 2018) was used \nto combine the fasta files and GFF3 files into a EMBL format for submission to ENA. EDTA  \n(Ou et al., 2019) was used to annotate transposable elements. \n \nTo characterize differences between haplotypes, we aligned homologous chromosomes \nusing minimap2 (Li, 2017). The resulting alignment was processed with the \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 30, 2026. ; https://doi.org/10.64898/2026.03.26.714473doi: bioRxiv preprint \n\nminimap2-included paftools.js producing a report listing the number of insertions, SNPs and \nindels.  \n \n \n \nFigure 1. Sequenced specimen and genome profile. A)  Photograph of the female Luscinia s. svecica specimen \nused for genome sequencing (Photo: Lars Erik Johannessen). B) GenomeScope profile of the ONT reads from \nthe sequenced individual. This analysis estimates a 1289 Mb genome, with 0.953 % heterozygosity and bimodal \npattern characteristic of a diploid genome. The left-hand peak of k-mers corresponds to k-mers from \nheterozygous regions of the genome, while the right-hand peak is from homozygous regions. \nMHC annotation \nMajor histocompatibility complex (MHC) genes were annotated using targeted manual \ncuration in addition to the standard genome annotation described above. Candidate loci \nwere first identified using Pfam domain annotations. Each phased assembly was translated \nin all six reading frames and screened against a curated set of Pfam HMMs using HMMER3 \n(Eddy, 2011). Hits to MHC-associated domains were mapped to gene models to identify \ncandidate loci. \nGene models overlapping expected MHC domains were retained for further inspection. Loci \ncontaining either the MHC class I domain (PF00129), the MHC class II beta domain \n(PF00969), or the MHC class II alpha domain (PF00993), together with the immunoglobulin \nsuperfamily C1-set domain (PF07654), were considered putatively functional. \nTo evaluate gene model structure, the standard genome annotation was compared with ab \ninitio predictions generated by Helixer (Holst et al., 2025) using the vertebrate-specific \nmodel (vertebrate_v0.3_m_0080). Domain architectures and coding sequences (CDS) were \ninspected iteratively. Where Helixer predictions better matched the expected MHC domain \norganization, these models were incorporated and manually refined by merging partial \npredictions, adjusting CDS boundaries, or removing spurious features. \nCurated CDS sequences from candidate loci were combined with full-length MHCI and \nMHCIIβ sequences from Westerdahl et al. (2022) to support locus classification. Putative \nMHCI sequences identified as CD1d-like based on BLAST similarity were excluded. Extracted \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 30, 2026. ; https://doi.org/10.64898/2026.03.26.714473doi: bioRxiv preprint \n\nCDS segments were concatenated to reconstruct full coding sequences for each locus and \nhaplotype. \nFor each MHC class, CDS sequences were aligned in R using the msa package with MUSCLE, \nand initial neighbor-joining trees were inferred using ape. Alignments and trees were \nvisualized together with exon boundaries and coding structure to support manual curation ( \nSupplementary Figures 3 and 4). \nComparative visualizations of the MHC region on chromosome 35 were generated for both \nhaplotypes by combining pairwise haplotype alignments, curated gene models, Pfam \ndomain annotations, read coverage, and assembly gap positions. Smoothed ONT and HiFi \nread coverage was calculated from bedGraph files, and assembly gaps were identified as \nruns of Ns in the genome sequence. \nTo compare curated loci with previously characterized MHC allele variation, exon sequences \nfrom amplicon-based genotyping studies were incorporated. MHC class IIB exon 2 \n(MHCIIβe2) sequences were obtained from Rekdal et al. (2018) (GenBank \nMF769842–MF769959) and Rekdal et al. (2019) (GenBank MN332585–MN333760), and \nMHC class I exon 3 sequences from Rekdal et al. (2018) (GenBank MF769960–MF769977). \nCurated loci were filtered using Pfam HMM coverage (≥85%) and score (≥65) thresholds to \nretain genes with strong support for complete MHC class I or MHC class II beta domains. A \nrepresentative reference sequence of median length was selected from each exon dataset, \nand the corresponding regions (MHCI exon 3: 239 bp; MHCIIβ exon 2: 267 bp) were \nextracted from curated loci by pairwise alignment. \nThe extracted exon regions were combined with previously published amplicon sequences, \ntranslated, and aligned at the amino acid level