SNP-based genetic signatures revealed breeding effects in indigenous Livni compared with Landrace and Large White breeds

: Livni is one of the Russian local pig breeds. We previously reported that this breed was more distinct from Duroc breed than from Landrace and the Large White breeds


INTRODUCTION
The pig is a major livestock species, and the global pork production primarily relies on the use of a limited number of international commercial breeds, specifically Duroc, Large White, and Landrace [1].Intensive implementation of commercial hybrid breeds characterized by high production standards led to an impoverishment of genetic resources which in the past had a fair distribution [2].However, recently a strong attention has been attracted to local breeds for improving genetic diversity and conservation of genetic resources.Local breeds are valued not only by adaptive traits, but also by the unique functional characteristics and intensi-vely studied in Asia, Europe, Africa, as well as North and Latin America [3][4][5][6][7][8][9][10][11][12][13][14].
Twenty-two local breeds were recorded in the Soviet Union in 1980, which were generated by crossing of native breeds adapted to the local climate and having appropriate constitution and disease resistance with highly-productive improved European breeds [15,16].As a result of the interbreeding of the imported breeds and crossing them with the native animals, many pig breeds were created during 1920-1990.For example, Ukranian White Steppe was created in Askania Nova and approved in 1932; North Siberian -in Novosibirsk and approved in 1942, Urzgum -in the Kirov region and approved in 1957 -by crossing native pigs with the Large White boars.Breeds, as Kemerovo (approved in 1961), Breitov (approved in 1948), Latvian White and Lithuanian White (both approved in 1967), Semirechensk (1978), Mirgorod (1940), Tsivilsk (a cross of native Chuvash pigs with Large White boars, the breed is not approved), Mangalitsa, Altai (approved in 2015), and others were created by multiple crossbreeding procedures.[15].According to the Department of Livestock and Breeding of the Ministry of Agriculture of the Russian Federation data, 98% of the total pig purebred population in 2020 included four breeds -Large White (66%), Landrace (15%), Yorkshire (13%) and Duroc (4%) [17].Other breeds' share was about 2%.Pavlova et al. consider 0.56% of the total pig population in the RF are of the local breeds -Livni, Altai, Tsivilsk, as of January 1, 2022.Four breeds make 99.46% of the RF pig herd, namely 56.9%Large White, 18.52% -Yorkshire, 18.18% -Landrace, and 5.83% -Duroc breed [18].The dramatically reduction of local pig breeds during last 30 years finally led to remaining only Livni, Altai Meat-type, Short-Eared White, and Tsivislk.The authentic Kemerovo breed has also been mentioned for a number of years.However, according to the Yearbook on breeding work in pig husbandry in establishment of the Russian Federation for 2021, the last time a breeding farm certificate for the Kemerovo purebred was issued in 2019 [19].It should be noted, that the certificate for the Tsivislk breed was last issued in 2021.
Livni is one of the Russian local pig breeds approved in 1949.Pigs of the Livni breed are large, white, black-mottled, black and red.At present, only a small population of Livni pigs is kept in a single farm in the Oryol region [20].According to the Yearbook on breeding work, in pig husbandry in establishment of the Russian Federation for 2021 one certificate is issued annually for the Livni purebred, but the total number of the Livni pigs is steadily declining.At the beginning of 2022, 547 heads were purebreds, including 348 sows with the share in the total livestock of 0.24% [19].For comparison, in 1949 the Livni livestock was 6757 purebreds (1334 sows), while in 1980 it was 27 200 purebreds (5500 sows) [21].It is noteworthy that at the age of 6 months, Livni correspond to the bacon (meat) type.Then the active accumulation of fat begins and at the age of 10 months Livni pigs are already belong to meat-and-fat type, and with further fattening lead to fat type [21].
We previously reported that Livni breed is characterized the highest level of genetic diversity compared with commercial breeds.The neighbor-joining tree showed that this breed was the most distinct from Duroc breeds, but formed the knot bounding the branches corresponding to the Landrace and the Large White breeds.This observation confirmed the participation of these two breeds in the Livni breed creation.The aim of our study was to determine the SNP-based genetic signatures in Livni breed common with Landrace and the Large White breeds, and ones that are affected by putative selection in the genome of Livni breed and could be associated with fatty tissue formation and breed specificity.

Samples and genotyping.
