The quail as an avian model system: its genome provides insights into social behaviour, seasonal biology and infectious disease response

The Japanese quail (Coturnix japonica) is a popular domestic poultry species and an increasingly significant model species in avian developmental, behavioural and disease research. We have produced a high-quality quail genome sequence, spanning 0.93 Gb assigned to 33 chromosomes. In terms of contiguity, assembly statistics, gene content and chromosomal organization, the quail genome shows high similarity to the chicken genome. We demonstrate the utility of this genome through three diverse applications. First, we identify selection signatures and candidate genes associated with social behaviour in the quail genome, an important agricultural and domestication trait. Second, we investigate the effects and interaction of photoperiod and temperature on the transcriptome of the quail medial basal hypothalamus, revealing key mechanisms of photoperiodism. Finally, we investigate the response of quail to H5N1 influenza infection. In quail lung, many critical immune genes and pathways were downregulated, and this may be key to the susceptibility of quail to H5N1. This genome will facilitate further research into diverse research questions using the quail as a model avian species.


INTRODUCTION
Japanese quail (Coturnix japonica) is a migratory bird indigenous to East Asia and is a popular domestic poultry species raised for meat and eggs in Asia and Europe. Quail have been used in genetics research since 1940 (Shimakura 1940), and are an increasingly important model in developmental biology, behaviour and biomedical studies (Minvielle 2009). Quail belong to the same family as chickens (Phasianidae) but have several advantages over chickens as a research model. They are small and easy to raise, have a rapid growth rate and a short life cycle, becoming sexually mature only seven to eight weeks after hatching (Huss et al. 2008). The quail embryo survives manipulation and culture better than chicken embryos making them ideal for this type of research (Huss et al. 2008). Quail have been used as a model for stem cell differentiation, for example a culture system that mimics the development of hematopoietic stem cells has been recently developed, as quail show greater cell multiplication in these cultures than chickens (Yvernogeau et al. 2016). Quail are also used to study the genetics underlying social behaviours (Mills et al. 1997 We have produced a high-quality annotated genome of the Japanese quail (Coturnix japonica), and herein describe the assembly and annotation of the quail genome and demonstrate key uses of the genome in immunogenetics, disease, seasonality and behavioural research demonstrating its utility as an avian model species. 6

Genome assembly and annotation
We sequenced a male Coturnix japonica individual from an inbred quail line using an To improve the quantity and quality of data used for the annotation of the genome, we sequenced RNA extracted from seven tissues sampled from the same animal used for the genome assembly. Using the same inbred animal increases the alignment rate and accuracy. 7 The amount of data produced for annotation from the 7 tissues is (in Gb): 18.9 in brain, 35 For further annotation, a set of genes unnamed by the automated pipeline were manually annotated. As part of an ongoing project to investigate hemogenic endothelium commitment and HSC production (Yvernogeau et al. 2016), transcriptomes were produced for two cultured cell fractions. Study of these cells is critical for developmental biology and regenerative medicine, and quail are an excellent model for studying these as they produce much more hematopoietic cells than similar chicken cultures. Approximately 8,000 genes were expressed in these cells lines which lacked gene names or annotation from the automated annotation pipeline. Using BLAST (Altschul et al. 1990) searches to identify homology to other genes, 3,119 of these were manually annotated (Supplemental File 1).
Genome completeness was also quantitatively assessed by analyzing 4,915 single copy, orthologous genes derived from OrthoDB v7 and v9   Figure S1). The quail genome has 10 more missing and 23 more fragmented genes than the Gallus gallus 5.0 assembly. However, relative to the total number of genes in the benchmarking set, these increases amount to just 0.2% and 0.5%, respectively. This indicates that the quail genome, like the chicken genome, is highly contiguous and in terms of its expected gene content, is close to complete. 1 All species-specific assembly metrics derived from the NCBI assembly archive.