in R using DECIPHER. Neighbor-joining trees \nwere inferred from amino acid distances with ape, rooted with zebra finch (Taeniopygia \nguttata) outgroup sequences, and visualized with ggtree. \nResults \nDe novo genome assembly and annotation \nThe genome of the female Luscinia svecica svecica (Figure 1), had an estimated genome size \nof 1.29 Gb, with 0.953% heterozygosity and a bimodal distribution based on the k-mer \nspectrum (Figure 1B). A total of 33-fold coverage in R10 Oxford Nanopore reads and 38-fold \ncoverage in Arima Hi-C reads resulted in two haplotype-separated assemblies. The final \nassemblies have total lengths of 1461 Mb and 1171 Mb (Table 2 and Figure 2), respectively. \nHaplotypes one and two have scaffold N50 size of 36.0 Mb and 40.3 Mb, respectively, and \ncontig N50 of 7.8 Mb and 8.2 Mb, respectively (Table 2, Figure 2). 40 autosomes were \nidentified in both haplotypes (numbered by length in haplotype one, with the homolog in \nhaplotype two receiving the same number). Sex chromosomes W and Z were added to \nhaplotype one. \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 30, 2026. ; https://doi.org/10.64898/2026.03.26.714473doi: bioRxiv preprint \n\n \n \nFigure 2: Metrics of the genome assemblies of Luscinia s. svecica. A) Haplotype one. B) Haplotype two. The \nBlobToolKit Snailplots show N50 metrics and BUSCO gene completeness. The two outermost bands of the circle \nsignify GC versus AT composition at 0.1% intervals. Light orange shows the N90 scaffold length, while the \ndeeper orange is N50 scaffold length. The red line shows the size of the largest scaffold. All the scaffolds are \narranged in a clockwise manner from the largest to the smallest and are shown in darker gray with white lines \nat different orders of magnitude, while the light gray shows cumulative count of scaffolds.  \n \nTable 2: Genome data for Luscinia s. svecica. \nProject accession data \nSpecies Luscinia svecica svecica  \nSpecimen bLusSve1  \nNCBI Taxonomy ID 52792 (subspecies 1434997)  \nBioProject PRJEB90386  \nBioSample ID SAMEA118366426  \nIsolate information Blood sampling from brachial vein  \nRaw data accessions \nPacBio HiFi reads* ERR15077406 1 PACBIO_SMRT (Sequel II) runs: 2.3 \nM reads, 34.4 Gb  \nHi-C Illumina reads ERR15077407 1 ILLUMINA (Illumina NovaSeq S4) \nrun: 183 M pairs of reads, 55.4 Gb \nONT reads ERR16909039 and ERR16912516 2 ONT (PromethION P2i) runs: 4.4 \nM reads, 48.8 Gb  \nGenome assembly metrics \nHiFi read coverage* 23x \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 30, 2026. ; https://doi.org/10.64898/2026.03.26.714473doi: bioRxiv preprint \n\nONT read coverage 33x \nAssembly identifier bLusSve1.1.hap1 bLusSve1.1.hap2 \nSpan (Mb) 1461 1171 \nNumber of Contigs 806 474 \nContig N50 length \n(Mb) \n7.8 8.2 \nLongest contig (Mb) 30.6 36.6 \nNumber of gaps 167 165 \nNumber of Scaffolds 640 312 \nScaffold N50 length \n(Mb) \n36.0 30.3 \nLongest scaffold \n(Mb) \n151.7 151.4 \nConsensus quality \n(QV) compared to \nHi-C (compared to \nONT; compared to \nHiFi) \n34.4 (64.6; 38.4) 34.4 (64.6; 38.2) \nBoth assemblies 34.4 (64.6; 38.3) \nk-mer completeness \n(percentage; \ncompared to ONT; \ncompared to HiFi) \n94.7 (82.2; 83.3) 93.8 (74.9; 75.9) \nBoth assemblies 96.7 (98.9; 99.2) \nPercentage of \nassembly mapped to \nchromosomes \n77.4 88.4 \nCompari\nsons \n(hap2 \naligned \nto hap1) \nBases in \nalignment \n985,401,588 \nSubstitution\ns \n(percentage\n) \n7,900,943 (1.65) \n1bp \ndeletions \n288,146 \n1bp \ninsertions \n287,689 \n2bp \ndeletions \n80,133 \n2bp 80,239 \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 30, 2026. ; https://doi.org/10.64898/2026.03.26.714473doi: bioRxiv preprint \n\ninsertions \n[3,50) \ndeletions \n236,491 \n[3,50) \ninsertions \n236,365 \n[50,1000) \ndeletions \n25,303 \n[50,1000) \ninsertions \n25,388 \n>=1000 \ndeletions \n1,254 \n>=1000 \ninsertions \n1,240 \n \nSex chromosomes Z, W  \nOrganelles MT  \nGenome annotation \nNumber of \nprotein-coding genes \n22,462 18,769 \n \nNumber of \nprotein-coding genes \nwith functional \ndomain \n21,900 18,408 \nNumber of \nprotein-coding genes \nwith gene names \n17,650 15,425 \n \nBUSCO** C:99.2%[S:98.8%,D:0.4%],F:0.3%,\nM:0.5%,n:8338,E:5.9% \nC:94.9%[S:94.6%,D:0.3%],F:0.3%,M:4\n.7%,n:8338,E:5.7% \n* PacBio HiFi reads were used for coverage comparison analyses, including microchromosomes, but not for \nprimary assembly. ** BUSCO scores based on the aves BUSCO set using v5.4.7. C  = complete [S = single copy, \nD = duplicated], F = fragmented, M = missing, n = number of