Assessment of genetic diversity and genetic relationships of farm and laboratory quail populations in Japan using microsatellite DNA markers

Abstract Background The Japanese quail (Coturnix japonica) is an important poultry species owing to their high economic efficiency and biological advantages. The genetic diversity of farm quail populations has rarely been studied. Objectives This study aimed to assess the genetic diversity of farm quail populations and their genetic relationships, which could provide important information for designing breeding programmes to maintain egg and/or meat production efficiency. Methods Molecular phylogenetic and STRUCTURE analyses were conducted for seven farm populations and six laboratory lines using 50 microsatellite markers previously developed by us. Results The genetic diversity within each farm population was relatively high despite long‐term breeding within closed colonies. However, the genetic variation between populations was absent. Twenty highly polymorphic markers, selected based on Ne, He and FST values, enabled the construction of reliable phylogenetic trees and STRUCTURE plots. Conclusions In the farm populations analysed in the present study, gene flow between genetically distant populations is needed to restore genetic diversity between farm populations, which could exploit heterosis and decrease the risk of inbreeding depression. Our findings demonstrate that these markers are useful for examining the genetic structure of farm quail populations.

In Japan, their commercial use started at the beginning of the 20th century (Wakasugi, 1984), and commercial quail were first exported to the United States and Europe in the 1930s (Minvielle, 2004). Following a severe decline in Japan during World War II, domestic quail populations were re-established from a few surviving individuals (Wakasugi, 1984;Yamashina, 1961). The re-established commercial quail were again exported and rapidly distributed worldwide. Most domestic Japanese quail that are currently bred worldwide are thought to have been derived from the populations that were re-established in Toyohashi city, Aichi prefecture, Japan, after World War II (Wakasugi, 1984;Yamashina, 1961). Toyohashi is still a centre of quail egg production, accounting for approximately 60% of commercial egg production in Japan. Even though 4.62 million birds were raised on 83 farms in Toyohashi area in 1980, these numbers decreased to 1.76 million birds raised on 12 farms by 2018 (unpublished data from the Department of Agriculture, Forestry, and Fisheries, Aichi Prefectural Government). However, the average number of quail per farm has increased from around 56,000 in 1974 to around 147,000 in 2018. Furthermore, farmers in Aichi prefecture have generally preferred closed breeding within a colony, without introducing sire quail from other farms for many years. Japanese quail are known to be very susceptible to inbreeding depression (Shinjo, Mizuma, & Nishida, 1971;Sittmann, Abplanalp, & Fraser, 1966), raising concerns for farmers that productivity will decrease as a consequence of the long-term breeding within a closed colony.
One way to escape inbreeding depression is cross-mating with genetically distant quail, which is an effective method of restoring genetic diversity in farm quail populations. For this, monitoring the genetic diversity and/or autozygosity within populations, as well as the genetic variation between populations, are required for determining which populations should be introduced for cross-mating.
In our previous study (Nunome et al., 2017), genetic characterisation was carried out for 19 domestic populations of Japanese quail, consisting of nine laboratory lines, six meat-type quail lines, one commercial population and three wild quail populations from East Asia, using 50 highly variable microsatellite DNA markers.
A phylogenetic tree constructed with the markers clearly represented the breeding history of laboratory lines established in the last 50 years, indicating that the microsatellite markers are effective for estimating the genetic diversity of farm quail populations and their genetic relationships. Recently, Shimma and Tadano (2019) examined genetic diversity of 12 laying-type quail lines collected from nine farms in five prefectures in Japan, using 50 microsatellite markers. In the study, although four farms in Aichi prefecture showed genetic diversity as high as farm quail in the other prefectures, they could not find clear genetic differences among lines.
In this study, to assess the genetic diversity within and between farm populations of Japanese quail, we conducted genetic monitoring for six farm populations from Aichi prefecture, and five laboratory lines using 50 microsatellite markers that were developed previously (Tadano et al., 2014). Then, we performed molecular phylogenetic and STRUCTURE analyses for seven farm populations and six laboratory lines, including our previous data of one farm population from Saitama prefecture and one laboratory line (the wild-derived line) to find genetically distinct candidates that can be used for restoring the genetic diversity of the farm quail populations. Then, we assessed the effective number of highly polymorphic markers that are needed to accurately estimate the genetic structure of farm quail populations.

