Biodiversity and genetics of beef quality, a review

Abstract

Preservation of biodiversity and genetic improvement of livestock populations are often considered to be antagonistic. Biodiversity affects meat quality (MQ) on different levels: among species (only three species yield 88% of the world’s meat); among breeds within species (there are more than 3000 cattle breeds worldwide, but about half are at an unknown risk of extinction, and only one-fourth of the others are not endangered); among animals within breed (in cattle populations with millions of individuals the median effective size is equivalent to only about 100 unrelated animals); and between alleles within animal genes (the inbreeding coefficients of individual animals are increasing with selection, and particularly genomic selection [GS]). Beef quality traits are very particular because they cannot be directly measured on living animals. The genetic improvement of beef quality can be pursued by different techniques. In chronological order, they are: phenotypic selection, which has created breed differentiation (not useful for MQ traits); selective breeding (heritability of beef quality traits varies considerably according to breed, trait and conditions; indirect selection through NIRS predictions, etc., (could be useful); the fixation of major gene mutations (the myostatin gene for double muscling, calpain (CAPN)-calpastatin (CAST) genes for beef tenderness; diacylglycerol O-Acyltransferase 1 for beef marbling, etc.); genomic and other omic approaches (strong increase in scientific studies, genome-wide association studies (GWAS), GS, gene network identification, etc.); the cloning of animals (not useful); and the cloning of tissues (cultivated meat).

HIGHLIGHTS

• The biodiversity of beef cattle populations is threatened by globalisation, intensification, and (genomic) selection, which increase inbreeding;

• Indirect phenotyping (NIRS) and genomic selection (GS) will allow the genetic improvement of beef quality traits.

• Animal cloning has no practical use, and cell cloning (cultured meat) needs further research and ethical debate.

Keywords:

• Meat quality
• heritability
• bovine meat
• hypertrophy
• inbreeding

Introduction

In this article, biodiversity refers to the genetic variability among all living beings (species, breeds and individuals), and genetics is the science of genes and heredity. In farm animal breeding, genetics is used especially for the genetic improvement of domestic populations, which is often considered antagonistic to maintaining biodiversity as it is based on the selection and breeding of the best animals, which reduces the genetic variability, and hence reduces biodiversity.

Beef quality traits are very particular because, unlike traits related to the growth, efficiency, conformation, reproduction and disease resistance of animals, they are not measurable on living animals, only on slaughtered animals, and cannot therefore be evaluated directly on animals destined for reproduction (candidate sires and dams). Further difficulties arise from the fact that these traits can be measured only once on a given animal when it is at the age at slaughter specific to the production system it is destined for (veal calves, young bulls, steers and heifers, culled cows).

The objective of this review article is obviously not to analyse in depth every aspect of this complex subject, but to provide an overall view of the various issues involved and to briefly present their relationships from a holistic perspective.

This review article has been structured in three main sections related to:

• Beef quality traits;

• Biodiversity and meat quality (MQ);

• Genetics of beef quality.

The search of relevant scientific articles was done using Scopus database. The objective of the work was not to review a large number of articles on every aspect of the issue, but to select relevant articles for outlining the most important concepts included.

Beef quality

MQ is a very complex matter with many different aspects and dozens of traits, as reviewed by Prache et al. (2022). As a first step, we can distinguish between traits related to the quality of meat products, on the one hand, and those that depend on production processes and that are characterised by a more comprehensive approach, on the other (Table 1). Sensory and technological traits can be assessed by instrumental characterisation or by a trained panels. The processing and overall quality traits of beef often receive certification (labelling) from specialised public or private organisations according to official regulations.

Table 1. Importance of genetic variability (Biodiversity) and genetic improvement (Genetics) for the phenotypic variability of the major traits affecting the product, process and overall quality of beef meat.

Meat quality trait	Related to:	Importance of:
		Biodiversity	Genetics
Product quality:
Chemical composition	Degree and distribution of marbling, leanness	^*****	***
Histological traits	Muscle fibre type, adipocytes	^****	*
Texture	Tenderness, connective tissue, enzymes	***	**
Colour	Dark or pale meat, haem iron,	**	**
Purge and cooking losses	Exudative meats, DFD	**	*
Ageing, shelf life	Fat oxidation, microbial alteration	**	*
Sensory properties	Smell, taste, juiciness, etc.	***	**
Volatile compounds	VOCs, smell	??	*
Hygiene	Safe or harmful meat	–	–
Human health claims	Favourable or unfavourable nutrients	^****	**
Process quality:
Animal welfare	Dehorning, castration, dystocia, etc.	**	***
Antibiotic free	Disease resistance and prevention	**	*
GMO free	Crop cultivars and origins	–	–
Carbon neutrality	Enteric methane, slurry management	*	**
Environmental impact	GHG, land use, water, pollutants, etc.	**	*
Biodiversity friendly	Rare breeds, no art. insemination	^*****	**
Overall quality:
Designation of origin	PDO, PGI, Slow-Food praesidia, etc.	^****	–
Farming system	Organic label, mountain product, etc.	***	**
Ecosystem services	Landscape, fires and social issues	***	*
Religious beliefs	Food taboos, halal, kosher, etc.	–	–
Cultural features^a	Animal rights advocates, veganism and vegetarianism	–	**
Gastronomic heritage	Typical products and recipes	***	*

^aHeritable in humans (Wesseldijk et al. 2023).

*, **, ***, ****, *****: asterisks represent a subjective scoring of the importance of Biodiversity and Genetics on the meat quality traits; ??: not known; DFD: dark, firm, dry meat; VOC: volatile organic compounds; GMO: genetically modified organism; GHG: greenhouse gases; PDO: protected designation of origin; PGI: protected geographical indication.

In Table 1, the most important quality traits are listed with a subjective scoring of the importance of biodiversity and genetics in explaining their phenotypic variability. In the next sections, various aspects of, firstly, biodiversity, and then of genetics/genetic improvement with respect to beef quality will be discussed.

Biodiversity and meat quality

Biodiversity is defined by the United Nations as ‘the variety of life on Earth, it includes all organisms, species, and populations; the genetic variation among these; and their complex assemblages of communities and ecosystems. It also refers to the interrelatedness of genes, species, and ecosystems and in turn, their interactions with the environment.’ (United Nations Environment Programme 2010). Here, especially genetic variation is discussed, whereas relationships with habitat and ecosystems are not.