orthologues in comparison. \n \nHaplotype one had 99.2% and haplotype two 94.9% complete BUSCO genes using the aves \nlineage set. When compared to a k-mer database of the Hi-C reads, haplotype one had a \nk-mer completeness of 94.7%, haplotype two of 93.8%, and combined they have a \ncompleteness of 96.7%. Further, haplotype one has an assembly consensus quality value \n(QV) of 34.4 and haplotype two of 34.4, where a QV of 40 corresponds to one error every \n10,000 bp, or 99.99% accuracy compared to a k-mer database of the Hi-C reads (QV 64.6 and \n64.6, respectively, compared to a k-mer database of the ONT reads). The Hi-C contact map \nfor the assemblies are found in Supplementary Figure 1, and show clear separation of the \ndifferent chromosomes. \n \nWhen comparing the two haplotypes using minimap2, there are 7,900,943 SNPs differences \n(1.65 % of the aligned sequence), 631,327 deletions in hap2 compared to hap1 ranging from \n1 bp to more than 1000 bp and 630,921 insertions from 1 bp to more than 1000 bp in size \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 30, 2026. ; https://doi.org/10.64898/2026.03.26.714473doi: bioRxiv preprint \n\n(Table 2). A total of 20,779 and 20,937 protein-coding genes were annotated in haplotype \none and two, respectively (Table 2). In both haplotypes 34.1% of the genome assembly was \nannotated as a repetitive sequence. \nGenomic organization and copy number variation of the MHC region \nPrevious work in bluethroat has shown that MHCIIβ diversity is linked to mate choice and \noffspring immune performance, with evidence that selection favors an intermediate allele \ncount (Rekdal et al., 2019). However, these inferences rely on amplicon-based allele counts, \nwithout knowledge of the underlying genomic structure. To interpret these patterns \nmechanistically, we first characterize the organization, copy number, and chromosomal \ndistribution of MHC genes in our haplotype-resolved assembly. Due to the difficulty \npresented in annotating paralogs we applied a targeted curation strategy integrating \ncomplementary annotation approaches and domain-based evidence to identify candidate \nMHC genes. \nIn total, we identified 12 and 5 MHCI loci in haplotype 1 and haplotype 2, respectively, that \ncontain the MHCI domain (PF00129) and a C1-set domain (PF07654) (Supplementary Figure \n3). All loci are located on chromosome 35, except for a single locus per haplotype found on \nchromosome 22. This single locus retains similarity to MHCI genes but shows disrupted \ncoding structure, including irregular exon boundaries and premature stop codons in the \nfull-length alignment. For MHC class II, we detected 29 and 26 MHCIIβ loci in haplotype 1 \nand haplotype 2, respectively, that contain both the MHCIIβ domain (PF00969) and a C1-set \ndomain (PF07654) (Supplementary Figure 4). All loci are located on chromosome 35, except \nfor a deviating single locus per haplotype found on chromosome 21.  \nMost MHC genes are located on chromosome 35 (Figure 3). In hap2, a large cluster of 22 \nMHCIIβ genes lies between BRD2 and MHCIIA, 11 on the BRD2 strand, and 11 on the \nopposite strand. Hap 1 has fewer MHCIIβ genes in this region, eight in the BRD2 orientation \nand two on the opposite strand. Additional MHCIIβ genes also occur with MHCI genes close \nto the flanking gene flot1. In a region 1MB further upstream there are 3 additional MHCIIβ \ngenes in hap2, two in the BRD2 orientation, and one on the opposite strand. In this region \nhap1 has 18, in four nearly equally sized clusters (three of which have the following pattern: \nthree in the BRD2 orientation and two on the opposite strand). In this second region is \nwhere we find MHCI loci, near flot1. These MHCI genes come in pairs oriented away from \neach other. MHCIIβ genes in this region are flanked by these MHCI pairs. In both haplotypes \nno gaps occur within the gene dense MHC regions, only a single assembly gap is present in \neach haplotype, occurring between the larger MHCIIβ-only cluster and the MHCI+ MHCIIβ \ncluster. \n \nTBXN, TAP1 and TAP2 are located on chromosome 39, separate from MHC genes on \nchromosome 35. This separation is expected in passeriforms, where the genomic block \ncontaining TAP genes and TNXB has been translocated away from the core MHC region \n(Westerdahl et al., 2022). \n \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 30, 2026. ; https://doi.org/10.64898/2026.03.26.714473doi: bioRxiv preprint \n\n \nFigure 3. Visualization of the MHC region for both haplotypes on chromosome 