Characterisation of white facial markings in Pura Raza Española horses (a worldwide population genetic study)

Abstract

White markings are characteristic of many equine breeds, being quite common in the Pura Raza Español horses (PRE). These white markings are the result of a lack of melanocytes in the skin and hair. In certain horse breeds, such as the PRE, the presence and extension of white facial markings is penalised in the breed’s patron and morphological competitions, so it would be interesting to include it in their selection genetics programs to select against the presence of this special feature. The aim of this study was to calculate the prevalence of white facial markings in a representative population sample of PRE and determine its prevalence depending on the coat colour, its genetic parameters and the influence of systematic effects. The white facial markings have been classified into 5 score. A total of 42,080 PRE horses were analysed. Genetic parameters were estimated using a Bayesian procedure with the BLUPF90 software. Systematic effects included in the model were: birth period, sex, birth stud geographical area and inbreeding coefficient. The pedigree information included 93,322 horses. The prevalence was 34.2%. Systematics factors were significant in the presentation of facial white markings. Heritability in real scale ranged from 0.53 for black to 0.32 for the chestnut coat colour population, both in the linear heterogeneity model. White facial markings were more prevalent in inbred chestnut males of Spain. The additive genetic base shows that the prevalence could be managed by genetic selection.

HIGHLIGHTS

• PRE with white marks produce financial losses and those that invade a large portion of the eye orbit cannot registered as animals of the breed

• The prevalence was 34.2%, being the large white markings more affected by systematic risk effects.

• Genetic base of white facial markings in PRE is very related with the coat colour, and presents high-moderate heritability.

• White markings could be reduced by genetic selection using breeding values in all the PRE horses as a selection criterion into its breeding program.

Keywords:

• coat colour
• disqualifying fault, equine
• genetic parameters
• heritability

Introduction

Coat colour is one of the most noteworthy features of an animal and has historically been the object of particular interest to humans (Pruvost et al. 2011). White markings in mammals may result from an altered embryonic development of the neutral crest melanocyte lineage and a lack of mature melanocytes in the unpigmented skin areas (‘leucism’). Although over 40 genetic variants are known to influence the white markings on the horse, many still have unknown genetic causes (Patterson-Rosa et al. 2022) finding genetic variants associated in several genes, including KIT, PAX3 and FMIT (Hauswirth et al. 2012; Hauswirth et al. 2013; Haase et al. 2013; Negro et al. 2017). The lack of white markings in the subspecies Przewalski horse (Equus ferus przewalskii) and in the Palaeolithic cave paintings illustrating the horse, as well as the markings rarely occurring in primaeval Asiatic horses, indicate white marks are characteristic of domestication (Stachurska and Ussing 2012), since domestication has altered the frequency at which certain characteristics, potentially detrimental in the wild, are manifested. The fact that current clones horses differ in their white markings also indicates that there must be an epigenetic component in the distribution of these white markings.

The Pura Raza Español (PRE) studbook accepts animals with a diversity of coat colour patterns. In PRE, the colour coat and peculiarities such as white marks are systematically recorded above the 6 months age. As with other horse breeds (Swiss Franches-Montagnes, Rieder et al. 2008; Polish Konik, Stachurska and Ussing 2012; Hucul, Pasternak et al. 2020; Menorquin horses, Perdomo-González et al. 2022), PRE horses with white marks produce themselves financial losses and those that invade a large portion of the eye orbit (surrounds a large part of the area around the eye) and the white marks on the body have a size ≥ 3 cm in diameter cannot registered as animals of the breed (ANCCE 2023). So, it could be useful to include this type of characters in their selection genetics programs to select against the presence of this special feature. Until now, very few studies have addressed the estimation of the heritability of this type of traits (Górecka et al. 2006; Rieder et al. 2008; Perdomo-González et al. 2022) and in none using the PRE population. For this reason, is crucial to understand better how different factors contribute to the development of white facial markings to improving the breeding selection of PRE. Therefore, the specific aims of this work were first, to calculate the prevalence of white facial markings in a significant population of PRE horses as a function of coat colour and the other environmental effects (year of birth, sex, birth stud geographical area and inbreeding coefficient); second, estimate the heritability, using different genetic approaches, to test its suitability for genetic selection.