Biodiversity affects beef quality at different levels. From the most general to the most particular, we can list the four major levels:

• biodiversity among species;

• biodiversity among breeds (lines, populations, etc.) within species;

• biodiversity among animals within breed (line, population, etc.);

• biodiversity among alleles within animal.

In the following paragraphs, we will briefly discuss these four levels of biodiversity in relation to beef quality.

Biodiversity among species: extinction of species

The number of species used as meat sources has dramatically decreased from prehistory to now. Our hunter-gatherer ancestors of the Palaeolithic era consumed tissues from several hundred different animal species (invertebrates included). Dozens of species are still hunted and fished for food. However, with the birth of agriculture and pastoralism and the domestication of animals and crops in the Neolithic era, the sources of animal foods were concentrated in a small number of species, a process that is still ongoing.

According to the FAO (2020), three species (swine, chicken and cattle) account for 88% of all meat produced worldwide (Figure 1), with bovine species representing approximately 20% of total. It should be noted that the FAO’s Domestic Animal Diversity Information System (DAD-IS) lists 35 mammal and avian species used for food production (Bittante 2011). Although these species, other than the major three, represent only 12% of meat production, the proportions vary considerably in different areas of the planet. For example, other species represent only 4% of total meat production in the Americas, but about 30% in Africa. It is evident that the development of modern agricultural systems has favoured the concentration of rearing on the most efficient species, with a huge loss in biodiversity in terms of the number of species reared. This concentration is even more acute in the case of milk production, where a single species accounts for almost 90% of the world’s production of milk and dairy products.

Figure 1. Contribution of different species to the worldwide production of meat (FAO 2020).

Alongside the reduction in species being reared, we are seeing a progressive globalisation of food habits with the (irreversible?) loss of many cultures related to typical local animal products, and traditional recipes and gastronomy. Loss of biodiversity is therefore also loss of culture.

Biodiversity among cattle breeds: extinction of breeds and crossbreeding

Unlike biodiversity among species reared for food production, biodiversity in terms of the number of breeds within species is very large. The FAO’s DAD-IS reports about 15,000 breeds of 35 species of mammals and birds used for food production (Figure 2) (Bittante 2011).

Figure 2. Number of breed entries of mammals and birds reared for food production worldwide retrieved from the FAO’s Domestic Animal Diversity Information System.

The number of bovine breed populations censed at the national level represents about 20% of all breeds of all species. Almost half of the about three thousands of national cattle breed populations recorded in 182 countries by DAD-IS are local breeds distributed in single countries, while the other half belongs to 110 international and 98 regional-transboundary breeds distributed in several countries, with a total of 1435 individual cattle breeds recorded (Commission on Genetic Resources for Food and Agriculture 2018).

It is worth noting that all bovine breeds are used for meat production, except for the majority of those reared in Southern Asia where a religious taboo forbids the consumption of beef meat (Narayanan 2023). A considerable proportion of bovine meat is not obtained from breeds classified as ‘beef breeds’ (or their crossbreds), but from dairy breeds, dual purpose breeds and the many unspecialised local breeds and populations characterised by their ability to adapt to harsh environments and low-input farming systems. With the exception of calves produced in intensive dairy farms, the vast majority of calves destined for beef production are produced by cows reared in extensive or semi-extensive farming systems, where pasture often plays an important role and adaptation to local conditions is of great value. A highly variable proportion of those calves, according to different areas of the planet, are destined for intensive fattening in feedlots or fattening barns. It is therefore evident that beef quality is highly variable across and within different regions of the world, and breed can be considered the primary source of variability.

This is a very different situation from that characterising milk production, where one breed (of one species) accounts for a very large and growing proportion of the world’s milk production.

In this context and particularly in the more extensive conditions, crossbreeding is a common practice in beef production due to the favourable effect of heterosis, especially on reproductive and fitness traits. However, the biodiversity represented by the large number of bovine breeds is endangered by widespread crossbreeding practices that do not adequately maintain the parental pure breeds, and by the intensification and globalisation of farming systems using a few ‘improved’ beef breeds. About half of the world’s cattle breeds listed in the FAO database have an unknown risk status; of the breeds with a known status, only about one-fourth are not at risk, half are at various levels of risk of extinction, and one-fourth are now extinct (Commission on Genetic Resources for Food and Agriculture 2018).

It is clear that preservation of the genetic resources represented by local breeds is an important objective. First, because they represent a reservoir of genes that could be valuable for maintaining or improving the resistance and resilience of many cattle populations to climate change and environmental modifications. MQ and value could gain from a large, and often unknown biodiversity whose benefits are for the most part not yet known and, therefore, remain unexploited. It is also important to note that saving local genetic resources is not simply a question of storing semen, oocytes, embryos and stem cells in liquid nitrogen tanks (ex situ conservation), but also of protecting farms and farmers, especially through the development of sustainable value chains for small-scale livestock producers (FAO Fernando GL& M 2019). Local breeds are, in fact, the source of many typical products, traditional recipes and gastronomy, and may also underpin traditions and culture (Marsoner et al. 2018; Ovaska et al. 2021).

Biodiversity among individual bovines: relatedness

The third level of genetic variation (biodiversity) is represented by the variability among different animals of the same population. It is well known that relatives share a portion of their DNA from their common ancestors, so they are more similar to each other than nonrelated individuals, i.e. they are less genetically variable. One widely used measure of genetic similarity for pairs of relatives is gene identity-by-descent (IBD) sharing. Two nonrelated individuals should have non-common ancestors. This means that, in practice, nonrelated individuals do not exist: every individual has two parents, four grandparents, … 1024 ancestors 10 generations back, more than 1 million ancestors 20 generations back, more than 1 billion 30 generations back, … They cannot always be fully unrelated animals (or humans).

The contribution of an individual to the genetic variability of a population is assumed to be 1.00 if it is not related to any other individual in the population; otherwise, it is <1.00 depending on the number of relationships and the level of relatedness it has with the others. The sum of the contribution of the actual independent genetic variability of each individual defines the ‘effective population size’ (N_e).