For the study, we used samples (ear tissue) of Livni pigs (n = 35).Only purebred animals registered in Russian swine herdbook were selected, the origin of which is confirmed by both the pedigree data and DNA analysis.For genotyping, we selected the most unrelated individuals.Samples of all breeds were sent to the Ernst Federal Research Center for Animal Husbandry.A parentage and breed assignment of those breeds were confirmed based on the microsatellites in the laboratory of the Ernst Federal Research Center for Animal Husbandry, which has a certificate of 2020-2021 ISAG Pig STR Comparison Test (2020-2021) and has a special license issued by the Russian Ministry of Agriculture.Commercial breeding farms and the Ernst Federal Research Center for Animal Husbandry collaborate based on the contracts.In the contract, a clause states the consent of the owners (breeding farms) to use the samples with research purpose.
Moreover, the study did not involve any endangered or protected animal and all procedures were conducted according to the ethical guidelines of the L.K. Ernst Federal Science Center for Animal Husbandry.The Commission on the Ethics of Animal Experiments of the L.K. Ernst Federal Science Center for Animal Husbandry approved the protocol No. 6 of May 10, 2021.The ear tissues were collected by trained personnel under strict veterinary rules in accordance with the rules for conducting laboratory research (tests) in the implementation of the veterinary control (supervision) approved by Council Decision Eurasian Economic Commission № 80 (November 10, 2017).
The genome-wide SNP genotyping was carried out using an iScan microarray scanner (Illumina Inc., Singapore) using the Porcine GGP HD BeadChip (Illumina Inc., San Diego, CA, USA), which contains ~ 80 000 SNPs.In our study, we used all the capital equipment required for SNP genotyping by Illumina SNP arrays.The equipment belongs to the Center for Collective Use "Bioresources and Bioengineering of Agricultural Animals" of the Ernst Federal Research Center for Animal Husbandry (https://www.vij.ru/infrastruktura/ckp,accessed on 10 May 2021).The SNPs genotypes of Large White (n = 53) and Landrace (n = 50) breeds were included in the data set and obtained from Center for Collective Use "Bioresources and Bioengineering of Agricultural Animals" of the Ernst Federal Research Center for Animal Husbandry.
Genetic diversity, PCA, Neighbor-Net and Admixture.To assess the within-population genetic diversity, the observed (H O ) and unbiased expected ( U H E ) heterozygosity, the rarefied allelic richness (А R ), and the unbiased in-breeding coefficient ( U F IS ) were estimated using the R package, diveRsity [24].Additionally, we computed the genomic inbreeding coefficient based on runs of homozygosity (ROH, F ROH ) as the ratio of the sum of the length of all ROHs per animal to the total autosomal SNP coverage; for ROH estimation, see the "Runs of Homozygosity Estimation" Section below).PCA was performed using PLINK v1.9 software.An R package, ggplot2, was used to visualize the results [25].Pairwise FST values were calculated in the R package, diveRsity, and used for the construction of the Neighbor-Net tree in SplitsTree software (version 4.14.5)[24,26,27].Admixture software (version 1.3.0)was employed for genetic admixture analysis and an R package, pophelper, was used for plotting the results [28,29].A crossvalidation (CV) procedure was used to calculate the number of ancestral populations (k) from one to five using Admixture software (version 1.3.0).
Selection signature analysis.Three different statistics were used for detecting the signatures of selection in the genome of pigs: the calculation of F ST values for each SNP when comparing pairs of breeds, the estimation of the ROH islands, which were overlapped among different animals within each breed, and hapFLK analysis.
F ST analysis.F ST values for all SNPs were estimated for pairs of breeds using PLINK 1.9 [24].Minor allele frequencies were below 5% (--maf 0.05) [30].The top SNPs corresponding to 0.1% of F ST values were used to represent a selection signature, according to Kijas et al. and Zhao et al. [31,32].
Runs of homozygosity estimation.Runs of homozygosity were detected according to the window-free method for consecutive SNP-based detection using the R package, detectRUNS [33].One SNP with a missing genotype and up to one possible heterozygous genotype in one run were allowed to avoid the underestimation of the number of ROHs that were longer than 8 Mb [34].The minimum ROH length was set to 500 kb to exclude the common ROHs.To minimize false-positive results, the minimum number of SNPs was calculated as it was proposed by Lencz et al. and later modified by Purfield et al. [35,36].