Material and methods

Description of the traits and database

The data base included white facial marks records of 42,080 horses (26,704 mares and 15,376 stallions). The average evaluation age of the horses was above the 6-month (to identify white facial markings in grey horses, the data were collected when the horses were foals, i.e. they still have a coat colour so that the white facial markings can be well identified). These records were taken between 2006 and 2021, from four different geographical areas (Spain, rest of Europe, North America and Central-South America) during the systematic veterinarian evaluation that all horses have to identify them before being registered in the PRE studbook (M.J.J. Sánchez-Guerrero et al. 2017). The genetic coat colour data of 18,375 PRE were also available.

To carry out this study, the white facial markings trait was analysed in two ways: a) as a dichotomy trait (no affected, class 0; affected, class 1); b) using a discrete scale, which includes five scores (no affected, class 0; affected, classes 1, 2, 3 and 4) in which the extremes represent the biological limits for this morphological trait (Table 1).

Table 1. Description of the white facial markings scoring system in the pura raza español horse.

Score 0	Score 1	Score 2	Score 3	Score 4
No white facial Markings	Small white markings on the forehead	Big white markings on the forehead	White markings on front-nose	White markings all over the head

Statistical and genetic model

The number of PRE horses affected by white facial markings and their proportion within the entire population, a Generalised Non-Linear Model (GLZ) analysed depending on coat colour (4 levels: grey, bay, black and chestnut). A study to observe the presence of white facial markings in the offspring depending of the coat colour and according to whether the mating between their parents both have markings, both do not have markings, one has markings and the other does not, and vice versa was carried out.

A univariate Generalised Non-Linear Model for an ordinal multinomial distribution (multinomial discrete distribution containing information on a rank scale) was computed to assess the significance of some potential risk factors (without interactions) that could have in white facial marking. For the statistical analyses, the Statistica software v. 11.0 (StatSoft STATISTICA) was used. The potential systematic risk factors studied were:

• Sex (2 levels): male and female.

• Birth period (2 levels): animals born between 2006-2013 and 2014-2021.

• Inbreeding coefficient (4 levels): non-hazard level ≤0.0625; low hazard level >0.0625 ≤ 0.125; hazardous level >0.125 ≤ 0.1875 and very hazardous level >0.1875. The inbreeding coefficient has been considered as a fixed factor, establishing a series of classes or levels, as in other studies in this breed (Ripollés-Lobo et al. 2023; Encina et al. 2023) to study de inbreeding depression hazard level.

• Geographical area (4 levels): Spain (due to its high PRE number), rest of Europe, North America and Central-South America.

The estimation of the genetic parameters was carried out with a univariate animal model, with white facial markings score as discrete traits or the presence/absence of white markings as dichotomy variable (binomial distribution). The heritability estimates of the binomial threshold model analyses were transformed into the observed scale in order to compensate for the underestimation of heritability when applying models to binomials traits (Dempster and Lerner 1950). The approaches included as systematic effects all the influencing factors that were statistically significant (p < .05) in the previous GLZ analysis: gender, the inbreeding coefficient, the coat colour, the birth stud geographical area and the birth period. Additive genetic and residual effects were included as random factors. For the analyses, two genetic models were applied and compared: two models without (for binomial and discrete trait) and one with heterogeneous residual variance according to the coat colour (for discrete trait). All the genetic models were analysed using a Bayesian approach via Gibbs sampling with the GIBBSF90+ module of the BLUPF90 software (Misztal et al. 2020). The Gibbs sampler was run for 250,000 rounds, with the first 10,000 considered as burn-in; then, every 10th sample was saved for later analysis. Posterior means and standard deviations were calculated with the POSTGIBBSF90 software (version 3.15) (Misztal et al. 2020) to obtain estimates of variance components. The equation in matrix notation to solve the mixed model was: $$y=\text{X}b+\mathrm{Z}u+e,$$ with $$\mathrm{u}\sim \mathrm{N}\left(0,\text{A}{\sigma }_{\mathrm{u}}{}^2\right),\text{e}\sim \mathrm{N}\left(0,\text{I}{\sigma }_{\mathrm{e}}{}^2\right)$$ where y is the vector of observations, X is the incidence matrix of systematic effects, Z is the incidence matrix of animal genetic effects, b is the vector of systematic effects, u is the vector of direct animal genetic effects, e is the vector of residuals, σ_u² is the direct genetic variance, σ_e² is the residual variance, I is an identity matrix, and A is the numerator relationship matrix.