| Specimens and genomic DNA extraction
Whole blood samples were collected from wing veins of 499 individuals from the following 11 populations (Table S1): four laboratory quail lines (AARC-B, -C, -BB and -WW) from Aichi Agricultural Research Center (AARC) (AARC-B and AARC-C have been maintained for 25 generations and AARC-BB and AARC-WW for 15 generations); one laboratory line, DY, that derived from a dominant yellow mutant which was found in a farm of Toyohashi in 1963 (this mutant may have been mated with commercial quail that derived from an unknown farm and then has been maintained as a closed colony for nearly half a century); four farm populations from Toyohashi city (Farms A-D) and one population from each of two neighbouring cities (Toyokawa, Farm E; Tahara, Farm F); Genomic DNA was extracted from 20 µl of whole blood using 300 µl of DNAzol BD reagent (Molecular Research Center, Tokyo, Japan). For construction of phylogenetic trees and STRUCTURE analysis, the genotyping data of the following two quail populations that were reported in our previous study were included (Tadano et al., 2014): 57 commercial quail from a farm in Saitama prefecture and 40 quail from a wild-derived line that was established from wild-captured quail at the National Institute of Genetics, Japan, and is now maintained at the National Institute of Livestock and Grassland Science (NILGS), National Agriculture and Bio-oriented Research Organization, Japan.

| PCR amplification
PCR amplification of the 50 microsatellite markers used in our previous study (Nunome et al., 2017) was performed using a 10-µl reaction mix containing approximately 50 ng of genomic DNA, 10 pmol of each primer and 5 µl of Taq Gold 360 Master Mix (Thermo Fisher Scientific-Applied Biosystems). The following PCR cycling conditions were used: initial denaturation at 95°C for 10 min, followed by 42 cycles at 95°C for 30 s, 50°C or 55°C for 30 s, and 72°C for 25 s, and a final extension at 72°C for 5 min. The nucleotide sequences and suitable annealing temperatures of all the primer sets were provided in our previous study (Tadano et al., 2014). Amplicons were electrophoresed with Hi-Di formamide (Thermo Fisher Scientific) and the GeneScan 600 LIZ Size Standard (Thermo Fisher Scientific) using the ABI PRISM 3130 Genetic Analyzer (Thermo Fisher Scientific).
Allele size was determined using GENEMAPPER version 4.1 (Thermo Fisher Scientific).

| Estimation of genetic diversity within populations
The total number of alleles (NA) per marker was counted using MICROSATELLITE ANALYSER 4.05 (Dieringer & Schlötterer, 2003).
The mean number of effective alleles (Ne, the number of equally frequent alleles at a marker), observed heterozygosity (Ho), expected heterozygosity (He), fixation indices (F IS , F ST and F IT ) and the chi-square statistic for Hardy-Weinberg equilibrium (HWE) were calculated for each marker using GENALEX 6.5 (Peakall & Smouse, 2012). Polymorphic information content (PIC) was calculated using CERVUS 3.0.7 (Kalinowski, Taper, & Marshall, 2007). Null allele frequency (NAF) was estimated based on the Expectation Maximisation (EM) algorithm (Dempster, Laird, & Rubin, 1977) using the FreeNA software (Chapuis & Estoup, 2006). Allelic richness (AR) in each population was calculated using MICROSATELLITE ANALYSER 4.05 (Dieringer & Schlötterer, 2003). The genetic diversity of each population was estimated based on the mean genetic distance (MGD) between individuals within a population determined using the relatedness analysis conducted with GENEALEX 6.5 (Peakall & Smouse, 2012). The codominant genetic distance (Smouse & Peakall, 1999) was used to calculate the MGD and the upper and lower limits of the 95% confidence interval of the MGD were determined using 1,000 bootstrap replicates.

| Estimation of genetic relationships among populations
Pairwise genetic distances between populations were calculated using MICROSATELLITE ANALYSER 4.05 (Dieringer & Schlötterer, 2003) based on three different genetic distances, namely, Nei's angular genetic distance based on allele frequencies (Da; Nei, Tajima, & Tateno, 1983), the genetic distance based on the proportion of shared alleles (Dps; Bowcock et al., 1994) and the Cavalli-Sforza chord distance based on allele frequencies (Dc; Cavalli-Sforza & Edwards, 1967). The wild-derived line was used as an outgroup to reconstruct the phylogenetic trees. The bootstrap values were calculated with 1,000 replicates. Bayesian clustering analysis was performed to examine genetic relationships among 13 populations using STRUCTURE 2.3 (Pritchard, Stephens, & Donnelly, 2000). Log probability values (Ln P[D]) were estimated for each K from 1 to 14 with sampling periods of 100,000 MCMC generations after burn-in periods of 300,000 generations, using the admixture and the correlated allele frequency models (Porras-Hurtado et al., 2013). Twenty independent MCMC runs were performed for each K. Among 20 runs, those that showed obviously large values for the variance of Ln likelihood were discarded because MCMC simulations do not sufficiently converge on an optimal Ln likelihood value when the variance is large. The results were averaged using CLUMPP 1.1.2 (Jakobsson & Rosenberg, 2007) and graphed with DISTRUCT 1.1 (Rosenberg, 2004). The optimal K value was estimated using the Evanno method (Evanno, Regnaut, & Goudet, 2005), implemented in STRUCTURE HARVESTER 0.6.94 (Earl & vonHoldt, 2012).

