All the procedures and animal works were conducted according to the guidelines for the care and use of experimental animals established by the Ministry of Agriculture and villages of China and were approved by the Ethics and Experimental Animal Committee of Northeast Agricultural University, China.

All the individuals (31♂, 63♀) from the same generation of the Min pig conserved population were selected from the National Conservation Farm for Min pig in Lanxi County, Heilongjiang Province, China. We traced 4 generations of pedigree records of these individuals.

Ear tissue samples were collected and placed in cryopreserve tubes for rapid preservation in liquid nitrogen. Genomic DNA was extracted by phenol-chloroform extraction method and stored at −20°C. The quality of DNA was determined by ultraviolet spectrophotometry (NanoDrop 2000; Thermo Scientific, ShangHai, China), and gel electrophoresis (DYY-6C & DYCP-31BN; Beijin Liuyi, Beijing, China). The DNA with OD value between 1.8 and 2.0 was diluted to a final concentration of 20 ng/μL.

The qualified DNA samples were genotyped using KPS Porcine Breeding Chip v1 50K (Compass Biotechnology Co., Ltd., Beijing, China), which is a chip designed based on the application of genome selection breeding. The chip meets the requirements of genome selection breeding for the chip that “involving as many traits as possible, distribute evenly on the genome, which has high polymorphism and strong specificity”. In the chip design, the single nucleotide polymorphism (SNP) variation loci related to the growth, reproduction and meat quality traits of Chinese local pig breeds were obtained by resequencing technology. The credibility of SNP loci was scored according to the requirements of minor allele frequency (MAF) and genome average distribution. Finally, the chip was customized by Illumina optical fiber microbead technology, and a total of 51,315 SNP was contained.

To reveal the fine population structure of Min pig from a global perspective, we extracted the common sites of KPS porcine breeding chip v1 (including Min pig, Hebao pig, Dapulian pig, and Laiwu pig) and Illumina 60K SNPs chip (including 3 European commercial pigs, 1 Asian wild boar and 6 Asia local pigs) according to the chromosome information and physical location information of SNPs loci, which were common SNP sites in pigs. A total of 23,708 sites were extracted for follow-up analysis [9,10]. The above quality control on the combined SNP data was performed by PLINK v1.90 software [11]. Quality control of the genotypic data was conducted with the following criteria: i) retaining the SNPs located on autosomes; ii) filtering out individuals with a call rate below 0.90; iii) removing the SNPs with a call rate less than 0.90; iv) removing the SNPs with a MAF below 0.01; and v) removing the SNPs with a Hardy Weinberg equilibrium p value less than 1×10⁻⁶.

Based on the genome-wide SNP quality control data, we used PLINK v1.90 software to calculate the MAF, the percentage of polymorphic marker ratio (PN), the observed heterozygosity (Ho) and the expected heterozygosity (He). We estimated the effective population content (Ne) based on the level of linkage disequilibrium [12,13]:

where r² is estimated by the linkage disequilibrium decline model, Ne represents the effective population content, and c represents the Morgan distance between SNP sites [14]. In this study, inter-SNP distances (kb) were binned into the following classes: 0.5, 1, 2, 5, 10, and 100 Mb, the Ne of the Min pig conserved population in the corresponding 200, 50, 20, 10, 5 generations and the current generation was estimated, where 1 cm is approximately equal to 1 Mb.

Runs of homozygosity (ROH) for each individual were estimated using “- homozyg” of PLINK v1.90 software. The default parameter –homozyg were used to define ROH and the following criteria were chosen: i) 50 SNPs were contained in each sliding window; ii) an ROH consisted of no less than 30 consecutive SNPs; iii) the density should be higher than one SNP per 1,000 kb; iv) SNPs with missing genotypes less than one and with a heterozygous genotype less than one were allowed in an ROH due to genotyping error. All homozygous segments filtered were classified into three length classes: 1 to 5, 5 to 10, and >10 Mb, identified as ROH 1 to 5 Mb, ROH 5 to 10 Mb, and ROH>10 Mb, respectively.

Pedigree-based inbreeding coefficients (F_PED) for all pigs were estimated using R package pedigree. Then, genomic inbreeding coefficient (F_ROH) of all individuals was estimated by PLINK v1.90 software, which was calculated according to the following formula [15]:

Where L_ROH was defined as the total length of the genome covered by all ROH segments for each individual, and L_AUTO was the length of the sequenced genome, which equaled 2,450,462.292 kb in this research. For each animal, four ROH estimates were calculated based on lengths from sequence data as the proportion of its genome: ROH>10 Mb (F_{ROH>10 Mb}), 5 to 10 Mb (F_{ROH 5–10 Mb}), 1 to 5 Mb (F_{ROH 1–5 Mb}), and ROH>1 Mb (F_{ROH>1 Mb}), corresponding to 5 generations, 5 to 10 generations, 10 to 50 generations, and 50 generations, respectively. In addition, SNP-based inbreeding coefficients were estimated using the option –ibc from the GCTA software [16]. The inbreeding coefficients obtained by the methods were compared using Pearson’s correlation.