Convergence was tested using the Z criterion of Geweke (Sorensen and Gianola 2002) and Monte Carlo sampling error, was computed using time-series procedures, as described in Geyer (1992). In order to identify the precision of the parameters, the 95% highest posterior density (HPD) intervals were determined from their marginal posterior distributions. Deviance Information Criterion (DIC) have been used in this work to choose the best model, which assessed the models’ goodness of fit. ENDOG software (Gutiérrez and Goyache 2005) was used to estimate the inbreeding coefficient of animals. The pedigree file information necessary for genetic evaluation, collected from the official PRE studbook, all available generations were included (93,332 horses; ranging the known equivalent complete generations between 7.60 and 12.65).

Results

A total of 42,080 PRE horses were evaluated. The most common class was had not white facial markings score 0 (65.8%). Within those affected by white facial markings (14,383 horses): the most common class was score 2 (60.0%) and the less common class was score 4 (3.3%) (Table 2). Coat colour was a significant risk factor for white facial marks. Consequently, there were differences in the proportions of horses with and without white facial markings in each of the coat colours analysed. A high difference in the expression of white markings was found between the chestnut and non-chestnut phenotypes. Besides, horses with black coats had the lowest percentage of horses affected by white facial markings. So, the biggest difference was between horses with black coats (70.6% no affected) and horses with chestnut coats (42.7% no affected).

Table 2. Number and proportion (%) of Pura Raza español horses according to the white facial markings scores and Generalised Non-Linear Model analysed depending on coat colour.

		Affected				p value
	Score 0 (No affected)	Score 1	Score 2	Score 3	Score 4
Total Population (n)	27,697	2,645	8,629	2,634	475	<.001
(%)	(65.8)	(6.3)	(20.5)	(6.3)	(1.1)
Bay (n)	11,091	1,322	3,762	601	175
(%)	(65.4)	(7.8)	(22.2)	(3.6)	(1.0)
Chestnut (n)	1,482	204	1,114	583	89
(%)	(42.7)	(5.9)	(32.1)	(16.7)	(2. 6)
Black (n)	3,851	431	1,022	115	33
(%)	(70.6)	(7.9)	(18.8)	(2.1)	(0.6)
Grey (n)	11,273	688	2,731	1335	178
(%)	(69. 6)	(4.3)	(16.9)	(8.1)	(1.1)

The number of offspring, within our studied population, affected by white facial markings, depending on the coat colour, when one or both parents have white facial markings was present in Table 3. A total of 4,484 foals with both parents evaluated have been studied, of which 1,505 of them have white facial markings. In the case of both parents with markings, the offspring shows that 61.8% have markings and 38.2% have no white facial markings. In the case of both parents without markings, 78.2% of the offspring have no markings and 21.8% have white facial markings. Depending of the coat colour the highest prevalence of white facial markings always happened in chestnut foals. Differences according to the gender of the parents seem not to occur.

Table 3. Results of matings of Pura Raza española horses regarding the presence of white facial markings according to coat colour.

Parents	Total population	Offspring
		Number of foals	% With markings	% Without markings
		4,484	33.6%	66.4%
With markings (father) × With markings (mother)	Bay	243	55.9%	44.1%
	Chestnut	242	72.3%	27.7%
	Black	30	50.0%	50.0%
	Grey	70	51.4%	48.6%
With markings (father) × Without markings (mother)	Bay	464	31.9%	68.1%
	Chestnut	257	50.6%	49.4%
	Black	91	30.8%	69.2%
	Grey	162	28.4%	71.6%
Without markings (father) × With markings (mother)	Bay	469	33.1%	66.9%
	Chestnut	214	57.0%	43.0%
	Black	103	34.0%	66.0%
	Grey	104	32.7%	67.3%
Without markings (father) × Without markings (mother)	Bay	1,125	19.2%	80.8%
	Chestnut	311	31.5%	68.5%
	Black	265	20.7%	79.3%
	Grey	334	22.77%	77.3%

The non coat colour systematic risk factors associated with white facial markings score are shown in Table 4. In the whole population, all the risk factors studied (coat colour, sex, geographical area, inbreeding coefficient and birth period) were significant risk factors, with significance coefficients below 0.05. Males had a higher percentage affected than females in the whole population (35.0% vs. 33.7%). There was also an association with the white facial markings score, and the geographical area in which the horse was birth. In general, the percentage of affected horses was higher in studs located in countries of Europe area and Spain (28.9% and 36.5%, respectively). There was also an association with the inbreeding coefficient that horses present. In horses with inbreeding exceeding 18.7%, the percentage of individuals with facial white markings was higher (39.2% affected). Besides, animal with more than 12.5% of inbreeding had the mayor prevalence into the score 4 (1.7% animals with more than 12.5% and 1.8% for those with more than 0.1875). Horses born in the last birth period tend to be more affected (35.8%) (Table 4).