It is worth noting that, with the exception of rare and very rare breeds, the N_e has little relationship with the census size (N_c) (Figure 3). This is mainly because large populations have often expanded in recent decades from a relatively modest number of founders.

Figure 3. Median effective population sizes (Ne) of cattle breeds calculated on the basis of genealogical relationships (DF, 130 breeds) and the genomic linkage disequilibrium (LD, 60 breeds) according to census size (Nc) (adapted from Hall’s (2016) data).

At similar N_c, the N_e of equine and ovine populations is much larger than that of bovine populations, which could have to do with the very modest use of artificial insemination in the former species (Hall 2016). On the other hand, the N_e of porcine populations is similar to bovines, probably due to their high prolificacy and, therefore, the high frequency of full sibs in the population.

The N_e of cattle breeds with several million individuals, but that have undergone artificial insemination for many generations using a large number of semen doses from a moderate number of bulls obtained from a much lower number of bull sires, as practiced in many countries of the world, is equal to a few dozen unrelated animals. This is the case of Holstein-Friesians. The N_e of this breed is rapidly decreasing and is now below 50 (Makanjuola et al. 2020), which is considered the critical value below which the population loses fitness and viability (FAO Commission on Genetic Resources for Food and Agriculture 2015).

The widespread use of natural mating and a much lower selection intensity are responsible for the better situation of many beef populations, and also of many local breeds, despite the modest number of individuals. This is because the average relationship coefficient (the proportion of alleles of common origin that are shared) among the individuals of the population is much lower.

Biodiversity of alleles within individuals: Inbreeding

The presence of many relatives in a population increases the probability of matings between relatives, which is why there are many individuals (their offspring) with less genetic variability in their genome, i.e. a greater than average proportion of homozygote loci. Inbred animals are the result of matings between relatives with a high coefficient of relationship (incest) or a moderate/low coefficient of relationship. The inbreeding coefficient of an individual is, by definition, half the relationship coefficient of its parents (Wright 1922). In the case of random matings, at the population level, the average inbreeding coefficient of all individuals is half the average relationship coefficient of their parental generation.

Again, the Holstein breed is the supreme example of the risks to genetic variability and the animal’s survival rate and performance caused by modern technologies and selection intensity in a large population. The steep increase in the genomic inbreeding coefficient of Holsteins in the last decade (Figure 4) is attributed to the intensive use of genomic selection (GS; Guinan et al. 2023). It is worth noting that the proportional increase in the average inbreeding coefficient is steeper than the proportional improvement in productive traits (milk yield, fat and protein %), whereas fitness traits (productive life, pregnancy rate and somatic cell score) benefitted.

Figure 4. Trends in the average genomic inbreeding coefficients of Holstein bulls and cows according to their year of birth (simplified from Guinan et al. (2023)).

Again, the situation with regard to beef and local breeds is not as critical because genomic tools are not so widely used. This allows us to combine the use of GS with the development of breeding schemes that have a greater regard for genetic variability and with the adoption of better reproduction strategies, possibly favoured by specific genomic tools.

It is also clear that the best way to limit the dangers of inbreeding is to expand cross-breeding (heterosis could be considered the opposite genetic effect to inbreeding depression). However, it is also clear that the long-term genetic improvement of beef production (and of MQ traits) must be built on the genetic improvement of many parental breeds and the maintenance of their genetic variability.

Genetics of beef quality

The first application of genetic knowledge to a farm animal population is to improve some of the characteristics of animals. Genetic improvement of beef quality can be pursued using different techniques. In chronological order, we can list:

• phenotypic selection;

• selective breeding;

• fixation of gene mutations;

• genomic selection;

• the cloning of animals;

• the cloning of tissues.

In the next section, we will consider these techniques in turn and their application for the improvement of MQ.

Phenotypic selection: differentiation of cattle breeds

Since the Neolithic era, with the domestication of plants and animals and the birth of agriculture and pastoralism, the evolution of domestic animal populations has been linked to human migrations, needs, societies, cultures and technologies. Natural selection was integral to it, but was then replaced by much more intensive (artificial) selection of candidates for reproduction by humans. The selection criteria for several millennia were the appearance and conformation of the candidates, but also their temperament and certain productive information on them or their parents. This phenotypic selection led to the differentiation of species subpopulations into what we now call breeds. With the exception of some ‘synthetic’ breeds, all existing dairy, beef, dual-purpose and local cattle breeds originated from phenotypic selection. This was a viable practice and it led to (very slow) genetic differentiation and improvement while largely maintaining genetic variability within and between breeds.

The fact that the majority of the traits important for beef production (adult body weight, growth rate, muscularity, skeletal frame, fatness, etc.) are moderately to highly heritable meant that phenotypic selection was effective in differentiating many beef breeds for these traits. Unfortunately, MQ traits are not measurable on live animals, and sometimes they are not measurable at all. Phenotypic selection indirectly affected (indirect selection) only those MQ traits that are genetically correlated − positively or negatively − with some trait visible on live animals (like meat marbling with body condition score, or beef lightness with muscularity). However, the vast majority of beef quality traits are not related to visible characteristics of live animals, so phenotypic selection has had only a minor effect on them.

Selective breeding of beef breeds: genetic variation, heritability, indirect prediction and breeding values of beef quality traits

In addition to pre- and post-slaughter conditions, meat characteristics are directly related to the muscle biology of live animals, which is regulated by genetic factors, and also by nutritional and biological factors (Hocquette et al. 2005; Mortimer and Przybylski, 2016; Leal-Gutiérrez and Mateescu 2019; Mwangi et al. 2019).

Heritability of beef quality traits

The scientific literature contains many estimates of the heritability of several beef quality traits. Unfortunately, there is often a very large variability in these estimates, which is due to many factors, particularly those related to the animals (size of the sampled population, breed/breed combination, sex, age, genetic population structure, connectedness of sires, etc.), the environment (climate, farming system, finishing, diet characteristics, etc.), pre-slaughter conditions (fasting, transport to abattoir, mixing, etc.), slaughter and post-slaughter conditions (slaughtering procedure, cooling, post-slaughter carcase treatments, length of ageing, etc.), the meat analyses procedures (meat sampling, sample storage and transport conditions, instrument type and set-up or panel characteristics, repeatability and reproducibility of measures or scores, etc.), and the statistical methodology (data editing, the use of genetic or genomic relationship matrix among animals, variance components estimation method, software used, etc.). An example of the variability in genetic parameter estimation (of shear force and tenderness of beef meat) reported in literature is depicted in Figure 5, which summarises the heritability coefficients collected in an early review by Burrow et al. (2001).