Galliforme genome synteny
Comparative mapping of the quail and chicken genomes revealed a high conservation of the chromosomal arrangement ( Orthologous genes between quail and closely related species were identified through reciprocal BLAST searches. One-to-one orthologs in chicken were identified for 78.2% of all quail genes and 91.8% of protein-coding quail genes (Supplemental Table S1), indicating a high degree of genic conservation in the quail genome. Fewer orthologs were seen between turkey and quail genes (69.3%), although the number of orthologs of protein-coding genes was similar (91.7%), so the discrepancy is likely due to missing non-coding gene predictions in the turkey genome. As expected, conservation of one-to-one orthologs was lower with the mallard duck (Anas platyrhynchos), with duck orthologs identified for 64.5% of quail genes (78.9% protein-coding genes).  Table S2), and similarly analysed passerine birds (Mason et al. 2016).
The majority of ERV sequences in all three genomes were short and fragmented, but 393 intact ERVs were identified in the quail, most of which were identified as alpha-, beta-or gamma-retroviral sequences by reverse transcriptase homology. It is possible that the smaller genome size of the quail compared to other birds reflects a more limited expansion of ERVs and other repeats (such as the LINE CR1 element; Supplemental Table S2 Despite variation in total and intact ERV content, the overall genomic ERV distribution in these three gallinaceous birds was highly similar. ERV sequence density was strongly correlated with chromosome length on the macrochromosomes and Z chromosome (r > 0.97; P < 0.001), but there was no significant correlation across the other smaller chromosomes.
Furthermore, ERV density on each Z chromosome was at least 50% greater than would be expected on an autosome of equal length. These results support the depletion of repetitive elements in gene dense areas of the genome, and the persistence of insertions in poorly recombining regions, as was seen in the chicken (Mason et al. 2016). This is further supported by the presence of clusters of intact ERVs (where density was five times the genome-wide level) on the macrochromosomes and sex chromosomes (Supplemental Table   S2).

Immune gene repertoire
We investigated the immune genes in the quail genome in detail due to the importance of quail in disease research. The MHC-B complex of the quail has been previously sequenced and found to be generally conserved compared to chicken in terms of gene content and order Several genes are thought to be crucial for influenza resistance in both humans and birds, including RIG-I, TLR and IFITM genes. RIG-I has not previously been identified in chicken, despite being present in ducks and many other bird orders, and is considered highly likely to be deleted from the chicken genome (Barber et al. 2010). In addition, an important RIG-I binding protein RNF135 has also not been identified in chicken (Magor et al. 2013).
Likewise, an ortholog of RIG-I or RNF135 could not be identified in the quail genome or transcriptomes through BLAST searches and therefore is likely missing in the quail also.
Orthologs of all five chicken IFITM genes (IFTIM1, 2, 3, 5 and 10) were identified in the quail genome and transcriptomes. In addition, orthologs of each chicken TLR, including key TLRs for viral recognition, TLR4 and TLR7, were identified in the quail genome, except that of TLR1A.

Selection for social motivation
Quail has been used as a model to study the genetic determinism of behaviour traits Further experiments will be required to examine the possible functional link between the selected genes and the divergent phenotype observed in these lines. Also, by analyses of genes known to be differentially expressed in the zebra finch during song learning we hope to comparatively understand molecular systems linked to behaviour in the avian brain.