The ancestral pedigree composition of all Min pigs was subdivided by ADMIXTURE analysis. When K = 2, clarified the proportion, average and standard deviation of individuals Chinese lineage in the Min pig conserved population. To select the nucleus conservation and the candidate conservation population, we divided the proportion values into four categories, including those greater than the mean value (Min1), between the mean and the mean minus one standard deviation (SD) (Min2), between the mean minus two SDs (Min3), and less than the mean minus two SDs (Min4).

We genotyped and quality controlled 94 Min pigs with the KPS Porcine Breeding Chip v1 50K. A total of 51,315 SNPs loci were detected, and the sample call rate was 98.38% to 98.84%, with an average of 98.66%. Among that, 44,739 SNPs were identified as autosomal. A total of 1,383 SNPs were deleted because the call rate was less than 0.90, 11,202 SNPs were deleted because the MAF value was less than 0.01, and 460 SNPs were filtered because they significantly deviated from the H-W equilibrium with the p value less than 1×10^–6. Finally, a total of 31,694 SNPs data that met high-quality control standards were obtained. After integrating SNP data with Illumina 60K (52,556 SNPs), 23,708 SNPs data from 224 pigs obtained by the same quality control standard were used for the following analyses.

The MAF distribution of all SNPs was calculated by PLINK v1.90 software. As shown in Figure 1, the MAF was distributed in each interval, with an average of 0.185. The percentage of PN in the Min pig conserved population was 0.663, the observed heterozygosity (Ho) was 0.335, and the expected heterozygosity (He) was 0.330. The Ho was slightly higher than the He.

Based on the autosomal genome information of the Min pig conserved population, we calculated the linkage disequilibrium (r²) between all SNPs pairs of each chromosome, and then evaluated the effective population content (Ne). In this study, a total of 2,514 million pairs of SNPs linkage disequilibrium estimates were obtained, inter-SNP distances (kb) were binned into the following classes: 0.5, 1, 2, 5, 10, and 100 Mb, and historical Ne was estimated from 200 to 0 generations ago. The Ne of the Min pig conserved population showed a continuously decreasing trend. The rate of decline was the slowest from 200 to 50 generations ago (r = 0.95), then accelerated slowly from 50 to 5 generations ago (1.40<r<1.50) and increased significantly in the last 5 generations (r = 2.6) (Figure 2).

The abundance and genomic distribution of ROH provides efficient information about the demographic history of livestock species. A total of 5,456 ROH were obtained from 94 individuals by PLINKv1.90 software, with an average of 58 ROH per individual. The total length of ROH was 507.83 Mb, the mean ROH length was 9.31 Mb, and all ROH segments account for 22.05% of the whole genome. The distribution of ROH according to length is shown in Figure 3. For chromosomes, the number of ROH per chromosome and the percentage of chromosomes covered by ROH are shown in Figure 4. The highest number of ROH per chromosome was on SSC1 (598 segments), accounting for 23% of the chromosome, whereas the lowest was on SSC18 (129 segments), accounting for 17% of the chromosome. Among them, the longest ROH segment was on SSC13 (132.9 Mb), containing 2,474 SNPs, and the shortest ROH segment was on chromosome SSC15 (1.13 Mb), containing 55 SNPs.

The descriptive statistics of ROH number and length by classes are given in Table 1. Among the different classes of ROH segments, the majority were short ROH segments (1 to 5 Mb), accounting for 45.16% of the total number of ROH segments, ROH segments (5 to 10 Mb) made up 29.80% of the total ROH length, the number of long ROH segments was the lowest, but it still had a coverage ratio of 25.04%.

The inbreeding coefficient of the Min pig conserved population estimated by pedigree, ROH of different lengths, and the SNP are shown in Table 2. The average inbreeding coefficient estimated by pedigree information was F_PED = 0.015 ±0.032, and the average inbreeding coefficients estimated by ROH of different lengths were F_{ROH > 10 Mb} = 0.133±0.038, F_{ROH 5–10 Mb} = 0.049±0.011, F_{ROH 1–5 Mb} = 0.037±0.007, F_{ROH>1 Mb} = 0.220±0.037. And the average inbreeding coefficient estimated by SNP was F_SNP = −0.014±0.040. Though there were obvious differences among F_PED, F_ROH, and F_SNP, there was also a certain correlation between them. Among them, the Pearson correlation between F_{ROH>1 Mb} and F_SNP was the highest and significant (p<0.05), while the Pearson correlation between F_{ROH>1 Mb}, F_SNP, and F_RED was not significant (p>0.05).

The proportion of Chinese lineage in Min Pig conserved population was analyzed by ADMIXTURE v1.30. The results showed that the average proportion of Chinese lineage in Min pig population was 0.70±0.04 (Supplementary Table S1). According to the proportion of Chinese lineage, the nucleus conservation (greater than mean minus one SD) and candidate conservation population (less than mean minus one SD) were 81 and 13, respectively (Figure 6).