Figure 5. Distribution of heritability estimates of Warner-Bratzler shear force (in blue, 44 studies) and of tenderness scores (in red, 19 studies) of beef meat (data taken from Burrow et al. (2001) review).

It is beyond the scope of this study to analyse and discuss all these causes of variation in detail. Review articles have recently been published also on meat colour traits (Mancini and Ramanathan 2019), and meat volatile organic compounds (Ni et al. 2022).

Clearly, there is a need to obtain heritability estimates specific of the population of interest or in very similar populations. There are, in particular, two major groups of estimates. The first consists of those obtained in extensive, pasture-based farming systems (mainly in the Americas and Australia) in steers and heifers with a relatively high degree of fatness (crosses from British breeds and Bos indicus breeds), often finished in feed-lot conditions, like the majority of studies reviewed by Burrow et al. (2001). The second group consists of estimates obtained mainly in Europe in young bulls of lean continental breeds (including double-muscled breeds and crosses), fattened and finished indoors in intensive farming systems. These two systems also differ in the value that the local market places on MQ traits: In the first case, marbling is the most important trait, in the second, it is leanness. This means that genetic improvement programmes can have very different objectives.

Table 2 presents a summary of the heritability estimates of several beef quality traits obtained in the last two decades, mainly in the second farming system.

Table 2. Summary of the descriptive statistics and heritability estimates of beef quality traits obtained from studies carried out in the last 20 years (22 studies^a, 31 populations and 161,574 head).

	Estimates N	Descriptive statistics:			Heritability:
		Median	Mean	SD	Median	Mean	SD
Heads	31	1254	5984	15,267	–	–	–
Raw meat samples
Fatness traits:
IMF, %	8	1.88	2.04	1.25	0.25	0.30	0.18
Marbling score	12	^b	^b	^b	0.30	0.31	0.19
Colour traits:
L*	15	36.73	37.04	4.23	0.18	0.21	0.10
a*	10	22.64	20.22	5.48	0.13	0.19	0.22
b*	10	11.45	11.85	2.78	0.09	0.09	0.04
C*	4	22.14	22.05	6.74	0.15	0.12	0.06
H*	4	37.96	34.27	11.35	0.31	0.35	0.27
Cooked meat samples
Cooking loss, %	11	21.78	21.18	4.92	0.08	0.11	0.07
Texture traits:
WBsf, N/cm^b	24	^b	^b	^b	0.24	0.23	0.12
Tenderness score	9	^b	^b	^b	0.17	0.21	0.14

^aAllais et al. (2014); Hornyák et al. (2008); Zwambag et al. (2013); Boukha et al. (2007, 2011); Savoia et al. (2019); Cecchi et al. (2004); Crews et al. (2004); Do et al. (2016); Gill et al. (2010); Johnston et al. (2003); Minick et al. (2004); Wolcott et al. (2009); Berry et al. (2021); Magalhães et al. (2019); Bedhane et al. (2019); Londoño-Gil et al. (2022); Feitosa et al. (2017); Grigoletto et al. (2020); de Nadai Bonin et al. (2021); Bhuiyan et al. (2017); (Magalhães et al. 2019).

^bData not reported due to heterogeneity in the technical procedures adopted in different studies.

IMF: intramuscular fat content; WBsf: Warner-Bratzler shear force; SD: standard deviation.

It shows that fatness and tenderness traits tend to have a higher heritability than cooking losses. Among the colour traits, the lightness (L*) and the hue (H*) are more heritable than the others. The variability in the estimates is always very large.

Genetic correlations of beef quality traits

If the estimates of heritability coefficients are variable, those of the genetic correlations within beef quality traits, and between these and other important traits, are even more variable and less accurate, so here, too, specific estimates for different populations are needed.

There are two main categories of genetic correlations relevant to the genetic improvement of beef populations. The first concerns the correlations among the different beef quality traits. Their synergistic or antagonistic relationships could have a strong effect on the overall quality of beef in response to a given selection programme, depending on the traits included in the selection index and their weights. The second category concerns the genetic correlations between the quality traits and the traits included in the overall selection index. This is fundamental for understanding the long-term effects on MQ of the current selection system.

Only some of the studies summarised in Table 2 also report estimates of some genetic correlations. Looking only at the correlations among the most important traits (intramuscular fat content [IMF]; Warner-Bratzler shear force [WBSF]; L*, meat lightness; and CL, cooking weight loss), we found low positive and negative estimates for each of them, and their average value is always lower than the corresponding standard deviation: −0.16 ± 0.42 for the genetic correlation between IMF and WBSF; +0.13 ± 0.24 for IMF-L*; −0.08 ± 0.30 for WBSF-L*; and +0.23 ± 034 for L*-CL.

Genetic correlations between beef traits and the traits included in the selection index obviously differ in different populations according to the traits selected and the characteristics of the population. As an example, in the Piemontese breed, the majority of genetic correlations between beef quality traits and other traits directly or indirectly included in the selection index are moderate/low, with some exceptions: carcase daily gain is favourably correlated with meat lightness, while carcase muscularity is correlated favourably with meat purge loss, but unfavourably with meat shear force (Savoia et al. 2019 and 2021).

Indirect selection for beef quality

Different countries of the world have developed systems for grading carcases, generally based on muscularity and fatness (Commission of the European Communities 1982), but, with the exception of Australia (Bonny et al. 2018), they do not include beef quality traits, with some exception for marbling (USA, Japan, Korea, etc.). The major fundamental problems of establishing a genetic programme for improving the quality of meat from every animal species are: i) that the gold standard methods (wet chemistry, physical measures and sensory evaluations) used for measuring MQ require carcase sampling, are expensive, time-consuming, and often have only moderate repeatability and reproducibility; and ii) that, consequently, MQ traits are not directly measurable on live animals.

