Head

Pork fatty tissues have many advantages, both for processing into processed products and for the sensory and nutritional qualities of fresh meat and pork products. A review of the situation of the genetic variability of adiposity in pigs in France in relation to product quality is necessary to define future breeding strategies and thus better meet the various expectations of industry and consumers¹.

1.1. Anatomical location, cellular and molecular characteristics

Within the body, there are several sites of fat deposit with a wide range of sizes. In pigs, the most quantitatively important depots are subcutaneous, with backfat as the main depot, and intermuscular, with the other fat depots (perirenal or leaf fat, intramuscular fat) representing only a small fraction of body fat (Henry, 1977). In an animal with 115 kg of body weight (commercial slaughter stage), subcutaneous and intermuscular fat depots represent 74% and 26% of total carcass fat, respectively (Monziols et al., 2005).

Structurally, adipose tissue is composed essentially of adipocytes, which are cells specialised in the storage of lipids (Louveau et al., 2016). These spherical cells, which can be more than 100 µm in diameter, are characterised by the presence of a single lipid vacuole that occupies up to 95% of the total cell volume. They are enclosed in a mesh of connective fibres that also support blood and lymphatic vessels and nerve endings. In addition to adipocytes, adipose tissue contains adult stem cells, pre-adipocytes, endothelial cells, fibroblasts and immune cells. Unlike many mammals, in which two types of adipocytes, white and brown, coexist or replace each other during growth, pigs have only white adipocytes in their adipose tissues (Trayhurn et al., 1989).

Chemically, subcutaneous adipose tissue in pigs contains on average 69-77% lipids and 14-22% water, with internal adipose tissues (perirenal fat) containing more lipids (Wood et al., 2003; Gondret et al., 2014). Muscle lipids are essentially triglycerides, which are reserve lipids (0.5-5%), phospholipids, which are structural lipids present in cell membranes (0.5-1%) and cholesterol (0.05-0.1%) (Lebret and Picard, 2015). Triglycerides are stored mainly in the lipid vacuole of adipocytes that develop along fibre bundles and, to a lesser extent (5-20%), in the cytoplasm of muscle fibres in the form of lipid droplets. The phospholipid content varies relatively little, while the triglyceride content varies greatly and depends strongly on the size and number of intramuscular adipocytes (Listrat et al., 2015).

Tissue lipids are made up of Fatty Acids (FAs), which are divided into three classes: saturated (SFA), which represent an average of 38% (muscle) to 40% (backfat) of FAs and are mainly C16:0 (palmitic) and C18:0 (stearic); monounsaturated (MUFA), which represent ca. 45% of the FAs (muscle and backfat) and are essentially C18:1 (oleic); and polyunsaturated (PUFA), which represent 15-20% of the FAs (Mourot, 2010). Within PUFAs, distinction is made between n-6 or omega-6 PUFAs, which are the most common, essentially C18:2 (linoleic or LA), and n-3 or omega-3 PUFAs, which include C18:3 (linolenic or ALA), C20:5 (eicosapentaenoic or EPA) and C22:6 (docosahexaenoic or DHA). Among these FAs, C18:2 and C18:3 are indispensable because they are not synthesised by the pigs. Their levels, as well as those of their derivatives, EPA and DHA, therefore depend directly on dietary intake.

1.2 Roles of adipose tissue in the body

The first roles attributed to adipose tissue were those of organ support and thermal insulation. Adipose tissue is also recognised for its essential role in energy storage. It has the capacity to store large amounts of FAs in the form of triglycerides in the adipocytes, thus ensuring the maintenance of energy homeostasis. This storage capacity is the result of a balance between the uptake of exogenous FAs, de novo synthesis of FAs or lipogenesis, their esterification into triglycerides, the hydrolysis of triglycerides and the re-esterification of the products of lipolysis. In pigs, adipose tissue is the major site of lipogenesis (Henry, 1977).