Putative ROH islands were defined as overlapping homozygous regions in analyzed individuals within each breed.A threshold of 50% (the minimum proportion of animals within the breed in which overlapping ROH were detected) was selected, as this was suggested in other studies [37,38].We applied the threshold of 0.1 Mb for the minimal overlapping length size and 5 SNP for minimum number in ROH island.
HapFLK analysis.In this study, a hapFLK analysis was performed to detect the selection signatures through haplotype differentiation among the studied breeds using hapFLK software (version 1.4.)[39].The number of haplotype clusters per chromosome was calculated in fast-PHASE by using cross-validation and was set to 35 [40].For detailed analyses, the hapFLK regions containing at least one SNP with a p-value threshold of 0.01 (−log10(p) > 2) were selected.
Identification of candidate genes.For candidate gene mining in the genomic regions under putative selection, the genomic localization of the regions as detected by three different statistics was used, i.e., the FST, ROH, and hapFLK methods.Regions that were overlapped and revealed by at least two different techniques were prioritized.Borders of these regions according to the 10.2 genome assembly were converted to genome assembly 11.1.Genes located on the selected regions were obtained from the Ensembl Genes Release 103 database based on the Sus scrofa gene sequence assembly [41].
Functional enrichment analysis.To understand the biological functions of the candidate genes, the Database for Annotation, Visualization, and Integrated Discovery (DAVID) was used for enrichment analysis [42].Significant annotation clusters of enriched Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways and Gene Ontology were selected using an enrichment score of more than 1.3 and a p-value of < 0.05.To learn the biological functions of annotated genes and genes not included in clusters, a comprehensive literature search including information from other species was carried out.

RESULTS AND DISCUSSION
Genetic diversity.The Livni pigs were characterized by higher level of genetic diversity assessed by the levels of observed heterozygosity, unbiased expected heterozygosity, and allelic richness as compared to the Landrace and Large White breeds.The negative value of the inbreeding coefficient U F IS indicates an excess of heterozygotes from the Hardy-Weinberg equilibrium in all the breeds (Table 1).In commercial breeds, the excess of heterozygotes was more significant compared to the Livni breed.
Breed relationship and admixture.The PCA-plot (Fig. 1a), the neighbor-joining tree (Fig. 1b) and cluster structure (Fig. 1c) showed the breed-specific distribution of individuals for all of the studied breeds.Obtained distribution indicated the insignificant participation of the Landrace and Large White breeds in the formation of the modern allelofund of Livni pigs and demonstrated that sampling is suitable for searching for loci under selection pressure in the studied pig breeds and their subsequent structural annotation.
The distribution of ROH island number and length in chromosomes is presented in Table 2. Forty-two ROH islands were detected in the Livni breed, which covered 34.415 Mb of the genome, while for Landrace and Large White, 126 and 224 ROH islands covered 161.792 and 282.402Mb of the genome, respectively.The average length of the ROH island in Livni breed was significantly lower than that of pigs of commercial breeds: 2.868 ± 0.822 Mb versus 8.988 ± 2.185 (Landrace) and 15.689 ± 2.770 Mb (Large White), respectively (p < 0.001).
The hapFLK analysis resulted in the identification of 13 putative regions affected by the selection (Fig. 3).These regions were distributed among 10 autosomes, including regions on SSA1, SSA3, SSA14, and SSA13 with     a statistical significance of p < 0.001.Four regions were Large White-specific, three -Landrace-specific, three -Large White and Landrace-specific, three -Livni and Large White-specific (Table 4).
Comparing the genomic localization of the regions under putative selection detected by three different statistics (F ST , ROHs, and hapFLK) revealed the presence of 13 overlapping regions, which were identified by at least two different methods (Table 5); 7 regions corresponded to the Large White breed, 2 corresponded to the Landrace breed, 2 were common to Large White and Landrace breeds, and 2 were common to Large White and Livni breeds.Additionally, in the list of genes for structural and functional annotation, we included ROH islands identified only in Livni pigs, as well as common ROH islands identified in the Livni and one or two compared breeds.Thus, 16 Livni-specific regions, and 39 regions, which are common for both two and more breeds were selected for the structural and functional annotation.
Candidate gene determination and functional enrichment determination.The structural annotation of these regions revealed the presence of 238 candidate genes: 50 genes were specific to Livni pigs, 62 to Livni   and Large White pigs, 35 to Livni and Landrace pigs, 36 to all studied breeds, and 55 were specific to Large White and Landrace pigs (Table 6).