In the first case, it means that gold standard methods cannot be used routinely for estimating breeding values at the population level, although they could be used on smaller population samples for research, or to estimate variance components/heritability, or to create calibration datasets for setting up indirect secondary or genomic methods of predicting MQ traits. In the second case, bull progeny testing would be mandatory, although these data could be used for pedigree and/or genomic predictions through indirect selection.

The various alternative indirect routes available to improve MQ traits at the population level are summarised in Figure 6.

Figure 6. Schema of approaches to improve meat quality (MQ) through: direct genetic selection from reference meat quality phenotypes at the nucleus level (red pathway); indirect genetic selection from NIRS MQ predictions (blue pathway); direct genetic selection of NIRS absorbances and genetic prediction of MQ (green pathway); indirect genomic selection through SNP MQ predictions (orange pathway) and direct genomic selection (brown pathway) (modified from Bittante, Savoia et al. (2021)).

Rapid, high-throughput, low-cost phenotyping of meat quality traits

As with predicting the composition and technological properties of milk and dairy products, for meat, too, the most commonly used alternatives to gold-standard wet-chemistry analytical methods are those based on visible infra-red (Vis-NIRS) spectrometry.

Many review articles have been published on this topic, but, interestingly, few of them deal with the ‘genetic’ use of infra-red predictions. As an example, we tested the prediction of beef quality traits in the Piemontese breed, measured by reference ‘gold standard’ laboratory methods, and by visible and infrared spectrometry using two different portable instruments (Savoia et al. 2020). With these instruments, spectra can be acquired in the abattoir from the freshly cut surface of the meat after separating the half carcase into two quarters, without any need for meat sampling, thus yielding real-time prediction results. The two spectrometers compared represented the two extremes of those available: one (Vis-NIRS) covers a spectrum twice as wide as the other (Micro-NIRS) with 12 times the wavelength density, but is also 40 times heavier and costs about 10 times more. The trial under operational conditions showed that both spectrometers were able to capture the major sources of variation in most of the MQ traits (Savoia et al. 2020). Application of these two spectrometers at the genetic level (Savoia et al. 2021) confirmed that the most important factor for selection purposes is not the accuracy of the predictions (phenotypic correlations between the predicted and measured MQ traits), but rather the genetic correlations between the predicted and measured traits (Figure 7).

Figure 7. Genetic (r_a, dark colours) and residual (r_e, light colours) correlations between reference ‘wet chemistry’ beef quality traits and their predictions obtained with a transportable visible-near infrared (Vis-NIRS, in blue) or a very small portable near-infrared (Micro-NIRS, in green) spectrometer (data taken from Savoia et al. 2021). (L* = lightness; a* = redness index; b* = yellowness index; C* = chroma; H* = hue; PL = beef purge loss; CL = beef cooking loss; WBSF = Warner-Bratzler shear force of cooked meat).

As Figure 7 shows, the genetic correlations are almost always higher than the residual correlations, bearing witness to the ability of infrared predictions to capture that part of the genetic information underlying the quality of beef meat. This is also evident from the fact that infra-red predictions of beef quality traits have been found to be heritable (Figure 8).

Figure 8. Heritability estimates of beef quality traits measured using reference ‘wet chemistry’ methods (dark red) and their predictions obtained with a transportable visible-near infra-red (Vis-NIRS, in blue) or a very small portable near infrared (Micro-NIRS, in green) spectrometer (data taken from Savoia et al. (2021)). (L* = lightness; a* = redness index; b* = yellowness index; C* = chroma; H* = hue; PL = beef purge loss; CL = beef cooking loss; WBSF = Warner-Bratzler shear force of cooked meat).

It is worth noting that the reference methods for analysing MQ often yield larger heritability coefficients than the corresponding infrared predictions. The real problem, however, is that reference methods cannot be applied at the population level as they are unacceptably complex and time-consuming, and sampling and analysing the meat is costly, whereas infrared predictions can be obtained in real time in the abattoir without the need for meat sampling. The heritability of predicted traits relies on the fact that the absorbance by the meat surface of all the wavelengths measured by both instruments is affected by a genetic component: while the heritability of absorbance in the near-infrared portion of the spectrum varies between 3% and 13%, in the ultra-violet and visible portions of the spectrum it is much higher, reaching 40% and above (Bittante, Savoia, et al. 2021).

Fixation of gene mutations: the major genes of beef quality

A relatively simple way to genetically improve a population is the fixation of a favourable allele of a ‘major gene’. A major gene is a single gene, often with simple Mendelian inheritance, that has a ‘wild’ allele and at least one mutant allele that exert a strong effect on a given trait (Lambe et al. 2015).

A well-known example in pigs is the Halothane/PSE gene, whose mutant allele is responsible for the so-called ‘pale, soft, exudative’ (PSE) meat (Barbut et al. 2008). It is worth noting that the opposite type, ‘dark, firm, dry’ (DFD) meat, is not caused by a major gene, but instead has an additive genetic background (many independent genes, each having a small effect) that increases the predisposition to this defect, in cattle and pigs, caused mainly by glycogen depletion as a consequence of stress and physical activity before slaughtering (Ijaz et al. 2020).

There are also examples of major genes in cattle directly or indirectly affecting MQ (Pećina and Ivanković 2021). The most important are:

1. MSTN, the myostatin gene, or the ‘double-muscle gene’;

2. CAPN1 and CAST, the calpain and calpastatin genes that affect tenderness;

3. DGAT1, the diacylglycerol O-acyltransferase 1 gene affecting marbling of beef.

MSTN, or the myostatin gene and the muscularity of beef and other traits

The myostatin protein is a regulatory factor (Growth and Differentiation Factor 8, [GDF8]) in muscle that limits its maximum growth (Kobolák and Gócza 2002). The gene responsible for myostatin production (MSTN or GDF8, chromosome 2) also controls the activity of other hormones, particularly growth hormones (GH). Aside from the normal ‘wild’ allele (‘+’), several mutations have appeared in different species and breeds (Bellinge et al. 2005), in some cases limiting or blocking the negative regulatory function of the gene. This leads to increased muscle growth resulting in muscle hypertrophy and/or hyperplasia (double-muscling) (‘mh’ alleles).