In addition to its various roles, adipose tissue has a secretory function. Since the discovery of leptin in 1994, many products secreted by fat cells have been identified (Komolka et al., 2014). Products of a peptidic nature are grouped under the term adipokines. These products, whose list continues to grow, are involved in various biological functions such as food intake, growth, inflammation and immunity. Through this secretory capacity, adipose tissue can interact with its close environment but also with the entire body.

1.3 Establishment and development of adipose tissues during growth phases of the pig

The chronology of appearance of fat depots varies depending on their anatomical location. The first groups of adipocytes appear in pigs during the foetal period. They are observed subcutaneously from 50-75 days of gestation (which lasts ca. 115 days) and develop in the pericardial, epididymal and perirenal regions at ca. 70 days of gestation, whereas the first intramuscular adipocytes do not develop until the first month of post-natal life (Bonnet et al., 2015). The development of adipose tissues during embryogenesis results from the engagement of embryonic stem cells in adipocyte differentiation pathways, followed by an increase in the number (hyperplasia) and volume (hypertrophy) of the constituent cells of these tissues during the post-embryonic growth phases (Bonnet et al., 2015).

Fat accretion occurs almost exclusively after birth in pigs and may continue throughout life. It increases from 1-2% at birth (Canario et al., 2007) to ca. 15% at 20 kg and 25% at 90 kg body weight in Large White male pigs (Karège, 1991). The growth of adipose tissue results primarily from adipocyte hypertrophy, with hyperplasia considered to contribute little to adipose accretion (Gardan et al., 2006; Bonnet et al., 2015).

2.1. The uses of fatty tissues in pork products

In pigs, the vast majority of the fatty tissues that make up the carcass is used as raw material for meat processing technologies, and the surplus is used for lard production (IFIP, 2016). They contribute to the taste and texture of meat and pork products and to the technological processing of products. However, depending on their anatomical origin, not all fatty tissues are as useful for product quality.

The subcutaneous tissues, which have the largest mass, are visible on the carcass or cuts. They can therefore be isolated easily and removed totally or partly before sale or processing. They are used to produce bacon, pork bard for roasts or galantines, dry or cooked pork products (sausages, pâtés, rillettes, etc.). In comparison, except for cooked ham, which is usually made from defatted ham pieces or muscles, intermuscular adipose tissues remain in the meat pieces marketed and are an integral part of the products offered to consumers (dry-cured ham, pork belly, etc.), as is intramuscular fat. They therefore play a major role in the visual acceptability of products (roasts, chops, dry-cured hams, smoked bacon, coppa, etc.) and their nutritional characteristics. Perirenal fat is, for its part, used mainly for the production of lard, marketed as is or used as an ingredient in processed products (IFIP, 2016).

2.2. Sensory quality: the challenge of invisible fat

The sensory properties of food products are the characteristics that consumers perceive through their senses. For meat, these are appearance (colour; presence of drip loss; amount of external, intermuscular, and intramuscular or marbled fat), texture (tenderness, juiciness) and flavour (odour, taste). It is well established that these parameters determine the appreciation of pork and influence the act of purchase or re-purchase by consumers, although marketing factors (packaging, price, availability) or psychological factors (values, socio-cultural aspects) also influence purchasing behaviour (Dransfield et al., 2005; Font-i-Furnols and Guerrero, 2014).

The sensory qualities of meat depend on complex interactions among (i) tissue characteristics at slaughter: diameter and types of muscle fibres, glycogen content, content and nature of IntraMuscular Fat (IMF), amount and distribution of the extracellular matrix; (ii) peri and post-mortem muscle metabolism: rate and amplitude of pH drop; temperature, duration and conditions of meat ageing (proteolysis and lipolysis, lipid oxidation); and (iii) meat preparation and cooking conditions or processing techniques and conditions (Ngapo and Gariépy, 2008; Lebret et al., 2015; Listrat et al., 2015; Warner and Dunshea, 2018).