Table 4. Generalised non-linear model (GLZ) between white facial markings and systematic risk factors and significance level.

Risk factors	Class	Population (%)						p-value
		n	Score 0	Score 1	Score 2	Score 3	Score 4
Sex	Females	26,704	66.3	6.3	20.3	6.0	1.1	<.05
	Males	15,376	65.0	6.3	20.9	6.7	1.1
Geographical area	Spain	32,470	63.5	6.3	21.6	7.3	1.3	<.05
	Europe	2,085	71.1	8.3	16.1	3.9	0.6
	North America	2,956	76.3	6.4	13.8	3.3	0.2
	Central South America	4,569	73.1	5.5	18.7	1.7	1.0
Inbreeding coefficient	≤0.625	24,355	65.1	6.8	21.1	5.9	1.1	<.05
	>0.0625 ≤ 0.125	13,160	67.6	5.6	19.5	6.3	1.0
	>0.125 ≤ 0.1875	3,109	66.3	5.1	18.7	8.2	1.7
	>0.1875	1,456	60.8	5.7	23.8	7.9	1.8
Birth period	2006–2013	23,649	67.1	6.3	19.3	6.2	1.1	<.05
	2014–2021	18,431	64.2	6.3	22.0	6.3	1.2

The genetic parameters of the four models carried out in this study are presented in Table 5. Heritabilities, in real scale, of white facial marking ranging from 0.53 (black coat colour) to 0.32 (chestnut population), both in the lineal heterogeneity model). The additive variance components were higher in dichotomic model than linear ones. The Monte Carlo standard errors were small (all of them lower or equal than 0.001). Lack of convergence was not detected by the Geweke test, ranging between −1.8 (dichotomic) and −0.9 (for chestnut in linear with heterogeneity variance). DIC values for all the models are shown in Table 4. Under this criterion, the models with a lower DIC value were understood to fit better. Dichotomic model, for instance, which only included the presence/absence of white facial marks, had the best fit. Heterogeneity variance model had slightly better-fitting than homogeneity variance model.

Table 5. Genetics parameters (mean and median for additive genetic variance (σ_u) and residual variance (σ_e) and heritabilities (h²) and fitting parameter deviance information criterion (DIC) for white facial markings according to coat colour, calculated with different genetic models.

Model		σu		σe		h^2	HPD 95%	DIC
		Mean	Median	Mean	Median
Dichotomic		1.7	1.7	1.0	1.0	0.63^1 0.38^2	0.60–0.66 0.36–0.41	−32298.34
Linear		0.4	0.5	0.6	0.6	0.43	0.41–0.46	113184.05
Linear with heterogeneity	Chestnut	0.5	0.5	1.0	1.0	0.32	0.30–0.34	112344.19
	Bay			0.5	0.5	0.47	0.45–0.49
	Black			0.4	0.4	0.53	0.51–0.56
	Gray			0.7	0.7	0.40	0.38–0.42

¹Heredability underlying scale 2 heredability real scale.

HPD 95%: 95% higher posterior density. All Monte Carlo sampling error were lower or equal than 0.0001; Geweke Z-Score ranged between −1.8 and −0.9 and p > 1.