Using the DAVID web tool and a list of 238 candidate genes found in the genomic regions with selection signatures, 182 genes with described functions were identified.The significant clusters are shown in Table 7. Annotated clusters with an enrichment coefficient -log10(p) > 1.3 (corresponds to p < 0.05) were not determined for the Livni breed and all three studied breeds (Livni, Large White and Landrace).Two reliably annotated clusters were identified for the Livni and Large White, Large White and Landrace breeds, and one annotated cluster for Livni and Landrace.For the list of Livni and Large White genes, the presence of two annotated clusters was revealed.Cluster 1 (enrichment coefficient = 2.11) included G6PC2, HKDC1, HK1 genes involved in carbohydrate metabolism.Cluster 2 (enrichment coefficient = 1.60) included SUPV3L1, SLC25A16, HKDC1, DDX21, PIK3C2A, MAP3K7, DDX50, and HK1 genes involved in the processes of DNA replition and repair.For Livni and Landrace one reliable cluster (enrichment coefficient = 4.49) was determined, including the genes CIART, HORMAD1, HOXD3, HOXD4, HOXD8, HOXD9, HOXD10, HOXD12, HOXD13, EVX2, NR2E1, and PLEKHO1.Genes under selection pressure in commercial pig breeds (Large White and Landrace) were combined into two reliable clusters.Cluster 1 (enrichment co-efficient = 1.74) combined the genes KCNA1, KCNA6, KV1.5, and SLC30A9 involved in the regulation of ion transmembrane transport, mainly potassium.The IBTK, KCNA1, KCNA6, and ZBTB10 genes regulating transcription repression and interaction with components of histone deacetylase co-repressor complexes were localized in cluster 2 (enrichment coefficient = 1.54).
Specific and overlapping sites in the genome of Livni, Large White and Landrace breeds that are under selection pressure have been identified.Positional candidate genes were identified and their annotation was performed.In the current study, three pig breeds were examined and compared.We previously reported that Livni breed is characterized the highest level of genetic diversity compared with commercial breeds.The neighbor-joining tree showed that this breed was the most distinct from Duroc but formed the knot bounding the branches corresponding to the Landrace and the Large White breeds.This observation confirmed the participation of these two breeds in the formation of the Livni breed during it artificial selection.We observed the highest level of genetic diversity in Livni pigs compared to commercial breeds (Table 1), which may be a consequence of the participation of various breeds in the development of the Livni breed, including Large White and Landrace.However, results of breed relationship and admixture revealed distribution indicated the insignificant participation of the Landrace and Large White breeds in the formation of the modern allelofund of the Livni pigs.
Using three different statistics (top 0.1 FST at pairwise breed comparison, ROH islands and hapFLK analysis), we selected 13 overlapping regions, which were identified by at least two different methods (Table 2); 7 regions corresponded to the Large White breed, 2 corresponded to the Landrace breed, 2 were common to Large White and Landrace breeds, and 2 were common to Large White and Livni breeds.Among 238 candidate genes, which were localized within selected genomic regions (Table 3), 182 genes had the described functions in GO-terms; among them, 50 genes were specific to Livni pigs, 62 were specific to Livni and Large White pigs, 35 were specific to Livni and Landrace pigs, 36 were specific to all studied breeds, and 55 were specific to Large White and Landrace pigs, (Table 3).