The sensory characteristics of meat can be assessed by trained panels that evaluate, under defined conditions, the intensity of multiple descriptive traits of appearance, texture and flavour, or by consumers in taste tests to measure the pleasure experienced when tasting a product (appearance, texture, flavour). Analysis of about twenty publications based on a trained panel shows that the association between IMF content and the sensory quality of pork (loin) is positive in most studies (Lebret, 2009) but varies depending on the IMF content considered (a minimum content of 2.5% is considered favourable for sensory quality), the level of other major meat quality indicators (pH), the method and temperature of meat cooking, etc (Listrat et al., 2015; Warner and Dunshea, 2018).

What about the influence of IMF content on consumers' appreciation of meat? Before consumption, consumer preferences (overall appreciation, consumption or purchase intention) are directed mostly towards meats (loins) with the lowest amount of IMF, whereas preferences are often reversed after consumption, with the most-marbled meats judged to be juicier, more tender and tastier (Fernandez et al., 1999; Font-i-Furnols and Guerrero, 2014). These studies confirm the optimum of 2.5-3.5% IMF to enhance overall acceptability of meat by consumers. However, this association depends on their food habits and cultures (Dransfield et al., 2005) but also their age. Thus, increasing the IMF content improves overall acceptability of dry-cured ham among consumers over 25 years of age, but has the opposite effect among younger consumers (Ventanas et al., 2007). The influence of IMF content on the perceived quality also depends on the product: while the hedonic appreciation of dry-cured ham increases with IMF content (Ventanas et al., 2007), the opposite is observed for cooked ham (Fernandez et al., 2000). Finally, unlike IMF, the amount of subcutaneous fat almost always has a negative effect on the appreciation of meat or dry-cured hams (Dransfield et al., 2005; Ventanas et al., 2007).

2.3. Nutritional quality: the search for healthy fat

The nutritional quality of meat corresponds to its capacity to satisfy human nutritional needs: protein (including essential amino acids), lipids, vitamins (including A, E, B1) and minerals (iron, zinc, selenium) (Lebret and Picard, 2015). It is now recommended to increase energy intake in the form of lipids, which ideally should constitute 35-40% of the energy ingested (ANSES, 2016), as well as omega-3 (n-3) PUFAs, with the recommendation of 1% of energy intake in the form of ALA (C18:3), 250 mg/day of EPA (C20:5) and DHA (C22:6), and a LA (C18:2):ALA ratio less than 5 (ANSES, 2011).

Concerning total fat content, pork is lean when the visible fat is removed: 3-4% lipids in a cooked roast, and ca. 15% in a grilled rib. In comparison, the fat content of processed products is varies greatly: less than 4% in premium cooked ham, 12% in dry-cured ham, 20% in raw sausages, ca. 30% in dry sausage and 35-40% in pure pork rillettes (Ciqual, 2017).

In terms of FA composition, oleic acid (C18:1) predominates in pig muscles and adipose tissues (35-40%). The percentage of n-3 PUFAs is generally lower than that of n-6 (0.8-1.5% vs. 12-18% of total FAs, respectively), leading to a n-6:n-3 ratio of ca. 15 (Mourot, 2010).

Another important phenomenon that influences these qualities is lipid peroxidation, to which PUFAs are particularly sensitive. These are radical reactions that lead to the formation of several end products, including volatile compounds. When peroxidation is low, the compounds formed have a beneficial effect on the flavour of meat and processed products. However, high peroxidation leads to the production of toxic compounds that alter the nutritional quality, but also the colour and sometimes the flavour (Gandemer, 1999; IFIP, 2018). It is therefore essential to control oxidation phenomena in meat and products to control their sensory and nutritional qualities. The nature and proportions of the compounds formed depend on several factors: nature of the FAs, iron content, presence of oxygen, pH, etc. In processed products, mechanical (mincing, grinding) or thermal treatments, addition of salt, and storage duration promote peroxidation (IFIP, 2018). Lipid peroxidation of meat and products can be limited by the presence, in fat and lean tissues, of antioxidants provided through animal feed: vitamin E, plant extracts high in polyphenols, selenium, etc., some of which interact (Falowo et al., 2014).

2.4. Technological quality: for each product its fat

The technological quality of adipose tissues corresponds to their aptitude for processing and conservation, i.e. their consistency and cohesion (anatomically separable adipose tissues: subcutaneous, internal and intermuscular) and sensitivity to oxidation (all adipose tissues, including intramuscular).