Discussion

White facial markings in horses have already been described by other authors as the result of a lack of melanocytes in the skin and coat (Woolf 1990; Stachurska and Ussing 2012), being a very characteristic feature of most breeds of domestic horses. In despite of this, this study has addressed for the first time the prevalence of white facial markings in a worldwide population (as PRE) to determine its prevalence, depending on the coat colour and the other systematic risk effects (birth period, sex, birth stud geographical area and inbreeding coefficient). White facial marks prevalence in PRE were the lowest found (Table 2). There were 85.0% of Quarter horses with spots on the head (Maciel et al. 2020). It has been also described Huculs horses with white markings, claiming that in the 1920_s and 30_s, this trait was most common in grey (93%) and chestnut horses (73%), bay (42%) and blue dun horses (67%), and least common in dun (33%), black horses (26%) (Hollander 1938). Menorcan horses (always black coated horses) also have a high percentage of white markings on the head (47.6%) (Negro et al. 2017). In PRE, as also shown in Arabians and Swiss Franches-Montagnes Horse (Woolf 1991; Rieder et al. 2008), a significant difference in the expression of white markings was found between the chestnut and non-chestnut phenotypes. In Swiss Franches-Montagnes horses (Rieder et al. 2008), chestnut-coated horses exhibited significantly larger white markings compared to bay and black-coated horses. These basic colour phenotypes result from different alleles for the Extension (E/e) and Agouti (A/a) genes, and the differences in white marking expression can be attributed to a pleiotropic effect of the alleles of the Extension and Agouti genes (Woolf 1992; Rieder et al. 2008). It is interesting that when the frequencies of the MC1R gene (melanocortin-1 receptor gene), chestnut (ee), and non-chestnut (E-) were included; the results of the marking analysis were influenced, showing a significant relationship with markings in horses (Rieder et al. 2008). It could be assumed the presence of the recessive E allele at the MC1R locus and the dominant A allele at the ASIP locus act as ‘major genes’ and lead to a greater extent of markings (Stachurska and Ussing 2012). So, a foal with a specific combination of coat colour genes from both parents had a biggest possibility to develop large white markings. Besides this clear association, the behaviour of the small spot assigned to score 1 was difficult to predict (according to our results the highest prevalence of this small white marks were found in bay and black horses, but all the prevalence were very similar) and does not follow the pattern of the rest of the class (being chestnut the horses with highest percentage in class 2, 3 and 4). Neither the most inbreeding horses had higher prevalence of them. So, a hypothesis could be that the small white markings (class 1) appearing to behave in a stochastic manner independent of coat colour and inbreeding. Maybe, these small marks are the most influencing by epigenetic factors while systematic risk factor associated with colour coat, sex or inbreeding are relevant in biggest white marks (class 2, 3 and 4). So, it could be concluded that the extent of these markings must be influenced by genetic, systematic risk factors and random events occurring during intrauterine development, affecting the survival, migration, and clonal proliferation of melanoblasts (Stachurska and Ussing 2012). Taking into consideration that a foal requirement to be accepted as a specimen of a particular breed in the register horse book is to reduce the extent of white markings, it is necessary to understand better the genetic allelic combination and systematic risk effects that are involved (assuming that there are stochastic and epigenetic effects that we cannot control in a standard animal breeding program). Probably, it could be due to the effect of the polygenic inheritance of white marks that the ultimate extension of markings was influenced by genes, as well as by intrauterine factors (Stachurska and Ussing 2012), being probably the score 1 mainly affected by this last according to our results. To confirm that hypothesis would be necessary more genomics studies. Woolf in one of his articles on the study of white markings in Arabian horses (Woolf 1993) explained the inheritance of white markings and that foals inherit from their parents a series of genes with equal or unequal effects on the presence of white markings (Stachurska and Ussing 2012). So, highly marked parents could carry more genes that produce markings and are homozygous at some loci, leading to offspring showing a tendency towards having markings (Woolf 1990). In contrast, less marked or unmarked parents could have fewer genes producing white markings, resulting in offspring showing a tendency to not have markings. Our results (Table 3) were similar to the findings of the study conducted with Polish horses (Stachurska and Ussing 2012), but this general tendency is hardly affected for the coat colour of the parents being always higher prevalence in chestnut horses.