35. Haplotype 2 (top) and \nhaplotype 1 (bottom) are shown as rounded grey segments. The central band shows pairwise alignments \nbetween haplotypes generated with minimap2. Two higher-order MHC blocks are indicated: a left block \ncontaining multiple MHCIIβ loci and a single MHCIIα locus (IIB+IIA), and a right block where MHCI and MHCIIβ \nloci are interspersed (I+IIB). Lightning bolt symbols mark assembly gaps. Points surrounding each haplotype \nindicate the positions of MHC genes and are colored by domain composition: purple = MHCI (PF00129) + C1 \n(PF07654), red = MHCIIA (PF00993) + C1, turquoise = MHCIIβ (PF00969) + C1. Desaturated purple and \nturquoise indicate MHCI or MHCIIβ genes lacking a C1 domain. Grey indicates genes containing only a C1 \ndomain. Black indicates named flanking genes. Exclamation marks indicate genes with low or partial HMM \nscores. Cartoon gene models show gene orientation and the positions of flanking genes. ONT and HiFi read \ncoverage tracks are shown above haplotype 2 and below haplotype 1.  \nAssembly-derived MHCI exon 3 and MHCIIβ exon 2 sequences were distributed across the \ndiversity of previously published bluethroat alleles rather than forming distinct clusters (Fig. \n4). The MHCIIβ tree shown in Fig. 4B is based on the full dataset but pruned to include only \nthe 2018 alleles, with the underlying topology preserved; the complete tree including all tips \nis provided in Supplementary Fig. 5. In both trees, assembly alleles were interspersed among \nPCR-derived alleles reported by Rekdal et al. (2018, 2019). Several loci showed identical \nnucleotide sequences to previously reported alleles, allowing specific amplicon alleles to be \nlinked to genomic loci in the assembly. These matches occurred across multiple loci but were \ngenerally haplotype-specific. The single exception was a MHCI locus that showed 100% \nnucleotide identity between haplotypes (hap1 FUNCG00000016529 and hap2 \nFUNCG00000016440 on chromosome 35), both corresponding to allele MF769973.1 \n(Supplementary figure 6). Within haplotypes, identical exon sequences were occasionally \nobserved across different loci, whereas most loci differed slightly between haplotypes. \nSimilarly, the MHCIIβ locus on chromosome 21 showed identical exon sequences between \nhaplotypes (hap1 FUNCG00000013298 and hap2 FUNCG00000013300). Within haplotypes, \nidentical exon sequences were occasionally observed across different loci, whereas most loci \ndiffered slightly between haplotypes. Supplementary figure 6 shows these connections. \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 30, 2026. ; https://doi.org/10.64898/2026.03.26.714473doi: bioRxiv preprint \n\n \n \nFigure 4. Neighbor joining trees of bluethroat MHC alleles. A) MHCI exon 3 alleles. B) MHCIIβ exon 2 alleles. \nTrees were inferred from amino acid alignments generated with DECIPHER and visualized in circular layout. The \ntrees illustrate similarity relationships among alleles rather than strict evolutionary history due to known gene \nconversion and recombination in MHC loci. Alleles recovered from the assembly are highlighted in blue and \npreviously published GenBank alleles are shown in light grey. The tree is rooted with the Westerdahl allele \nTgut-DAB1045_1 (MHCIIβ) or Tgut-UA1045_1 (MHCI) shown in orange. The MHCIIβ topology was inferred from \na dataset including alleles from both Rekdal et al. (2018) and Rekdal et al. (2019); for visual purposes the tree \nshown here includes only 2018 alleles after dropping 2019-specific tips from the full tree. \nDiscussion \n \nWe present a haplotype-resolved genome assembly of the bluethroat (Luscinia svecica \nsvecica) that enables detailed investigation of complex genomic regions. The assemblies are \nhighly contiguous and near chromosome-level, with scaffold N50 values of 36.0 Mb and 40.3 \nMb and up to 88.4% of sequence assigned to chromosomes. Gene completeness is high \n(BUSCO 99.2% and 94.9%), supported by k-mer completeness of 98.9% and a consensus QV \nof 64.6. \nGC-rich and repetitive regions enriched in microchromosomes have been challenging to \nassemble, and are therefore often underrepresented in avian genome assemblies (Peona et \nal., 2018). Recent work has shown that Oxford Nanopore reads exhibiting less severe \ncoverage dropouts than PacBio HiFi in GC-rich sequence contexts improving recovery of \nthese regions, with (Formenti et al., 2025). Consistent with this, in the bluethroat we \nobserve improved continuity across complex regions using Oxford Nanopore