Firmness depends on chemical composition: lipid and water content, extent of the supporting collagen framework and FA composition. A low lipid content and subsequently high water content leads to a lack of consistency of the adipose tissues, whereas a high protein content ensures a certain firmness at room temperature (Lebret and Mourot, 1998). The nature of the FAs plays a major role in the consistency of adipose tissues: the more unsaturated the FAs, the lower their melting point (< 0°C for PUFAs). Several studies have shown the predominant role of the proportion of SFAs (especially C16:0 and C18:0) compared to MFAs or PUFAs on the firmness of adipose tissues (Lebret and Mourot, 1998; Wood et al., 2003; Hugo and Roodt, 2007). The shelf life of adipose tissues is limited by the development of lipid peroxidation reactions, which are favoured by high water and PUFA contents associated with the absence of antioxidant agents (see above), and can also occur in the frozen state.

The physical (firmness, colour) and biochemical (sensitivity to peroxidation) characteristics of adipose tissues are essential to be able to process products (Hugo and Roodt, 2007). In cooked (with rind) and dry-cured hams, fat tissue that is poorly oxidisable and white is required; for dry ham, minimum subcutaneous fat thickness helps to avoid too rapid and excessive drying that would worsen the texture of the product (IFIP, 2014). For dry sausage processing, white fat with a high melting point, low melting and rancidity risk is required to optimise the processing, as well as the flavour and texture of the product. Similarly, firm fats, i.e. high in SFAs, are required to produce emulsified sausages and rillettes, while fats high in low-melting-point PUFAs hinder the binding of the fat to the lean and the homogenisation of the mixture (IFIP, 2014).

Thus, the technological qualities of pig adipose tissues involve having a high proportion of SFAs and a high lipid content (low water content). Requirements for pig feed (< 1.9% linoleic acid in the diet of animals more than 12 weeks old) and the quality of the fat ("white and firm fat") are set out in the specifications for Label Rouge pork production (Official Bulletin No. 31, 2017) as well as for other products under official quality labels. These properties are also favourable for organoleptic qualities (limiting oxidation) but go against the nutritional quality. It therefore seems that management of these antagonisms would consist of directing fat tissues to different uses: fresh products (bard for roasts, chops, etc.) or processing of raw or cooked sausages that contain whole or minced fat tissues, depending on their physical and biochemical characteristics.

3.1. Variation according to genetic type

The genetic selection performed since the early 1970s in French pig populations, oriented towards producing increasingly leaner animals, has led to the reduction of BFT by nearly 3 Phenotypic Standard Deviations (PSD), i.e. -6.3% PSD per year on average in purebred populations (Bidanel et al., 2018). This is also observed at the production stage in commercial pigs, as shown by performance test results of crossbred pigs: -10 mm BFT from 1977-1997, i.e. -40% in 20 years (Monin et al., 1998).

Since the end of the 1990s, the reduction in BFT has slowed in all selected French pig populations compared to that during the two previous decades, due to lower selective pressure on this trait, which had already reached very low levels. Unsurprisingly, the Piétrain (Pi) breed has the thinnest subcutaneous fat cover: depending on sex, 7-7.5 mm BFT adjusted to 100 kg body weight in 2017 (Figure 1). Duroc is more like a dam line, such as Large White Female and French Landrace, with 10.5-11.5 mm BFT adjusted to 100 kg body weight on average. The Large White Male line has an intermediate BFT of 9.5-10.5 mm, depending on sex. Although purebred animals have achieved very low BFT levels due to genetic selection, there is still high variability between genetic types for this trait.

The relation between BFT and IMF content has been the subject of multiple studies, both on purebred and crossbred animals, but their results are not easy to summarise, as their experimental conditions are not always similar (differences in feeding management; sex, age and weight at measurement; site and method of measurement; etc.). Ten studies that measured both mean BFT and IMF content of Longissimus muscle in animals of different pure breeds with contrasting adiposity were selected (Figure 2). The four main breeds selected in France (Large White, Landrace, Piétrain and Duroc) are represented, as is the Meishan breed, used in Sino-European crossbreeds, and two local French breeds: Gascon and Basque. For the selected breeds, recent references were favoured to limit bias due to phenotypic evolution of the traits under the effect of selection.