The genetic diversity analysis is very important for the protection of the population. Through different indicators of genetic diversity, we can evaluate the current state of the population. In this study, the whole-genome high-density SNP markers were used to evaluate the conservation status of Min pig for the first time. The heterozygosity analysis results elucidated that the expected heterozygosity (He) of the Min pig conserved population was 0.330, which was lower than Jinhua pig (He = 0.429) [18], Laiwu pig (He = 0.380) [19], Erhualian pig (He = 0.378) and Meishan pig (He = 0.382) [20]. The low expected heterozygosity of the Min pig conserved population indicates that many SNPs loci tended to be pure and may be undergoing inbreeding. In addition, the observed heterozygosity (Ho) of the Min pig conserved population was 0.335, which was slightly higher than that of He, indicating that there may be differentiation or introduction of foreign lineage in the history of the evaluated population.

Effective population size (Ne), as another important indicator of genetic diversity, plays an important role in the process of animal genetic breeding [21]. The result of this study showed that the decline rate of the Ne of Min pig was the slowest from 200 to 50 generations ago, then accelerated slowly, and increased significantly in the last 5 generations. We indicated that many European commercial pig breeds have been introduced since the construction of the Middle East Railway in the early 20th century, resulting in genetic introgression of Min pig. With the establishment of the Min pig National Conservation Farm in the 1970s, genetic introgression was blocked. However, due to the founder effect, the Ne of small-scale Min pig population has declined slowly since the 50th generation. In the past 5 generations, the inbreeding of Min pig may have increased due to poor management of breeding farms or other factors, which has led to a sharp increase in the decline rate of their Ne. In addition, some studies have confirmed that if the Ne value of a small population is less than 65, it means that the population is endangered [22]. The Ne of the Min pig conserved population is gradually decreasing, with a Ne value of 19 in the last 5 generations, which is in a threatened state. In the long term, it may lead to increase inbreeding and even genetic drift in the population. Therefore, the conservation farm should consider the ratio of male and female individuals and the lineage compositions of pigs to ensure the dynamic balance of the conservation population as much as possible.

Defining the inbreeding coefficient of the population also plays an important role in conservation work. The reduction of genetic diversity can be avoided by controlling inbreeding. Traditional estimation of inbreeding coefficient is based on pedigree, however, since it is incomplete, it is more accurate to use genomic information to estimate the inbreeding coefficient [23]. Among them, ROH has become a popular method to evaluate the inbreeding of populations [24,25]. Different length fragments of ROH can distinguish the inbreeding history of ancient and recent, respectively [26,27]. In this study, the FROH value of the Min pig conserved population has continued to increase from 50 generations ago, which directly reflects that the Min pig conserved population is undergoing inbreeding and has had an impact on the genome of Min pig, which should be taken seriously. Among them, the F_{ROH 1–5 Mb} value was the lowest, which could also reflect the possible genetic introgression of exotic pig breeds in the process of domestication. The overall inbreeding coefficient (F_{ROH>1 Mb}) value was high, which may be due to the genetic relationship among individuals of Min pig, and the influence of pedigree effect on inbreeding analysis.

Since the beginning of the 20th century, a large number of European commercial pigs have been introduced into China, crossed with indigenous breeds and pose a serious threat to Chinese indigenous pig breeds [28]. In this study, the results of Admixture analysis clearly showed that the European commercial pigs had historically flowed to the Min pig conserved population, which led to the mixed European lineage in Min pig. Among the 94 Min pigs, 81 Min pigs have a Chinese lineage ratio greater than mean minus one SD (66.0%). It is suggested that 81 Min pigs should be taken as the nucleus conservation population, then the remaining 13 Min pigs whose Chinese lineage ratio less than mean minus one SD (66.0%) are suggested as the candidate conservation population. During the conservation process, family rotation mating is adopted to ensure 90% genetic diversity of livestock. Future research will focus on the effects of the introgression of European commercial pig breeds on the traits of Min pig, define the domestication history of Min pig, further explore the list of candidate genes for introgression, and deepen the understanding of the genetic background of the unique trait mechanism of Min pig.

In the breeding process, there will inevitably be errors and deficiencies in the record of pedigree. Especially for the conservation farms, it may be difficult to evaluate the true kinship between individuals in the subsequent analysis, resulting in a decrease in the accuracy of analysis [29]. Meanwhile, since to the pedigree in this study can only be traced back to the effective information of 4 generations of the Min pig conserved population, the true kinship between individuals cannot completely and truly represented. The genomic kinship matrix constructed by whole genome SNP markers can more truly reflect the individual kinship [30] when the marker density is appropriate. Therefore, while constructing the A matrix based on pedigree, we also adopted the strategy of constructing the molecular pedigree G matrix and the pedigree -genome joint matrix H matrix based on the genomic information and calculated the correlation coefficient between different matrices. The results showed that the A matrix based on pedigree and the G matrix and H matrix based on the genome are strongly correlated and extremely significant, indicating that the constructed molecular pedigree is more accurate than incomplete pedigree information, and the molecular pedigree can be divided into 9 families, in which the distribution of male individuals is relatively uniform, and will lay a foundation for further breeding of rational utilization of families.