data, \nexemplified by the largely continuous assembly of the MHC region on chromosome 35, \nwhere HiFi coverage is low and uneven (Figure 3). \n \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 30, 2026. ; https://doi.org/10.64898/2026.03.26.714473doi: bioRxiv preprint \n\nEarly descriptions of avian major histocompatibility complex (MHC) architecture established \ntwo ends of a continuum. In the chicken, the MHC was defined as a “minimal essential” \nsystem with a compact structure and few loci (Kaufman et al., 1995), whereas passerines \nshow extensive duplication and high allelic diversity (Westerdahl, 2007). Resolving this \nexpanded passerine architecture has remained difficult, as highly duplicated and repetitive \nregions are typically fragmented in avian genome assemblies. Long-read sequencing now \nmakes it possible to directly resolve locus organization, copy number, and synteny in these \nregions, providing the context needed to interpret the complex MHC structure observed in \npasserines such as the bluethroat. \n \nThis haplotype-resolved bluethroat genome assembly provides a substantially improved \nview of the genomic organization of the MHC region. Previous studies of bluethroat MHC \ndiversity relied on amplicon sequencing, leaving the underlying genomic structure unknown. \nIn our assembly, the MHC region on chromosome 35 is largely continuous (only interrupted \nby one gap), allowing the organization and copy number of MHCI and MHCIIβ loci to be \nexamined directly. Such regions are typically difficult to reconstruct because avian \nmicrochromosomes are enriched for GC-rich and repetitive sequence that often leads to \nfragmented or missing sequence in genome assemblies (Li & Durbin, 2024; Peona et al., \n2018).  \n \nRead coverage across chromosome 35 reveals a region containing multiple MHCIIβ loci \nwhere HiFi coverage drops markedly, while ONT reads continue across the region. This \nsuggests that an assembly relying on HiFi reads would likely fragment this locus cluster. \n \nThe number of MHC loci recovered in this assembly is consistent with the extensive \nduplication typical of passerine birds (Minias et al., 2019; Westerdahl et al., 2022). The locus \ncounts observed in the current assembly fall within the range expected for the bluethroat \n(Rekdal et al., 2018, 2019). Our phased assembly reveals substantial structural differences \nbetween haplotypes. Copy-number variation is evident for both MHCI and MHCIIβ loci, with \nclear differences in the number and arrangement of loci between the two haplotypes. \nBecause the assemblies are close to continuous across the gene-dense regions, these \ndifferences likely reflect biological variation rather than assembly artifacts. A single assembly \ngap remains within the chromosome, leaving four possible structural configurations across \nthe breakpoint. However, all imply substantial divergence between haplotypes. Hi-C contact \npatterns support the configuration shown here. \n \nThe MHCIIβ locus on chromosome 21 differs from the duplicated loci within the main MHC \ncluster in genomic location as well as sequence similarity. In the gene tree, this locus groups \nwith homologous sequences from other species rather than with the species-specific \nduplicated loci, consistent with the distinction between conserved single-copy loci and \nlineage-specific tandem expansions described in other passerines (Westerdahl et al., 2022). \nIn the great reed warbler (Acrocephalus arundinaceus), this type of locus occurs on the same \nchromosome as the main MHC region (Westerdahl et al., 2022), whereas in the bluethroat \nassembly it is located on a different chromosome.  \n \nThe interspersed MHCI and MHCIIβ pattern found in bluethroat is interesting and has to our \nknowledge not been seen in other published or available genomes. Since this is a deviating \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 30, 2026. ; https://doi.org/10.64898/2026.03.26.714473doi: bioRxiv preprint \n\npattern from previously reported assemblies, it remains to be seen if this has any functional \nimplications. \n \nThese haplotype-resolved assemblies provide a detailed overview of the genomic \norganization and orientation of the bluethroat MHC but are based on a single individual. \nAdditional genome assemblies from multiple individuals will therefore be important to \nbetter characterize population-level variation in this region. Nevertheless, the \nhaplotype-resolved assemblies presented here reveal clear copy number variation in MHC \nloci, indicating