Among the indicators of adiposity, BFT appears to be a better indicator than IMF content for distinguishing local breeds, which are not subject to a breeding programme, from selected breeds. This observation reflects the fact that the European conventional breeds (Large White, Landrace and Piétrain) have been selected against BFT for many years, whereas the interest in selecting for IMF content is more recent.

European conventional breeds have low overall adiposity, both in terms of subcutaneous fat and IMF. While Duroc pigs have a moderate BFT and an IMF content that varies among studies (i.e. depending on the line considered), Meishan pigs have a high BFT and moderate IMF content. The French local breeds have high overall adiposity (BFT and IMF content), although IMF content varies between local breeds (Figure 2).

These studies illustrate the high variability in IMF content, which varies from 1% to over 4%, depending on the breed. Average IMF content in Iberian pigs can reach up to 10% (Pugliese and Sirtori, 2012). These results confirm that the Duroc breed has a higher IMF content than the other selected breeds. In the latter, the IMF contents are adapted to the fresh meat market and the production of cooked ham. However, the threshold of 2.5% IMF, considered the minimum necessary to produce meat with satisfactory sensory quality, is not reached.

In commercial crossbred pigs, which are intended for meat production, adiposity depends on the parental lines. In Sino-European lines, the more Meishan genes in the crossbreed scheme, the higher the IMF content and BFT (Jiang et al., 2012). In addition, comparisons between crossbred pigs from Duroc vs. Large White or Piétrain sires indicate that the IMF content of Longissimus muscle is 0.55-0.64 percentage points higher in Duroc crossbred animals (Alonso et al., 2009).

3.2. Variation according to sex type

In French collective populations, intra-genetic type variability due to sex type is observed in subcutaneous fat deposition (Figure 1): regardless of genetic type, females are fatter than entire males. This difference, well established in the literature, is also associated with a difference in the composition of adipose tissue, which contains more water and protein and less lipids in entire males than in females (Lebret and Mourot, 1998). Conversely, the adipose tissues of castrated males contain more lipids and less water than those of entire males and females, due to the increased adiposity after castration. Immunocastrated pigs, as an alternative to surgical castration of male piglets, have an intermediate adiposity between those of entire and castrated males (Batorek et al., 2012). Sex type influences the lipogenic activity of adipose tissues, which varies in the same way as the adiposity of the animals: consequently, for a given live weight, the degree of lipid unsaturation is higher in entire males than in females or castrated males (Lebret and Mourot, 1998).

Differences in IMF contents among sex types are sometimes smaller than those observed in carcass fat. However, several studies or meta-analyses show that castrated males have a higher IMF content than entire males, whereas females have an intermediate content (Trefan et al., 2013; Font-i-Furnols et al., 2019).

3.3. Variation according to anatomical location

Characteristics of fat depots, in particular their lipid content and FA composition, vary depending on their anatomical location (backfat, perirenal fat). In general, PUFAs are located preferentially in subcutaneous adipose tissue, and SFAs of endogenous origin in internal adipose tissues (Lebret and Mourot, 1998). Differences in BFT are observed along the carcass, depending on the anatomical zone: BFT profiles obtained by X-ray tomography (X-ray) show that BFT decreases from shoulder to ham, with the anatomical transition zones from shoulder to loin and then from loin to ham marked by greater decreases in BFT (Mercat et al., 2016).

Carcass muscles also show differences in IMF content: it is lower in loin than in ham, whose muscles have different IMF contents (Maignel et al., 2013; Font-i-Furnols et al., 2019). Furthermore, IMF content can vary within the same muscle depending on the anatomical site, as in the Longissimus muscle (Faucitano et al., 2004; Schwob et al., 2018). Magnetic Resonance Imaging (MRI) analysis of the entire loin shows clear rib/inter-rib variation in IMF content with, in extreme cases, up to 1 percentage point difference depending on whether the cut is centred on the rib (minimum value) or between two ribs (Figure 3). However, IMF content is relatively constant along the Longissimus muscle, except at the ends, which are fatter (Faucitano et al., 2004; Schwob et al., 2018).