Among common genes for three studied breeds, MLANA and JAK2 were previously observed in Livni and Duroc breeds and involved in adipogenesis [20,43].It was reported that FKBP7 is highly expressed in subcutaneous adipose tissue of mature Erhualian pig, while CHAT is essential for macrophages as a source of acetylcholine for the regulation of adaptive thermogenesis [44,45].HSPH1 is a known marker of both human and mouse brown adipocytes and was upregulated in young and old brown adipocytes after acute cold exposure [46].HSPA1L were found to be differently expressed between the low and high drip loss groups in the Duroc pigs [47].NCOA4 may play a role in early events of adipocyte differentiation and were found in Pudong White pigs [48,49].PLGRKT coordinately regulates multiple aspects of adipose function and was found to be related to obesity [50,51].According to Gene Ontology terms, ALOX5 is strongly associated with immunity, lipid metabolism and fat cell differentiation, insulin secretion, and oxidative stress.Interestingly, this gene was also very highly significantly associated with feet and leg structure soundness traits in pigs [52].OSBPL6 linked with lipid and sterol transport and encoded by miR-33, which may also regulate adaptive thermogenesis [53].PLEKHA3 is also associated with lipid metabolism, and mutations were identified for this gene in the Puławska pig breed, which is characterized by thicker backfat and better meat quality values [54].PARG is linked with carbohydrate metabolic process and could be involved in lipid metabolism [55].According to Gene Ontology terms, ERMP1 involved incellular response to oxidative stress, HEYL -in skeletal muscle cell differentiation.INSL6 was linked with male fertility in Enshi pigs and reproduction in Anhui pigs [56,57].MSMB was closely related to body weight, body height, abdominal circumference, and chest depth in Xiangsu hybrid pigs [58].OGDHL was up-regulated in the liver in pigs with higher backfat thickness of Songliao black female pig population [59].Although PRKRA is strongly associated with immune response, including piglets, this gene plays unexpected role in the regulation of mitochondrial biogenesis and energetics in cells and brown adipocytes [60,61].ZFAND4 gene encodes stress proteins and was detected in Pudong White pigs, as well as ZNF22 [49,62].C10orf10 is involved in adipose tissue thermogenesis and was observed in heavy Iberian  Pigs [63,64].PDL1 was determined as candidate biomarkers for predicting residual feed intake in Yorkshire pigs, as well as U1 [65,66].RLN is a candidate gene for reproductive traits in pigs and was found to regulate adipose tissue development through stimulating adipogenesis and modulating adipocyte metabolism [67][68][69].SNORA19 could be involved in body temperature regulation [70].U6 was associated with litter traits in Yorkshire and Landrace pigs and was a selection signature gene in Meishan population [71,72].
In the Livni and Landrace breeds, 35 common genes were detected, which formed one cluster with enrichment coefficient = 4.94 and predominant HOXD genes.According to Gene Ontology terms, HOXD10 and HOXD9 are involved in various developmental processes, such as single fertilization, skeletal muscle tissue development, adult locomotory behavior, embryonic skeletal system morphogenesis, peripheral nervous system neuron development, neuromuscular process, etc. HOXD10 is required systemically for secretory activation in lactation [73].Expression level of HOXD10 was increased in animals with high marbling [74].HOXD9 and HOXD10 are associated with such traits as growth, body weight and composition, abdominal fat, organogenesis, and feed intake and consumption [75].They also play an active role in chondrogenesis and the development of adipose depots [76,77].HOXD3, HOXD8, HOXD12, and HOXD13 are also associated with skeletal system development.HOXD12 is differently expressed between large and small piglet size [78].HOXD3 is also associated with nervous system development, considered as predictors for feed efficiency traits [79,80].It was reported that HOXD4 and HOXD8 are up-regulated in differentiated adipocytes [81].HOXD8 gene is involved in patterning the lower thoracic and lumbar vertebrae, in the urogenital tract development, also of mesoderm origin [82,83].HTR6, associated with nervous system, was identified as interesting candidate genes involved in axonogenesis and synapsis in Iberian breed [84].ADAMTSL4 was found to evolve under positive selection and exhibited significant downregulated mRNA expression in the Tibetan pigs [85].ABCC1 is expressed in adipose and skeletal muscle, upregulated in obesity, and involved in the embryo development of pig; it was also detected in Northeast wild boar [86][87][88][89].HORMAD1 is linked with embryo development and productivity.Z. Zeng et al. noted HORMAD1 to belong to growth-related Meishan pig genes [90].HORMAD1 was under heavy selection based on runs of homozygosity in a Large White pig population and associated with obesity [91].SEC63 was determined as candidate genes for estimated breeding values feed conversion ratio in Maxgro boars [92].It was found an association between the CIART genotype and backfat thickness in Duroc pigs, and its expression is affected by food intake [93,94].According to Gene Ontology terms, ENSA is associated with regulation of insulin secretion and related to adipocyte development [95].ECM1 is involved in immunity and bone development.It was reported to be an important gene highly expressed in subcutaneous white adipose tissue (sWAT) as compared to brown adipocytes, and was determined in Korean Wild Boar, up-regulated in Congjiang Xiang pigs with large litter size and in testis tissue from Duroc boars [96][97][98][99].KLHL1 could be linked with Landrace and Yorkshire pig backfat thickness in Korea and involved in environmental adaptation [100,101].NR2E1 is involved in developmental processes and linked with environmental adaptation concerning behavioral defense response in Xiang pigs [102].It showed significant associations with feed conversion efficiency and growth rate in pigs [92].PRPF3 gene is differentially expressed in the Longissimus dorsi muscle being more abundant in Large White than in Wujin pigs [103].VPS45 could be linked with growth trait [104].FOXO3A promotes metabolic adaptation and stress resistance in hypoxia, associated with carcass length, backfat thickness and drip loss, related to muscle development in Iberian pigs [105][106][107].FOXO3A could promote lipid accumulation as well [108].Ssc-mir-10b was downregulated in Tibetan pigs, related to hypoxia adaptation, play important roles in fat-related processes in adipose tissue, had been frequently reported highly expressed in skeletal muscle during porcine prenatal and postnatal developmental stages and abundantly expressed in subcutaneous adipose tissue in pigs [109][110][111][112][113].