substantial structural variation within the MHC, consistent with the high \nallelic diversity and copy number variation previously reported from PCR-based studies. \nFunding  \nThis project was funded by the Research Council of Norway project 326819 (The Earth \nBiogenome Project Norway) to KSJ. \nAcknowledgements \nThis project received data management and infrastructure support from ELIXIR Norway, \nsupported by the Research Council of Norway’s grant 270068, the University of Bergen, the \nUniversity of Oslo, the Arctic University of Norway in Tromsø, the Norwegian University of \nScience and Technology and the Norwegian University of Life Sciences: NMBU. The authors \nacknowledge support from the National Infrastructure for High Performance Computing and \nresources provided by Sigma2 as well as Data Storage in Norway (project NN8013K) for \ncomputational work. The Norwegian Sequencing Centre generated the sequencing data \nused in this project (http://sequencing.uio.no).  \nData Availability \nData generated for this study are available under ENA BioProject PRJEB90386. Raw PacBio \nsequencing data for the bluethroat (ENA BioSample: SAMEA118366426) are deposited in \nENA under ERX14482045, while Illumina Hi-C sequencing data is deposited in ENA under \nERX14482046. Base-called ONT reads are found in ERX16294347 and ERX16297870. \n Pseudo-haplotype one can be found in ENA at PRJEB89911, while pseudo-haplotype two is \nPRJEB90385. Please note that the first version of the assemblies are based on PacBio HiFi \nreads, while the second version is based on ONT.  \n \nRaw ONT data is deposited at NIRD Archive: https://doi.org/10.11582/2025.nu3dqs6m. \n \nThe genome assemblies and transposon and gene annotations are available at Zenodo: \n1https://doi.org/10.5281/zenodo.19220545 \n \n \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 30, 2026. ; https://doi.org/10.64898/2026.03.26.714473doi: bioRxiv preprint \n\n \nReferences \nAstashyn, A., Tvedte, E. S., Sweeney, D., Sapojnikov, V., Bouk, N., Joukov, V., Mozes, E., Strope, P . K., \nSylla, P . M., Wagner, L., Bidwell, S. L., Clark, K., Davis, E. W., Smith-White, B., Hlavina, W., Pruitt, \nK. D., Schneider, V. A., & Murphy, T. D. (2023). Rapid and sensitive detection of genome \ncontamination at scale with FCS-GX. In Bioinformatics. Bioinformatics. \nhttps://doi.org/10.1101/2023.06.02.543519 \nBrůna, T., Li, H., Guhlin, J., Honsel, D., Herbold, S., Stanke, M., Nenasheva, N., Ebel, M., Gabriel, L., & \nHoff, K. J. (2023). Galba: genome annotation with miniprot and AUGUSTUS. BMC Bioinformatics, \n24(1), 327. \nBuchfink, B., Xie, C., & Huson, D. H. (2015). Fast and sensitive protein alignment using DIAMOND. \nNature Methods, 12(1), 59–60. \nCheng, H., Concepcion, G. T., Feng, X., Zhang, H., & Li, H. (2021). Haplotype-resolved de novo \nassembly using phased assembly graphs with hifiasm. Nature Methods, 18(2), 170–175. \nCramp, S. (Ed.). (1988). The Birds of the Western Palearctic, Volume 5: Tyrant Flycatchers to Thrushes. \nOxford University Press. \nEddy, S. R. (2011). Accelerated Profile HMM Searches. PLoS Computational Biology, 7(10), e1002195. \nFormenti, G., Abueg, L., Brajuka, A., Brajuka, N., Gallardo-Alba, C., Giani, A., Fedrigo, O., & Jarvis, E. D. \n(2022). Gfastats: conversion, evaluation and manipulation of genome sequences using assembly \ngraphs. Bioinformatics (Oxford, England), 38(17), 4214–4216. \nFormenti, G., Jain, N., Medico, J. A., Sollitto, M., Antipov, D., Barcellos, S., Biegler, M., Borges, I., \nChang, J. K., Chen, Y ., Cheng, H., Conceição, H., Davenport, M., De Oliveira, L., Duarte, E., \nDurham, G., Fenn, J., Forde, N., Galante, P . A., … Jarvis, E. D. (2025). The complete genome of a \nsongbird. In bioRxivorg. https://doi.org/10.1101/2025.10.14.682431 \nFossøy, F., Johnsen, A., & Lifjeld, J. T. (2007). MULTIPLE GENETIC BENEFITS OF FEMALE PROMISCUITY \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 30, 2026. ; https://doi.org/10.64898/2026.03.26.714473doi: bioRxiv preprint \n\nIN A SOCIALLY MONOGAMOUS PASSERINE: MULTIPLE GENETIC BENEFITS. Evolution; \nInternational Journal of Organic Evolution, 62(1), 145–156. \nGirgis, H. Z. (2015). Red: an intelligent, rapid, accurate tool for detecting repeats de-novo on the \ngenomic scale. BMC Bioinformatics, 16(1), 227. \nHaas, B. J., Salzberg, S. L., Zhu, W., Pertea, M., Allen, J. E., Orvis, J., White, O., Buell, C. R., & Wortman, \nJ. R. (2008). Automated eukaryotic gene structure annotation using EVidenceModeler and the \nProgram to Assemble Spliced Alignments. Genome Biology, 9(1), R7. \nHoff, K. J., & Stanke, M. (2019). Predicting genes in single genomes with AUGUSTUS. Et Al [Current \nProtocols in Bioinformatics], 65(1), e57. \nHogner, S., Laskemoen, T., Lifjeld, J. T., Pavel, V., Chutný, B., García, J., Eybert, M.