The correlation between the average IMF content of the Longissimus muscle and that predicted by MRI at each position of the loin is high (R² of 0.76-0.88), with the highest correlations near the 13^th rib. A sample taken at this level thus represents the average IMF content of the loin. However, this study shows that it is possible to choose a sampling site closer to the cranial end of the Longissimus muscle to limit devaluation of the loin while maintaining good representativeness of its average lipid content (Schwob et al., 2018).

4.1. New technologies for phenotyping fat depots

The imaging technologies now available allow in vivo study of body composition and anatomical tissue distribution in pigs. Advantages and disadvantages of the main non-invasive technologies have been compared (Carabus et al., 2016): ultrasound (US), dual-energy X-ray absorptiometry (DXA), X-ray and MRI. All of these methods are used to analyse the internal composition of the animal. They provide 2D (US, DXA) or 3D (X-ray, MRI) images taken of different anatomical sites or tissues (see inset). X-ray and MRI technologies have the highest image resolution, followed by DXA and US. However, the accuracy of the images and the predictions obtained depend on several parameters: calibration of the device, data source used for prediction (thickness, surface area, volume, density, etc.), prediction equations (which may depend on genetic type, sex, weight, etc.), predicted parameter (carcass or cut weight, tissue weight, tissue volume, fat or lean content, IMF content, etc.). X-ray technology is suitable for predicting changes in carcass and cut composition during growth (Carabus et al., 2016) but is not suitable for predicting IMF content in vivo because the image quality is too low and the prediction error is too high (Font-i-Furnols et al., 2019).

Imaging technologies are also being tested in pigs to predict post-mortem the IMF content of Longissimus muscle samples at high analysis rates. For example, MRI provides the opportunity to predict the IMF content of ca. 400-500 loin samples per day with high accuracy (Davenel et al., 2012). Near InfraRed Spectroscopy (NIRS) is also a promising method for the meat industry. Using it to predict IMF content is now well established and applicable in industrial conditions (Andueza et al., 2015). Other studies have also shown the utility of NIRS measurements on the slaughter line for predicting the FA content, particularly omega-3, in pig adipose tissues (Andueza et al., 2015; Kucha et al., 2018).

4.2. Genetic determinism of adiposity traits in pigs

Additive genetic variance, i.e. the part of phenotypic variation in a trait that depends on genes and thus on which selection can act, is high for adiposity traits. For example, BFT has high heritability (≥ 0.44) when measured in vivo by US (Table 1), which explains why it is relatively easy to change through selection. While meat quality traits are moderately heritable (0.10-0.30), IMF content is an exception, with mean heritability close to 0.50. Thus, both BFT and IMF content are highly heritable, making them suitable traits for classic breeding methods.

The genetic correlation between BFT and IMF content, which induces the indirect selection response, is moderate, ranging from 0.13-0.40 (Table 1). Consequently, selection to increase IMF content, which is often the goal sought for the loin, will cause BFT to increase slightly. Thus, there is moderate antagonism between carcass quality (lean content) and meat quality indicators, including IMF content (Ciobanu et al., 2011). However, from a phenotypic viewpoint, it is possible to decrease the fat content of the carcass while limiting the decrease in IMF content through constrained selection (Ros-Freixedes et al., 2013).