A model for avian seasonal biology
Quail is an important model for studying seasonal biology. Seminal work in quail established that pineal melatonin (Ralph et al. 1967;Lynch 1971) is regulated by the circadian clock 1 8 (Cockrem and Follett 1985). In mammals, photo-sensing is dependent on a single retinal photoreceptor melanopsin (OPN4) that regulates pineal melatonin release. Nocturnal melatonin is critical for mammalian neuroendocrine response to photoperiod and is likely to target melatonin receptors in the pars tubularis (PT; Wood and Loudon 2014). Birds have a distinct non-retinal mechanism for photoreception through deep-brain photoreceptors (Menaker 1968) and melatonin does not appear to be critical for most avian seasonal cycles interact to determine the MBH transcriptome ( Fig. 2A).
We identified 16 temperature dependent DEGs with a large modulating effect of temperature (log 2 FC>1) (Fig. 2E). With the exception of aldehyde dehydrogenase (ALDH1A1), the temperature-dependent photoperiod effected DEGs were down-regulated in LD. There was an equal division of genes between temperature dependent amplification and suppression of LD down-regulated genes.
The MBH shows strong TSHβ induction in LD (Fig. 2C-D . We also noted down-regulation of the neuronally important GPCR GPR20 (Fig. 2G). In mice, deficiency of GPR20 is associated 2 0 with hyperactivity and may play a role in cAMP-dependent mitogenesis (Hase et al. 2008).
There was a strong enrichment of collagen biosynthetic processes and extracellular matrix organisation processes (Fig. 2F) and a large body of genes associated with cell differentiation and development (Fig. 2H).
We observed photoperiod-dependent regulation of a single clock gene, CRY4. CRY4 is upregulated in LP (log 2 FC=0.85 at 23°C, 1.37 at 9°C). This is consistent with the finding of We detected photoperiod effects on OPN4 transcripts, which were up-regulated in LD.
Photoperiod-dependent expression in OPN4 may well play a role in the photoperiodrefractory response. Encephalopsin (OPN3) was found to be highly expressed in the MBH In conclusion, we confirm the importance of temperature and photoperiod-dependent regulation of thyroid hormone metabolism in the avian MBH (Fig. 3). Temperaturedependent amplification and suppression of the photoperiod response may indicate qualitative differences in the MBH pathways or simply reflect different stages of progression through seasonally phased processes. This could be further investigated by contrasting across time series at different temperatures. We also observed concurrent regulation of multiple hormonal signalling pathways, this may reflect a diversity of pathways and cell types in the MBH or reflect a corrective mechanism to account for cross-talk with other GPCR pathways. We   To provide an overview of the response to LPAI and HPAI in quail we examined pathway and GO term enrichment of DEGs (see Supplemental File S9, Supplemental File S10 and Supplemental Figures S4-7). In response to LPAI infection, pathways enriched in the ileum included metabolism, JAK/STAT signalling, IL6 signalling and regulation of T-cells (Supplemental Figure S4). In the lung, pathways upregulated included complement, IL8 signalling, and leukocyte activation (Supplemental Figure S5). In the lung at 3dpi highly enriched GO terms included 'response to interferon-gamma', 'regulation of NF-kappaB', 2 5 'granulocyte chemotaxis' and 'response to virus' (Supplemental Figure S6), which are key influenza responses. This indicates an active immune response occurs to LPAI infection in quail, involving both ileum and lung, but with the strongest immune response occurring in the lung.
Genes upregulated in response to HPAI in the ileum were related to metabolism and transport, while inflammatory response was downregulated at 1dpi (Supplemental Figure   S6). Downregulated pathways at 1dpi included IL-6, IL-9 and neuro-inflammation signalling pathways (Supplemental Figure S6). In the quail lung many genes were downregulated after To compare the response of quail, duck and chicken, clustering of gene counts was examined using BioLayout 3D (Theocharidis et al. 2009). This revealed a cluster of 189 genes that were strongly upregulated at 1dpi in the duck, which showed no response in chicken and quail (Supplemental Table S3). This cluster was dominated by RIG-I pathway and IFN response genes including IFNG, DDX60, DHX58, IRF1, IRF2, and MX1. Pathways associated with this cluster includes MHCI processing and death receptor signalling (Fig. 4).
Thus, the lack of this early anti-viral response may be key to the susceptibility of Galliformes to HPAI. 2 6 To further compare the responses between the three species, enrichment of pathways in each species was examined (Fig. 5). This revealed very few commonly regulated pathways between the three species. However, at 1dpi in the ileum and 3dpi in the lung there were many pathways that were downregulated in the quail, not altered in chicken, and upregulated in the duck. In the ileum at 1dpi, this included pattern recognition and death receptor signalling. In the lung at 3dpi this involved host of immune related pathways including production of NOS by macrophages, pattern recognition, B and T cell signalling and NK-KB, IL8 and IL2 signalling.
The proportion of genes commonly regulated between quail, chicken and duck to HPAI infection was also examined (Fig. 6). Consistent with the heatmap comparison (Fig. 5), the response of chicken, quail and duck were largely unique, with few genes commonly differentially expressed. There was a large set of genes that were upregulated in duck, while being downregulated in quail at 3dpi, in both ileum and lung. In lung these genes were related primarily to innate immune system pathways, including pattern recognition pathways, cytokine production, leukocyte adhesion, TNF production, interferon production, B cell signalling and response to virus (Supplemental File S10). Genes with the greatest differential expression included RSAD2 which inhibits viruses including influenza, IFIT5 which senses viral RNA and OASL which has antiviral activity. These differences further highlight that the anti-viral immune response is dysregulated in quail. Additionally in both ileum and lung the apoptosis pathway was enriched in duck, but not quail (Supplemental File S10). Apoptosis is known to be a critical difference in the response of chickens and ducks to HPAI infection (Kuchipudi et al. 2010).
Lastly, we examined the response of key families involved in influenza and immune response, focussing on the lung (Supplemental Table S4 and Supplemental File S11).
IFTIM genes have previously been found to have a crucial role in HPAI resistance (Smith et Consistent with previous findings in the chicken (Smith et al. 2015), quail showed no significant upregulation of IFTIM genes, while these genes in duck were strongly upregulated, (Supplemental Table S4). TLRs and MHC receptors are involved in recognition of foreign molecules and triggering either an innate (TLR) or adaptive (MHC) immune response. TLR3, 4 and 7, which bind viral RNAs, were upregulated in response to LPAI in quail. A reversal was seen in response to HPAI, with TLR4, and 7 substantially downregulated. Likewise, genes of both MHC class I and II were upregulated in response to LPAI and downregulated in response to HPAI. By comparison there was no perturbation of TLR and MHC genes in chicken and upregulation of class I genes in duck. The quail seems to have a highly dysfunctional response to HPAI infection with key innate and adaptive immune markers downregulated at 3dpi, which contrasts with the strong immune response mounted by the duck and minimal immune response in the chicken.
In conclusion, we found that quail have a robust immune response to infection with LPAI, allowing them to survive the infection. However, they show dysregulation of the immune response after infection with HPAI, and this may explain their susceptibility to HPAI strains. IFITM response was not seen against HPAI while genes associated with apoptosis were downregulated, potentially allowing the virus to easily enter cells and spread early in infection. Antiviral and innate immune genes, including those involved in antigen recognition, immune system activation, and anti-viral responses were downregulated at 3dpi, which would prevent an effective immune response and viral clearance once infection is established. This study provides crucial data that can be used to understand the differing response of bird species to AIV, which will be critical for managing and mitigating these diseases in the future.