In the Livni and Large White breeds the largest amount of common genes was detected and averaged 62, which formed two clusters.Cluster 1, with enrichment coefficient = 2.1, was characterized with genes involved in glucose metabolism.Among them, G6PC2, HKDC1 and HK1 are critical for glucose homeostasis.HK1 effects on growth and meat quality in Polish Landrace [114].It is important for sperm motility in Duroc, enriched in brown adipocytes of aged mice, up-regulated by severe cold and essential for brown adipocytes thermogenesis [115][116][117][118]. Cluster 2, with enrichment coefficient = 1.60, demonstrated helicase genes.DDX21 is associated with immunity and belongs to the top 4 lymphocyte associated genes in pigs [119].SUPV3L1 is important for the maintenance of the skin barrier and related to percentage of certain fiber types [120,121].MAP3K7 is also linked with immunity and strongly associated with neuropsychiatric processes [122].It was reported to be associated with growth traits and adipocyte differentiation [100,123].PIK3C2A gene is related to hepatic insulin resistance and steatosis, average daily gain and lean meat percentage, intramuscular fat and backfat thickness in two Duroc populations, being under positive selection in all high-altitude species [124][125][126][127].According to Gene Ontology terms, ABCB11 is associated with fatty and bile acid metabolic process and could be involved in gene networks for intramuscular fatty acid composition in porcine [128,129].ABCC8 was reported to be selection region for intramuscular fat and backfat thickness in two Duroc populations, and the most down-regulated genes in the group with higher backfat thickness in Yimeng black pigs [126,130].DEF6 is linked with average backfat thickness [131].FKBP5 is associated with immunity, backfat thickness and leaf fat weight, significantly contributed to residual feed intake [79, 131 -133].Expression of this gene is inversely associated with the expression of lipolytic, lipogenic and adipogenic genes [134].According to Gene Ontology terms, LDB3 is associated with heart development and muscle structure development, related to muscle growth traits in pig and may have potential roles in environmental adaptation [135,136].ARMC12 regulates spatiotemporal mitochondrial dynamics during spermiogenesis and is required for male fertility [137].BMPR1A is associated with numerous developmental processes, identified as a novel candidate gene affecting the number of thoracic vertebrae in pigs, and regulates the development of hypothalamic circuits that are critical to the feeding behavior [138,139].Additionally, BMPR1A is important in brown fat development and involved in browning of white adipose tissue [140,141].CCAR1 positively regulates adipocyte differentiation [142].CLPSL2 and CLPS are linked with digestion, lipid catabolic process, and response to food.CLPSL2 could be involved in the regulation of acrosomal integrity, spermatozoa motility, and male fertility, while CLPS demonstrated effect on some characteristics connected with lean content of the carcass and fat content and affected intramuscular fat content [143][144][145].It may be associated with former selection toward reduced fat content in carcass [114].According to Gene Ontology terms, COL13A1 is associated with skeletal system development, was under significant positive selection Yorkshire pigs and associated with fat deposition, as well as HNRNPH3 [146][147][148].JMJD1C is potentially associated with cold adaptation [149,150].It demonstrates the positive selection in regulation of various reproductive traits in pigs [151][152][153][154]. JMJD1C was identified in Tibetan pigs that are well adapted to the high altitude [155].On the other hand, this gene have been associated with white blood cells in Large White pigs, identified as a novel regulator of adipogenesis and contributed to browning [156][157][158].According to Gene Ontology terms, MYOD1 and MYPN are strongly involved in skeletal muscle tissue development.It was reported about potential role of MYOD1 in body-fat distribution regulation [159].Mutations in the MYOD1 gene show a significant effect on the pork meat quality and single nucleotide polymorphisms in the porcine MYOD1 affected on meat quality traits and carcass traits in heavy pigs [160][161][162].MYPN is related to body composition and can be considered as candidate for meat and carcass traits in pigs [163][164][165].NRBF2 is linked with energy metabolism and was specific selective for Tibetan pig [155].NUCB2 is expressed in fat depots of the pig and that level of expression is sensitive to stimulation of appetiteregulating pathways in the hypothalamus [166].It plays an important role in whole-body energy homeostasis and body weight at puberty by regulation of appetite of Jinhua Pigs [126].NUCB2 is also involved in cold adaptation, indicating that central nesfatin-1 regulates ther-mogenesis [167,168].REEP3 