The effect of the mutated mh allele was well demonstrated through a two-generation crossbreeding experiment (Short et al. 2002) in which F2 crossbreds were obtained by randomly mating F1 bulls and cows, all heterozygotes for the mh gene (mh/+) having been sired by homozygote mutated Piemontese sires (mh/mh) and conventional dams (+/+). The F2 steers and heifers were all 50% Piemontese, but could have inherited 0, 1 or 2 copies of the Piemontese mutated mh allele. The main results are summarised in Table 3.

Table 3. Summary of the major differences in F2 crossbred steers and heifers obtained from F1 bulls and dams (heterozygotes mh/+) sired by Piemontese bulls (homozygotes mh/mh) according to the number of mutated mh alleles inherited from the Piemontese grandsires.

	Genotype of F2 at MSTN locus:			mh/mh vs. +/+	mh/+ vs. (mh/mh + +/+)/2
	+/+	mh/+	mh/mh
Birth weight, kg	35.7	37.0	40.1	***	–
Carcase weight, kg	269	278	291	***	–
Dressing, %	57.9	59.7	63.2	***	***
Longissimus dorsi area, cm^2	74.3	86.4	109.0	***	***
Marbling score	6.0	5.4	4.0	***	***
Fat depth, mm	6.3	5.6	2.6	***	**
Shear test, kg	4.2	4.1	4.0	–	–
Gain efficiency, g/Mcal	43.7	42.5	42.4	–	–
Primal cut efficiency, g/Mcal	12.6	13.2	15.2	***	–

Data summarised and adapted from Short et al. (2002).

**p < 0.01; ***p < 0.001.

It is immediately evident why MSTN is a ‘major’ gene: where animals all have the same grandparental breed composition, the mutated mh allele has a strong effect on many aspects of their appearance, reproduction, fitness, tissue development and MQ. It is also clear that the mutant allele is not recessive but for some traits has an additive effect (mh/+ heterozygotes are no different from the average of the two homozygotes) and for others a partial additive effect (mh/+ is more similar to +/+ homozygotes than to mh/mh homozygotes).

A large American project (the Germplasm Evaluation programme) studied in detail the effect of the mh mutation on MQ (Wheeler et al. 2001) and found that the mh allele improves tenderness, ease of fragmentation and the amount of connective tissues in four different muscles, but not juiciness, beef flavour intensity and cooking losses.

Several different mutations of the MSTN gene (mh alleles) have appeared in many breeds of cattle, including dairy breeds (Dunner et al. 2003). Mutant heterozygote animals were often excluded from reproduction because of unfavourable side effects, the most important of which is an increase in calving difficulty (Short et al. 2002); double-muscled animals also have greater nutrient requirements (Schiavon et al. 2012; Schiavon and Bittante 2012; Fiems et al. 2015). Two former dual-purpose breeds (Belgian Blue and Piemontese) have fixed the mutant alleles at the population level, whereas a synthetic double-muscled strain was created in France (INRA 95). Other breeds of cattle have double-muscled strains (Dunner et al. 2003; Wiener et al. 2009). The Belgian-Blue breed consists of an 11 bp deletion of nucleotides, nt821 (del11), while the Piemontese causes a substitution of amino acids (C313Y mutation).

The different mutations and different selective histories of the two main double-muscled breeds have differentiated them in several ways. The Belgian Blue selection objectives were concentrated on muscularity and growth rate (Coopman et al. 2007), and impressive results were achieved for these traits. The drawback was a worsening of maternal calving ability, so that almost all cows of this breed now calve only by caesarean section (Kolkman et al. 2007). On the contrary, the selection index of the Piemontese breed includes maternal and direct calving ability with a weight of 60%, with the result that caesarean sections in this breed are comparatively rare (5% for heifers and 3% for cows). These different selection histories have modified the sexual dimorphism of these breeds such that it is much less evident in the Belgian Blue than in the Piemontese, but their MQ traits were unaffected (Bittante et al. 2018). An important difference between the two major double-muscled breeds is that the increase in the muscle cross-sectional area in the Belgian Blue arises mainly from an increase in the sectional area of each muscle fibre (hypertrophy), whereas in the Piemontese it arises more from the muscle fibres increasing in number (hyperplasia) than in thickness (Verdiglione and Bittante 2013). However, the quality traits of the beef of these breeds are similar (Kobolák and Gócza 2002; Fiems 2012; Bittante et al. 2018).

Double-muscled breeds play an important role as terminal paternal breeds in crossbreeding, especially in areas where consumer demand is for muscular lean beef, as in continental Europe. New opportunities have opened with the widespread use of sexed semen from dairy bulls to produce replacement heifers on dairy farms and the consequent large increase in the availability of dairy cows for terminal crossbreeding with beef bulls (Foraker et al. 2022; Lauber et al. 2023). The use of beef semen increases the fertility of dairy cows (favourable heterosis of crossbred embryos), which compensates for the lower fertility of cows inseminated with purebred sexed semen (Bittante et al. 2020a, 2020b). Terminal crossbreeding of dairy cows with double-muscled bulls yields crossbred calves that have a much higher value than purebred dairy calves (Bittante et al. 2020a, 2020b), and similar fattening performances, carcase traits and also MQ traits to the calves of purebred conventional beef breeds and crossbreds from beef sires and suckler cows (Bittante, et al. 2021; Bittante et al. 2023). A future scenario for combined dairy and beef production could be the use of rotational crossbreeding of dairy breeds, whereby crossbred heifers and a few cows are inseminated with the sexed semen of purebred dairy bulls (three breeds), while all other cows are inseminated with conventional beef semen. This combined dairy-beef system also has the advantage of minimising the environmental burden per unit of product (milk + beef) obtained (de Vries et al. 2015).

CAPN1 and CAST, calpain and calpastatin, genes and beef tenderness

Further examples of the major genes that affect beef quality are those that regulate the calpain (CAPN) system (three different calpains and their specific endogenous inhibitor, CAST). Calpains are intracellular calcium-dependent cysteine proteases found in muscles and are involved in post-mortem proteolysis of myofibrillar proteins and cytoskeletal anchorage complexes, and also in the tenderisation process during meat ageing (Bhat et al. 2018).