DISCUSSION
Here we describe the assembly, annotation and use of a high-quality quail genome, an important avian model in biological and biomedical research. This genome will be crucial for future avian genome comparative and evolutionary studies, and provides essential genetic and genomic reference information, molecular information for making precise primers and 3 1 nucleic acid probes, and accurate perturbation reagents including morpholinos, RNA inactivation tools, and CRISPR-Cas9 constructs. We have demonstrated the utility of this genome in both infectious disease and behavioural research providing further confirmation of the importance of quail as a research model, and for its role in agricultural and animal health studies. Specifically, the availability of this genome has allowed us to make significant discoveries in the unique response of quail to highly pathogenic avian influenza infection, helping elucidate the basis for extreme susceptibility seen in this species. It has also allowed us to identify and confirm genes and genomic regions associated with social behaviour, showing many similarities to genes associated with autism in humans and thus represents a possible biomedical model for autism. Furthermore, we have shown that genome-wide transcriptomics using this genome facilitated further insights and hypothesis into the mechanism of photo-periodism in avian seasonal biology. Moving forward, the availability of a high-quality quail genome will facilitate the study of diverse topics in both avian and human biology, including disease, behaviour, comparative genomics, seasonality and developmental biology.

Whole Genome Sequencing and Assembly
To facilitate genome assembly by avoiding polymorphism, we produced an individual as inbred as possible. We started with a quail line previously selected for early egg production and having a high inbreeding coefficient (Minvielle et al. 1999 Finally, all contaminating contigs identified by NCBI filters (alignments to non-avian species at the highest BLAST score obtained), and all contigs < 200 bp were removed prior to final assembly submission.

Gene Annotation
Specific RNA-seq data for the genome annotation was produced from the same animal used for the genome assembly. RNA was extracted from heart, kidney, lung, brain, liver, intestine, and muscle using Trizol and the Nucleospin® RNA II kit (MACHEREY-NAGEL), following the manufacturer's protocol. 3 3 The Coturnix japonica assembly was annotated using the NCBI pipeline, including masking of repeats prior to ab initio gene predictions, for evidence-supported gene model building. We utilized an extensive variety of RNA-Seq data to further improve gene model accuracy by alignment to nascent gene models that are necessary to delineate boundaries of untranslated regions as well as to identify genes not found through interspecific similarity evidence from other species. A full description of the NCBI gene annotation pipeline was previously described (Thibaud-Nissen et al. 2013). Around 8,000 lacked gene symbols from this pipeline, and these were further annotated manually by using BLAST searches using the corresponding sequences and extracting protein names from Uniprot.

Comparative analyses
A set of single copy, orthologous, avian-specific genes were selected from OrthoDB v.

Sociability selection study
The data and methods used have been described previously (Fariello et al. 2017

DATA ACCESS
All data generated or analysed during this study are included in this published article (and its additional files), or in the following public repositories. Data has been submitted to the public databses under the following accession numbers: genome sequence data, NCBI Genome (https://www.ncbi.nlm.nih.gov/bioproject/?term=PRJNA490454). This quail reference genome will be available on the Ensembl platform as part of release 96.

DISCLOSURE DECLARATION
The authors declare that they have no competing interests