Some of the genetic variability in BFT and IMF content is explained by certain genes that have variable effects on these traits. For example, several genes whose polymorphism is associated with variation in adiposity are known to influence BFT, such as HAL, IGF2 and MC4R (Ciobanu et al., 2011; Mercat et al., 2012), or on IMF content, such as H-FABP (Gerbens et al., 1999). In addition, since the early 2000s, many studies have identified Quantitative Trait Loci (QTLs) that explain some of the genetic variability in traits. Indeed, more than 3,000 QTLs related to criteria of adiposity (BFT (n = 408), fat content of carcass or cuts, IMF content (n = 652), marbling, etc.) have been identified, and almost twice as many QTLs that influence FA composition (content of SFAs (n = 800), MUFAs (n = 729), stearic acid (n = 796) or oleic acid (n = 762)) are referenced in the PigQTLdb database (Hu et al., 2005; https://www.animalgenome.org/cgi-bin/QTLdb/SS/index). QTLs related to adiposity traits with a high significance level have been identified on all pig chromosomes. Several examples are detailed in the literature (Ciobanu et al., 2011).

4.3. Use for selection purposes

Objectives of genetic improvement are defined by breeding companies to produce animals adapted to market needs, which differ among countries. In France, the carcass payment scale is based on carcass weight and the muscle content of the cuts; nearly 75% (by volume) of pork is processed by the pig industry into a multitude of products, among which cooked ham and cooked meats account for 27% of pork processed products (IFIP, 2019). This specificity of the French market pushes breeders to reduce carcass adiposity as much as possible, both in terms of subcutaneous and intramuscular fat, to produce commercial pigs with a low fat:lean ratio and a low degree of marbling. However, there are many and sometimes contradictory expectations in the pork industry, including about the quantity and quality of fat (e.g. the need for higher or lower IMF content depending on whether fresh-meat, cooked or cured-ham processing is considered).

Genetic relationships among the traits of interest should be considered when defining breeding objectives. Breeding programmes give preference to traits that are measurable in vivo because they can be recorded for all selection candidates (not just slaughtered siblings) using non-invasive methods. In addition, collecting data is simpler than taking measurements in slaughterhouses, which requires individual traceability of each carcass or cut.

Thus, in the pig breeding programmes on the French market, genetic improvement in adipose tissue is based mainly on measuring of two major traits (Table 2): BFT, measured exclusively in vivo by US, and the IMF content of the Longissimus muscle, estimated by different methods on either live animals or carcasses, as described above. Although some indicators of adiposity are also recorded in dam lines to estimate the genetic value of reproducers, the selection effort on BFT and IMF content is greater in sire lines.

** NPPC : National Pork Producers Council

As early as the 1950s, selection for BFT was greatly facilitated by the development of non-invasive, US-based methods to measure BFT accurately in live animals (Monin et al., 1998). More recently, tools based on the same technology have been developed to estimate IMF content in live pigs (Newcom et al., 2002; Maignel et al., 2009). It is an interesting alternative to estimating IMF content in meat using an expensive chemical method, which devalues the carcass and is not feasible at a large scale, as well as to assessing marbling visually, which has many constraints (subjective evaluation, little discriminating when the average content is lower than 2%, traceability at the slaughterhouse, etc.). Other imaging technologies, described above, allow analysis of the internal composition of the animal (X-ray, DXA, MRI) but have some constraints in using them for breeding purposes: need to fast pigs and administer anaesthetics or sedatives before taking measurements, non-portable and expensive equipment, and need for an experienced operator to acquire images.

In addition, BFT is measured at the slaughterhouse on carcasses to classify them, at the dorsal level with the fat/lean evaluation system or, more recently, at the split (loin-ham junction) with the Image-Meater (Daumas, 2008; see inset). Tools for tracing individual carcasses in slaughterhouses (RFID chips), currently being implemented, should soon facilitate collection of routine measurements from a larger number of animals. These developments will offer new perspectives on carcass composition for breeding programmes.

Finally, genomic tools are currently available to identify new regions of the genome that have significant effects on traits of interest, via genome-wide association analyses, in order to include this information in breeding programmes and thus increase the accuracy of estimated breeding values of reproducers and selection candidates. Genomic selection is particularly useful for polygenic traits that are costly to measure and require slaughtering animals, such as indicators of meat or carcass quality and adipose tissue composition (Samorè and Fontanesi, 2016).

NOTES

• This article was presented at the 51st Journées de la Recherche Porcine (Schwob et al., 2019).