mediates adipogenic differentiation [169].According to Gene Ontology terms, SIRT1 is linked with regulation of lipid storage, white and brown fat cell differentiation, adipose tissue development, etc.It is implied in the browning of white adipose tissue, promotes lipid metabolism and mitochondrial biogenesis in adipocytes and coordinates abiogenesis by targeting key enzymatic pathways [170,171].Apart from that, it negatively correlates with intramuscular fat content and demonstrates protective role in skeletal muscle's adaptation to cold stress [172,173].SNCG controls metabolic functions in fat cells and belongs to white adipose tissue-selective genes [174,175].ZNF76 is very close to peroxisome proliferative activated receptor delta (PPARD) at 35 Mb, which is a positional and physiological candidate for affecting backfat thickness [176].RNF213 is involved in adipogenesis and emerged as a link between obesity, inflammation, and insulin resistance [177,178].SNORD14 were more expressed in Large White heavy pigs with high intramuscular fat content [179].U3 was identified as a promising candidate gene for average backfat thickness in multiple pig breeds and populations [180].
Annotated clusters with an enrichment coefficient -log10(p) > 1.3 (corresponds to p < 0.05) were not determined for the Livni breed.However, 50 candidate genes were specific to Livni pigs.DLGAP5 is a stillbirth associated gene involved in lipid deposition-related pathways and significantly associated with intramuscular fat content [181][182][183].ERCC4 is also associated with intramuscular fat content, presented in Tibetan wild boar and related to "response to UV" [151,184,185].GPR63 has been identified as a receptor for intercellular lipid messengers and associated with reproduction traits [186,187].According to Gene Ontology terms, LDB3 is involved in heart and muscle structure development, while PBX3 -in various important developmental processes.PDZRN3 and ATG14 could affect intramuscular fat content in Suhuai pigs [183,184].They are involved in adipocyte differentiation, demonstrating negatively influence [188,189].RBM15B is linked with average daily gain in Italian Large White pigs, while TBPL2 -with fertility [190,191].WDHD1 is associated with stillbirth in Large White sows and residual feed intake [79,181].According to Gene Ontology terms, BMPR1A is associated with immunity, bone, lung and heart development.BMPR1A is reported to be associated with obesity and important for brown adipocytes, candidate gene affecting the number of thoracic vertebrae in a Large White × Minzhu intercross pigs [138,[192][193][194]. DOCK3 is linked with fatness and growth in Huainan pigs [195].
LGALS3 is linked with immunity, sensitive to cold exposure, associated with stillbirth, and involved in adipogenesis [181,[196][197][198]. GLUD1 is an important gene for metabolic process, increased by cold exposure and essential for brown adipocytes [199,200].MANF positively regulates thermogenesis, resists obesity, as well as regulates hypothalamic control of food intake and body weight [201][202][203].MAPK1IP1L and SOCS4 are likely candidate genes for stillborn [181,204].According to Gene Ontology terms, MEF2C is involved in numerous developmental processes, may be a key gene in insulin-induced adipocyte differentiation, involved in fat deposition in pigs, important for foetal developmental, and associated with total number born and number born alive [205][206][207].WISP3 is linked with sketetal and muscular development [208].PDE10A is associated with chest circumference in Yorkshire Pigs, back fat thickness at 100 kg in Landrace pigs, and contributes to the regulation of energy homeostasis [209][210][211].PARN was identified as candidate genes associated with age at 100 kg in Large White pigs [212].TFAM promotes mitochondrial DNA content, which necessary for increased fusion during cold adaptation [213].Its amount significantly elevated after cold exposure and essential for thermoregulation [214,215].Mutation in the TFAM gene effects on fattening and carcass traits in commercial pig populations [216].TFAM gene expression abundance in particular tissues such as liver and L. dorsi revealed some strong correlations with carcass and meat quality traits including marbling [217].SNORD22 is associated with trimmed thigh weight in Italian crossbred pigs [218].U4 and ssc-mir-9-2 were previously determined in pigs [219,220].Genes associated with reproductive, meat and fat quality, carcass, and immunity traits in pigs were found in genomic regions affected by putative selection.Along with fatting genes, ones linked with thermogenesis were unexpectedly detected which oppositely should led to fat reduction.However, pigs could not have brown adipocytes but could have beige ones, which are very important for maintaining alternative mechanisms of thermoregulation in pigs that possibly avoid fat reduction [221][222][223][224].