The μ-CAPN1 gene is known to encode μ-CAPN1, while the CAST gene encodes the inhibitor of that protease (Koohmaraie 1996). Markers associated with significant effects on meat tenderness have been identified for the CAPN1 gene on bovine chromosome 29, and for the CAST gene on chromosome 7 (reviewed by Lambe et al. 2015). Commercial DNA tests to identify CAPN1 and CAST genotypes in cattle have been validated and found to be associated with a strong effect on meat tenderness (Van Eenennaam et al. 2007). Note that the effects and associations with genetic markers vary in different breeds. The strong association between the CAST and CAPN1 genes and the tenderness of beef meat has been reviewed by Węglarz et al. (2020). Aside from tenderness, these two genes have also been associated with other MQ traits: marbling (Cheong et al. 2008; Shi et al. 2011; Li et al. 2013), cooking losses (Juszczuk-Kubiak et al. 2004), meat colour (Juszczuk-Kubiak et al. 2004; Li et al. 2013) and meat flavour (Casas et al. 2006; Reardon et al. 2010).

Diacylglycerol O-acyltransferase 1 gene and beef marbling

Diacylglycerol O-acyltransferase 1 is a microsomal enzyme involved in the metabolism of lipids and, in particular it catalyses the final step in the synthesis of triacylglycerol (Lambe et al. 2015). This enzyme is encoded by the DGAT1 gene on chromosome 14 in cattle and is known for having a strong effect on milk production, but also meat production (Khan et al. 2021). In dairy cows, the DGAT1 gene is strongly related to milk fat production in the udder, on the one hand, increasing the fat content and fat daily yield, but also the protein content of milk, and, on the other hand, reducing the daily production of milk, protein and lactose (Bovenhuis et al. 2015).

Regarding beef quality, several studies found associations between DGAT1 gene polymorphisms and back-fat thickness (Avilés et al. 2013), intramuscular fat (IMF) content (Thaller et al. 2003) and marbling score (Li et al. 2013; Yuan et al. 2013). Interest in the DGAT1 gene as a way of modifying the fatness of carcases and meat depends on the local market, particularly whether or not consumers favour this trait, and, obviously, by the frequency of DGAT1 alleles in the cattle population of interest.

Genomic selection: genome-wide association studies, genomic prediction and other ‘omic’ sciences

Since the discovery of the double helix structure of DNA (Watson and Crick 1953), ‘omic’ sciences have made great strides in furthering our knowledge of the functions and reproduction of living beings, i.e. life on the planet. Our task here is not to illustrate these advances but simply to offer a reminder that knowledge of the genetic mechanisms of beef quality also relies on these sciences and their continually evolving research methods.

A simple bibliographic search of the Web of Science database for the period 2001–2022 using the keywords ‘Genom *’, ‘Beef’ and ‘Quality’ retrieved 753 items, 39 of which are classified as review articles (Hocquette et al. 2012; Leal-Gutiérrez and Mateescu 2019; Juárez et al. 2021). Since the first few publications at the beginning of this millennium, the number of studies has quadratically increased year on year, approaching a rate of a hundred articles per year (Figure 9).

Figure 9. Number of new publications per year related to genomic research on beef quality.

Animal genotyping by SNP has been very successful in recent years. Many of the articles involved genome-wide association studies (GWAS) that aimed to identify areas of the cattle genome and individual genes controlling the most important beef quality traits. Several articles concerned GS, i.e. the use of genomic calibration equations for indirectly predicting the breeding value of animals for beef quality traits by combining the genotyping and phenotyping datasets (see previous Figure 6).

GS has the potential to be a very powerful tool for improving the efficiency of MQ selection (genetic progress by year) because it could be used for: i) improving the accuracy of estimates of breeding value, ii) raising the selection intensity by increasing the number of sires (and dams) evaluated and iii) reducing the generation interval, allowing very young sire candidates to be evaluated and used without the need to wait for progeny testing of sires. It is well known that GS is particularly useful for improving those traits, such as the majority of beef quality traits, that have a modest heritability, are not measurable early in life, and, in particular, are not measurable on living animals, and hence on breeding stock.

On the other hand, it is worth remembering the great danger that incorrect use of this technology poses to the genetic variability of the population by increasing relatedness among the animals and inbreeding coefficients of individuals (see previous Figure 4).

Cloning of bovines: agamic multiplication of identical animals

Replicating an excellent animal to produce an unlimited number of copies is an attractive idea to maximise performance, quality and standardisation in the animal industry. Agamic multiplication is an alternative to sexual reproduction to obtain new, genetically identical, living beings. Since the birth of Dolly in 1996 (Campbell et al. 1996), cloning individuals from adult somatic cells became one of the possible techniques for multiplying mammals, particularly farm animals. Since then, several bovines have been successfully cloned. Aside from dairy breeds (Holstein and Jersey), some clones of beef breeds (Brahman, Angus, Hereford, Spanish Fighting Bull, Boran, etc.) have also been obtained.

Regarding the consumption of food obtained from cloned animals, the US Food and Drug Administration has concluded that food from cattle, swine, and goat clones is as safe to eat as food from any other cattle, swine or goat (FDA 2021).

Clones have the same alleles as the donor animal, but are different from identical twins because they are obtained by transferring the nucleus (DNA) of a somatic cell (not the embryo) taken from a donor animal into an oocyte taken from a female animal and from which the nucleus (which contains its genes) has been removed. From this, the embryo forms and is implanted in a different surrogate mother. Clones are often born in different periods and different herds. Clones may not be genetically identical because different somatic cells of the same individual can undergo different DNA mutations and DNA methylation, and the contribution of the maternal mRNA from the oocyte also differs. Lastly, the telomeres of cloned animals are often shorter than those of their DNA donor and of conventionally obtained animals, meaning an accelerated genetic ageing of animals. This, the different DNA methylation pattern, and large offspring syndrome due to epigenetic changes contribute to the low rate of success, increased pre- and post-birth mortality, and the frequently attested shortened lifespan of cloned animals (Keefer 2015).