| Selection of 20 informative microsatellite markers
First, phylogenetic trees were constructed for 13 populations with three genetic distances (Da, Dps, and Dc) using 48 markers, excluding two monoallelic markers (see the Results section), and we determined the genetic distance needed to construct the phylogenetic tree that reflected the breeding histories of the populations most accurately. Then, sets of the top 20 markers were selected for each Ne, He and F ST that satisfied the conditions of not significantly deviating from Hardy-Weinberg equilibrium (HWE) and presenting null allele frequencies less than 0.1 (the markers shown in bold in Table S1).
The molecular phylogenetic trees were constructed with each set of 20 markers using Cavalli-Sforza chord distance based on allele frequencies (Dc) (see the details in the Results section). STRUCTURE analysis was also performed using three sets of 20 markers, and the results were compared with those generated using 48 markers.

| Genetic diversity of quail populations
Two markers, NGJ0023 and NGJ0045, were monoallelic (Table S1), and the remaining 48 were used for subsequent population genetic analyses. The highest number of alleles was observed for NGJ0050 (NA = 23) and the lowest for NGJ0007, NGJ0022 and NGJ0044 (NA = 3) ( Table S1). The Ne was the highest for NGJ0050

| Genetic relationships between quail populations
Three phylogenetic trees for 13 populations based on Da (Figure 1a STRUCTURE HARVESTER analysis indicated that the highest Delta K value was 2 ( Figure 1d); however, the STRUCTURE plot at K = 2 only separated farm and farm-derived populations (Farms A-F, DY and Saitama) and the rest (four AARC lines and the wild-derived line) (data not shown), which was not informative for our research. Therefore, we constructed the STRUCTURE plot with the second highest K value (K = 5) because it was more informative. All

| Genetic relationships of quail populations estimated using 20 selected markers
Twenty highly polymorphic markers that were selected based on the Ne and He values were the same (Table S1)

| Genetic relationships and breeding histories of farm and laboratory quail populations
The STRUCTURE analysis of 15 laboratory lines in our previous study (Nunome et al., 2017), including the NIES-Br and the wild-derived lines, showed a high genetic affinity between the NIES-Br and wildderived lines; this may explain why the AARC-BB line is also closely related to the other three AARC laboratory lines (AARC-B, AARC-C and AARC-WW) that share a common origin from the wild-derived line.
The phylogenetic trees and STRUCTURE plots constructed with the 48 markers clearly supported the breeding histories of the farm and laboratory quail populations examined in this study, even though null allele frequencies were higher than 0.20 for five of the markers.
In particular, the tree based on the Dc genetic distance matched better to known histories than those based on the Da and Dps genetic distances. The positions of the DY line and Saitama population and the monophyly of the six farm populations from Aichi prefecture on the Dc tree indicated that the Dc tree is more reliable. In our previous study on the genetic characterisation for laboratory chicken lines, the Dps genetic distance constructed less realistic phylogenetic trees than that of Da genetic distance (Nunome et al., 2019). The Dc and Da genetic distances are known to be better estimators for topology construction (Takezaki & Nei, 1996). Recently, a simulation study by Sere, Thevenon, Belem, and De Meeûs (2017) confirmed that the Dc genetic distance is the most robust, even in the presence of null alleles, which is similar to that observed by Carlsson (2008), who revealed that STRUCTURE analysis also produces reliable results in the presence of null alleles. As to the STRUCTURE analysis, 300,000 MCMC generations were discarded as burn-in periods in every run, which was three times larger than the length of burn-in periods in Shimma and Tadano (2019

| Validity of 20 microsatellite markers for genetic assessment
Microsatellites are one of the most popular markers for estimating the genetic diversity in animal and plant populations. Microsatellite markers for Japanese quail were developed comprehensively in previous studies (Kayang et al., 2000(Kayang et al., , 2002(Kayang et al., , 2004Mannen et al., 2005). They have been used to discover gene loci associated with plumage colour, body weight, meat quality and egg production (Minvielle et al., , 2006Miwa et al., 2005

| CON CLUS IONS
Japanese quail are expected to become more widespread in developing countries as a useful source of animal protein. When breeding programmes are formulated to improve egg and/or meat production efficiency of poultry, genetic assessment of farm populations will be important to avoid the loss of genetic diversity or the accumulation of homozygosity that could result in decreased productivity. In this study, using the 50 microsatellite markers, we performed a range of molecular phylogenetic and STRUCTURE analyses that clearly demonstrated the breeding histories and genetic diversities of both the laboratory lines and farm populations of quail from the Aichi prefecture, Japan. The two laboratory quail lines, the AARC-BB and AARC-WW, were found to be genetically relatively distant from the farm populations. Based on our data, the Aichi Agricultural Research Center has started to introduce the two laboratory lines into farm populations in Toyahashi city.
Furthermore, 20 highly polymorphic microsatellite markers selected from among the 48 markers also accurately estimated the genetic diversity of the quail populations and their genetic relationships, indicating that these microsatellite markers are useful for the genetic assessment of farm quail populations.

Peer Review
The peer review history for this article is available at https://publo ns.com/publo n/10.1002/vms3.328.