In order to attain a precise and comprehensive comprehension of the patterns of white facial marks, it is imperative to take into account factors may exert an influence on the incidence and intensity of these marks. Although most widely published works have addressed the genetic perspective, beyond the coat colour, no detailed other systematic factor analyses for white facial marks in horses have been extensively conducted until now. In our study, we examined not only the influence of the coat colour but also the sex, the individual rank of inbreeding, the birth stud geographical area and the birth period (Table 4). Maybe the greatest proportion on white marks in the last birth period and in Spain could be due to the highest sport use of PRE, mainly in dressage. In this discipline, the presence of white markings within the permitted extent may not impact the commercial value of these horses, as their value is typically not significantly influenced by morphological scores. Additionally, within the population analysed, there has been a rise in the prevalence of chestnut coat colour (from 0.063% to 0.097%), whereas black and grey horses experienced a decrease in percentage. The percentage of inbreeding does not always exhibit a linear relationship with the trait under study. Sometimes, it reaches certain thresholds (breakpoint) where inbreeding depression becomes evident (Laseca et al. 2024). The presence of large white markings (class 3 and 4), have a higher probability to occur due to the level of inbreeding. Very small spots (class 1) do not seem to be related to the level of inbreeding, and class 2 of white facial markings only appear to be related to a dangerous level of inbreeding (>0.1875). So, in broad terms, one could assert that the highest inbreeding level does result in the greatest proportion of white markings (except for score 1), which could be due to a highest homozygosity for specific genes related to the extension of white markings. Overall, all these systematic effects were statistically significant for the white facial marks analysed. The effects mentioned in this study have previously been implemented in estimations of genetic parameters for morphological abnormalities in the PRE breed, such as cresty neck (Sánchez-Guerrero et al. 2017), ewe neck (Ripolles et al. 2020) or limb deformities (Ripollés-Lobo et al. 2023) as well as those associated with medical conditions, like obesity (Sánchez-Guerrero, Ramos, et al. 2019) or the presence of cutaneous melanomas (Sánchez-Guerrero, Solé, et al. 2019). Additionally, they have also been included in the assessment of various PRE variables, including morphological (Poyato-Bonilla et al. 2020) variables. In the biggest study of vitiligo in PRE horses (Sánchez-Guerrero, Solé, et al. 2019), it was found that males also had a higher prevalence of this disorder were the grey coat was the most affected followed by chestnut coat colour, and Central - South America stud farms showed a higher percentage of vitiligo affected horses followed by North America.

The estimation of genetic parameters is essential for an efficient development of a breeding program, since they are direct indicators of its success, as they condition the type of selection that can be carried out (Sánchez et al. 2013). White facial marking also shows lower heritabilities to previously white facial marking studies (Woolf 1990; Rieder et al. 2008) in Swiss Franches-Montagnes and Arabian horse; h²= 0.68 and 0.69) but higher than white facial markings for the Pura Raza Menorquin (Perdomo-González et al. 2022; h² = 0.23) where only black coat colour is accepted in their studbook. A heterogeneous variance model was used to allow the residual variance to be estimated based on coat colour. As in a prior study focusing on horse coat colour (Bartolomé et al. 2022), here the residual variances could be considered heterogeneous and can be divided according to the coat colour into four subclasses for all the horses studied. This involves the classification of classes, within which the variance is assumed to be constant, and where the change of the residual variance is continuous over time (Jaffrezic et al. 2000). In our study, the model containing heterogeneous variance was the best convergence, which aligns with previous studies where heterogeneous models tend to give preferable results to classic homogeneity models (Gutiérrez et al. 2006; Bartolomé et al. 2013; Cervantes et al. 2020). Conversely, the heterogeneity model residual variance models assume that genotypes at individual loci differ not only in their effect on the mean but also on the variance of the trait evaluated (SanCristobal-Gaudy et al. 2001; Hill 2002). The heterogeneity model showed that genetic variability in the residual variance of white facial markings variable due to the horse’s coat colour is relevant as previously was obtained in other behaviours and coat featured traits studies (Bartolomé et al. 2022; Encina et al. 2023). Another explanation that could corroborate these differences found could be attributed to the extension gene and to the agouti, depending on whether the animals are homozygous or heterozygous for these genes (Rieder et al. 2008). While the results (Tables 3 and 5) support the model that additive genes (polygenes) influence the presence and extent of common white markings, they also indicate (Table 4) that males have slightly more white markings than females and that chestnut horse have more white markings than bay horses.

Conclusions

Genetic base of white facial markings in PRE is very related with the coat colour, and presents high-moderate heritability, being the prevalence large white markings (score 2, 3 and 4) affected by systematic risk effect (were more prevalent in chestnut, males of Europe, including Spain, and with levels of inbreeding greater than 18.75%). Despite the influence of coat colour on the size of white markings, it is possible to conduct a selection without affecting the proportion of different coat colours in the Pura Raza Español horse.

Ethical approval statement

The approval of the work by an ethics committee was not required.

Declaration of interest statement

No potential conflict of interest was reported by the author(s).

Data availability

The data underlying the results presented in the study are available from https://zenodo.org/records/10623667 Sánchez-Guerrero, M.J. (2024), ‘white markings dataset’, doi:10.5281/zenodo.10623666.

Additional information

Funding

This research has been partially financed by the VI PPIT-US. Resolución definitiva de la Convocatoria 2020 de Contratos de Acceso al Sistema Español de Ciencia, Tecnología e Innovación para el Desarrollo del Programa Propio de I + D + i de la US. (II.5B). Fase II.