The main obstacles preventing the expansion are: the very low rate of success (and very high related costs), ethical concerns about the principle of artificially ‘multiplying’ individuals (Häyry 2018), and the poorer welfare (evilfare) of clones and their DNA donors, oocyte donors and surrogate mothers (Sinha et al. 2019). It seems clear that cloning animals is not a sustainable method for producing food from animals in the future, although cell cloning may well be.

Cloning of muscle tissues: ‘cultured meat’ or ‘artificial meat’, end of the game (?)

‘Cultured meat’ is meat obtained not by slaughtering animals but by culture of animal cells in vitro. It is also known as: artificial meat, healthy meat, slaughter-free meat, in vitro meat, vat-grown meat, lab-grown meat, cell-based meat, clean meat, cultivated meat, and synthetic meat. This is not a recent idea, for in 1931 Winston Churchill wrote: ‘We shall escape the absurdity of growing a whole chicken to eat the breast or wing, by growing these parts separately under a suitable medium’ (International Churchill Society 2018).

This very ‘hot topic’ is widely discussed in the 69th Congress ICoMST2023 (Thorrez 2023; Xiong, 2023). Regarding biodiversity and genetics, cultured meat is cloning of muscle fibres through agamic multiplication of an individual animal cell, so all the cells produced are genetically identical. So, the only possible genetic variability is that arising from using different ‘ancestor’ cells taken from different animals.

It is almost two decades since cultured meat was first suggested as a substitute for meat obtained from animals (Edelman et al. 2005). It attracts huge research and industrial investments, and many start-up companies in this field have been created around the world. The most common reasons in favour of cultured meat as a substitute for ‘natural’ animal meat are: i) environmental advantages, related to the high impact of meat production; and ii) ethical concerns, due to slaughtering animals. The second reason is clearly important, but not essential, as demonstrated by vegetarian and vegan food habits based on conventional foods but excluding meat. The first reason must be demonstrated. No credible ecological footprint is available for cultured meat because it has not yet reached the stage of being a ‘common’ food produced on an industrial scale.

The origins of humans are linked to those of all other living beings through the evolution of species, and humans have always existed in different ecosystems, from the most natural to our modern ‘Anthrome’ (contraction of anthropogenic biome, from the ancient Greek ἄνθρωπος -γενής ‘human generated’ and βίος–ωμα ‘living mass’). All living beings (species biodiversity) in every ecosystem depend on each other through food chains (autotrophs, herbivores, carnivores of different orders, etc.). Energy from the Sun is captured through photosynthesis and converted to chemical energy to fuel the life of all living beings through cellular respiration, maintaining biodiversity. The cycle of matter through every ecosystem in equilibrium could be considered net zero because the quantities of CO₂ and H₂O captured by autotrophs to produce organic substances during photosynthesis are equal to those produced by all living beings in ecosystems when they consume organic substances (feed and food) with cellular respiration. At the same time, the amount of O₂ released by autotrophs during photosynthesis is equal to the quantity consumed by all living beings to oxidise organic substances during cellular respiration.

We know that alterations to anthropic ecosystems are caused mainly by inputs that are external to the natural cycle of matter and, particularly, by the use of fossil fuels and nitrogen sources.

Like cultured meats, other substances essential for human nutrition (carbohydrates, lipids, vitamins, etc.) can also be artificially produced in labs and factories. The gradual substitution of ‘natural’ foods from agriculture with synthetic/cultured foods from factories means removing humans from ecosystems and from functional relationships with other living beings and creating an abiotic (from ancient Greek ἀ-βιωτικός, non-living) system in which humans depend on factories and power plants (and, of course, houses, buildings, roads, infrastructures, etc.) consuming large amounts of energy and non-renewable resources. This system is not an ecosystem, being based on only one living species, the humans, and is characterised by a null biodiversity among species. This also means returning to linear economics (whose motto is ‘take, make, waste’), and abandoning the circular economy (European Parliament 2015), mimicking the cyclical flow of typical ecosystems matters (recycling) allowed by the many living species involved.

Is this the future of mankind?

It is possible.

But personally, I hope for and work towards maintaining and improving our natural, anthropic ecosystems!

Conclusions

The biodiversity of beef cattle is endangered by the processes of market globalisation and intensification of farming systems, creating the risk that the vast majority of cattle breeds around the world will go extinct. Furthermore, genetic and genomic selections lead to a reduction in genetic diversity within breeds and an increase in the frequency of inbred animals. The major obstacles to the genetic improvement of beef quality traits are i) the fact that they are not directly measurable on live animals (sire and dam candidates), and ii) the complexity and high cost of meat sampling and analysis. Indirect phenotyping of beef quality traits through rapid, non-destructive, low-cost secondary prediction methods (NIRS, etc.) can contribute to solve the second problem, while genomic predictions can solve the first. The heritability (and efficiency of selection) of directly measured beef quality traits is highly variable according to breed, trait and condition. The same is true for indirect secondary predictions. The fixation of favourable alleles of some major genes (myostatin for double-muscling, CAPN-CAST for tenderness, DGAT-1 for marbling, etc.) represents an easy and efficient tool for modifying MQ, but comes with some unfavourable side effects that need to be taken into account. ‘Omic’ technologies are powerful instruments for understanding the genetic/biological basis of MQ, but their use is complex and needs to be carefully evaluated. Cloning animals seems not to be of practical value, whereas the cloning of cells (cultured meat) is much more promising, but has currently high costs and yet unknown environmental burdens and is, in particular, highly controversial from an ethical and a biodiversity point of view.

Ethical approval

Animal Care and Use Committee approval was not obtained for this study because the data were obtained from published scientific articles.

Acknowledgements

The author wishes to acknowledge his colleagues prof. Alessio Cecchinato, Luigi Gallo and Stefano Schiavon (Department of Agronomy, Food, Animals, Natural resources and Environment–University of Padova, Italy), and prof. Antonella Dalle Zotte (Department of Animal Medicine, Production and Health–University of Padova, Italy), for their support and advice during article preparation and their review of the manuscript.

Disclosure statement

No potential conflict of interest was reported by the author.

Data availability statement

The data presented in this article are freely available being obtained from cited published articles.