-C., Matsyna, E., & \nJohnsen, A. (2013). Rapid sperm evolution in the bluethroat (Luscinia svecica) subspecies \ncomplex. Behavioral Ecology and Sociobiology, 67(8), 1205–1217. \nHolst, F., Bolger, A. M., Kindel, F., Günther, C., Maß, J., Triesch, S., Kiel, N., Saadat, N., Ebenhöh, O., \nUsadel, B., Schwacke, R., Weber, A. P . M., Bolger, M. E., & Denton, A. K. (2025). Helixer: ab initio \nprediction of primary eukaryotic gene models combining deep learning and a hidden Markov \nmodel. Nature Methods. https://doi.org/10.1038/s41592-025-02939-1 \nJohnsen, A., Andersen, V., Sunding, C., & Lifjeld, J. T. (2000). Female bluethroats enhance offspring \nimmunocompetence through extra-pair copulations. Nature, 406(6793), 296–299. \nJohnsen, A., Andersson, S., Fernandez, J. G., Kempenaers, B., Pavel, V., Questiau, S., Raess, M., Rindal, \nE., & Lifjeld, J. T. (2006). Molecular and phenotypic divergence in the bluethroat ( Luscinia \nsvecica ) subspecies complex. Molecular Ecology, 15(13), 4033–4047. \nJohnsen, A., Andersson, S., Örnborg, J., & Lifjeld, J. T. (1998). Ultraviolet plumage ornamentation \naffects social mate choice and sperm competition in bluethroats (Aves: Luscinia s. svecica): a \nfield experiment. Proceedings. Biological Sciences, 265(1403), 1313–1318. \nJohnsen, A., & Lifjeld, J. T. (2003). Ecological constraints on extra-pair paternity in the bluethroat. \nOecologia, 136(3), 476–483. \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 30, 2026. ; https://doi.org/10.64898/2026.03.26.714473doi: bioRxiv preprint \n\nJohnsen, A., Lifjeld, J. T., Rohde, P . A., Primmer, C. R., & Ellegren, H. (1998). Sexual conflict over \nfertilizations: female bluethroats escape male paternity guards. Behavioral Ecology and \nSociobiology, 43(6), 401–408. \nKaufman, J., Völk, H., & Wallny, H. J. (1995). A “minimal essential Mhc” and an “unrecognized Mhc”: \ntwo extremes in selection for polymorphism. Immunological Reviews, 143(1), 63–88. \nKokot, M., Dlugosz, M., & Deorowicz, S. (2017). KMC 3: counting and manipulating k-mer statistics. \nBioinformatics (Oxford, England), 33(17), 2759–2761. \nKrokene, C., Anthonisen, K., Lifjeld, J. T., & Amundsen, T. (1996). Paternity and paternity assurance \nbehaviour in the bluethroat,. Animal Behaviour, 52(2), 405–417. \nKuznetsov, D., Tegenfeldt, F., Manni, M., Seppey, M., Berkeley, M., Kriventseva, E. V., & Zdobnov, E. M. \n(2023). OrthoDB v11: annotation of orthologs in the widest sampling of organismal diversity. \nNucleic Acids Research, 51(D1), D445–D451. \nLaetsch, D. R., & Blaxter, M. L. (2017). BlobTools: Interrogation of genome assemblies. \nF1000Research, 6, 1287. \nLi, H. (2013). Aligning sequence reads, clone sequences and assembly contigs with BWA-MEM. \nhttp://arxiv.org/abs/1303.3997 \nLi, H. (2017). Minimap2: pairwise alignment for nucleotide sequences. Bioinformatics , 34, \n3094–3100. \nLi, H. (2023). Protein-to-genome alignment with miniprot. Bioinformatics (Oxford, England), 39(1), \nbtad014. \nLi, H., & Durbin, R. (2024). Genome assembly in the telomere-to-telomere era. Nature Reviews. \nGenetics, 25(9), 658–670. \nLi, H., Handsaker, B., Wysoker, A., Fennell, T., Ruan, J., Homer, N., Marth, G., Abecasis, G., Durbin, R., \n& 1000 Genome Project Data Processing Subgroup. (2009). The Sequence Alignment/Map \nformat and SAMtools. Bioinformatics (Oxford, England), 25(16), 2078–2079. \nManni, M., Berkeley, M. R., Seppey, M., Simão, F. A., & Zdobnov, E. M. (2021). BUSCO update: Novel \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 30, 2026. ; https://doi.org/10.64898/2026.03.26.714473doi: bioRxiv preprint \n\nand streamlined workflows along with broader and deeper phylogenetic coverage for scoring of \neukaryotic, prokaryotic, and viral genomes. Molecular Biology and Evolution, 38(10), \n4647–4654. \nMathers, T. C., Paulini, M., Sotero-Caio, C. G., & Wood, J. M. D. (2025). MicroFinder: Conserved \ngene-set mapping and assembly ordering for manual curation of bird microchromosomes. In \nbioRxiv. https://doi.org/10.1101/2025.05.09.653066 \nMinias, P ., Pikus, E., Whittingham, L. A., & Dunn, P . O. (2019). Evolution of Copy Number at the MHC \nVaries across the Avian Tree of Life. Genome Biology and Evolution, 11(1), 17–28. \nNorling, M., Jareborg, N., & Dainat, J. (2018). EMBLmyGFF3: a converter facilitating genome \nannotation submission to European Nucleotide Archive. BMC Research Notes, 11(1), 584. \nOu, S., Su, W., Liao, Y ., Chougule, K., Agda, J. R. A., Hellinga, A. J., Lugo, C. S. B., Elliott, T. A., Ware, D., \nPeterson, T., Jiang, N., Hirsch, C. N., & Hufford, M. B. (2019). Benchmarking transposable \nelement annotation methods for creation of a streamlined, comprehensive pipeline. Genome \nBiology, 20(1), 275. \nPeona, V., Weissensteiner, M. H., & Suh, A. (2018). How complete are “complete” genome \nassemblies?