CONCLUSION
The dramatic reduction of local pig breeds during last 30 years finally led to 0.56% of the total pig population in the RF, mainly Livni, Altai, and Tsivilsk breeds.There are several reasons for the reduction of local pig breeds: a trend to the reduction the total amount of fat on pork carcass and in meat and the aggressive implementation of the Western commercial breeds.Commercial breeds were bred without taking into account Russia environment, the quality and composition of feed and drinking water.Local pigs bred in the USSR are characterized by unpretentiousness to feed, stress and cold resistance, as well as precocity and high productivity.Livni is one of the Russian local pig breeds.Landrace and the Large White breeds participated in creation of the Livni breed, but obtained breed relationship and admixture results indicated the insignificant participation of these breeds in the formation of the modern allelofund of Livni pigs.The largest amount of common genes was detected between the Livni and Large White breeds.Genes involved in glucose metabolism, namely G6PC2, HKDC1, and HK1 are critical for glucose homeostasis, which could effect on the growth and meat quality traits, as well as on thermogenesis.Other genes were associated with immunity, related to percentage of certain fiber types, growth traits, average daily gain and lean meat percentage, intramuscular fat and backfat thickness, etc.Among 35 common genes of the Livni and Landrace breeds, enrichment with HOXD genes was observed.HOXD genes are involved in various developmental processes, such as single fertilization, skeletal muscle tissue development, adult locomotory behaviour, embryonic skeletal system morphogenesis, lipid metabolism, etc., and are associated with traits such as growth, body weight and composition, fat development, organogenesis and feed intake, etc. Candidate genes associated with various growth, carcass and reproductive traits and essential for thermoregulation were specific to Livni pigs.Livni breed belongs to the meat-and-fat type, but during development pigs could be also meat and fat types.The analysis of genetic architecture confirmed the unique structure of local breed that was bred using commercial Landrace and the Large White breeds.During formation own allelofund, the Livni breed fixed important traits, including flexibility during growing and feeding.

Figure 1 3 PC2Figure 2
Figure 1 Genetic relationships between Landrace, Large White and Livni pig populations: (a) Principal component analysis (PCA) plot showing the distribution of Landrace, Large White and Livni individuals in two-dimensional coordinate system, i.e., the first (PC1; X-axis) and second (PC2; Y-axis) principal components, with percentage of total genetic variability, which can be explained by each of the two components, being indicated within the parentheses; (b) Neighbor-Net tree constructed based on the IBSdistances among the studied populations; (с) Admixture plot representing cluster structure of the studied populations if the number of clusters K = 3

Figure 3
Figure 3 Signatures of selection in the genomes of the studied breeds based on the hapFLK statistics.Values for the X-axis are pig autosomes, and those for the Y-axis are values of statistical significance (−log10 p-values).The red line indicates the threshold of significance at p < 0.01 (i.e., −log10(p) > 2) CONTRIBUTION I.M. Chernukha conceived and designed the study.I.M. Chernukha, L.V. Fedulova and E.A. Kotenkova designed the methodology.L.V. Fedulova and E.A. Kotenkova analysed and described the results.I.M. Chernukha and E.A. Kotenkova wrote the manuscript.All authors contributed to data interpretation.

Table 2
The distribution of ROH island number and length in chromosomes * SSA -Sus scrofa autosomes; # n -number of SSA

Table 3
Common ROH islands identified in genome of two or three studied breeds

Table 4
HapFLK regions identified in the genome of the studied breeds * SSA -Sus scrofa autosomes

Table 5
Overlapped genomic regions and/or SNPs under putative selection identified by at least two different statistics in the Duroc

Table 6
Genes within the overlapped genomic regions affected by putative selection

Table 7
Functional Gene Ontology terms enriched with candidate genes