-An avian perspective. Molecular Ecology Resources, 18(6), 1188–1195. \nQuestiau, S., Eybert, M. -C, & Taberlet, P . (1999). Amplified fragment length polymorphism (AFLP) \nmarkers reveal extra-pair parentage in a bird species: the bluethroat ( Luscinia svecica ). \nMolecular Ecology, 8(8), 1331–1339. \nRanallo-Benavidez, T. R., Jaron, K. S., & Schatz, M. C. (2020). GenomeScope 2.0 and Smudgeplot for \nreference-free profiling of polyploid genomes. Nature Communications, 11(1), 1432. \nRekdal, S. L., Anmarkrud, J. A., Johnsen, A., & Lifjeld, J. T. (2018). Genotyping strategy matters when \nanalyzing hypervariable major histocompatibility complex-Experience from a passerine bird. \nEcology and Evolution, 8(3), 1680–1692. \nRekdal, S. L., Anmarkrud, J. A., Lifjeld, J. T., & Johnsen, A. (2019). Extra-pair mating in a passerine bird \nwith highly duplicated major histocompatibility complex class II: Preference for the golden \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 30, 2026. ; https://doi.org/10.64898/2026.03.26.714473doi: bioRxiv preprint \n\nmean. Molecular Ecology, 28(23), 5133–5144. \nRhie, A., Walenz, B. P ., Koren, S., & Phillippy, A. M. (2020). Merqury: reference-free quality, \ncompleteness, and phasing assessment for genome assemblies. Genome Biology, 21(1), 245. \nStanke, M., Schöffmann, O., Morgenstern, B., & Waack, S. (2006). Gene prediction in eukaryotes with \na generalized hidden Markov model that uses hints from external sources. BMC Bioinformatics, \n7(1), 62. \nUliano-Silva, M., Ferreira, J. G. R. N., Krasheninnikova, K., Darwin Tree of Life Consortium, Formenti, \nG., Abueg, L., Torrance, J., Myers, E. W., Durbin, R., Blaxter, M., & McCarthy, S. A. (2023). \nMitoHiFi: a python pipeline for mitochondrial genome assembly from PacBio high fidelity reads. \nBMC Bioinformatics, 24(1), 288. \nUniProt Consortium. (2023). UniProt: The universal protein knowledgebase in 2023. Nucleic Acids \nResearch, 51(D1), D523–D531. \nWesterdahl, H. (2007). Passerine MHC: genetic variation and disease resistance in the wild. Journal of \nOrnithology, 148(S2), 469–477. \nWesterdahl, H., Mellinger, S., Sigeman, H., Kutschera, V. E., Proux-Wéra, E., Lundberg, M., \nWeissensteiner, M., Churcher, A., Bunikis, I., Hansson, B., Wolf, J. B. W., & Strandh, M. (2022). \nThe genomic architecture of the passerine MHC region: High repeat content and contrasting \nevolutionary histories of single copy and tandemly duplicated MHC genes. Molecular Ecology \nResources, 22(6), 2379–2395. \nZhou, C., McCarthy, S. A., & Durbin, R. (2023). YaHS: yet another Hi-C scaffolding tool. Bioinformatics \n(Oxford, England), 39(1), btac808. \n \nList of figures and tables: \nFigure 1. Sequenced specimen and genome profile \n \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 30, 2026. ; https://doi.org/10.64898/2026.03.26.714473doi: bioRxiv preprint \n\nFigure 2: Metrics of the genome assemblies of Luscinia s. svecica \n \nFigure 3. Visualization of the MHC region for both haplotypes on chromosome 35 \n \nFigure 4. Neighbor joining trees of bluethroat MHC alleles \n \nSupplementary Figure 1: Hi-C contact map of genome assemblies of Luscinia s. svecica \nhap1 and hap2. \n \nSupplementary Figure 2: BlobToolKit GC-coverage plots of genome assemblies of Luscinia \ns. svecica hap1 and hap2 \n \nSupplementary Figure 3. Alignment and distance-based clustering of MHCI coding \nsequences \n \nSupplementary Figure 4. Alignment and distance-based clustering of MHCII β coding \nsequences. \n \nSupplementary Figure 5. Neighbor joining tree of bluethroat MHCIIβ exon 2 alleles \n \nSupplementary Figure 6. Visualization of the MHC region for both haplotypes on \nchromosome 35, shown as a gene-level comparison. \n \nTable 1. Software tools: versions and sources \n \nTable 2: Genome data for Luscinia s. svecica. \n \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 30, 2026. ; https://doi.org/10.64898/2026.03.26.714473doi: bioRxiv preprint","source_license":"CC-BY-4.0","license_restricted":false}