Predictive ability of the feeding system in Iberian pig by means of several analytical methods

Se ha contrastado la fiabilidad de varios métodos analíti­ cos para determinar el tipo de alimentación recibida durante el cebo por cerdos Ibéricos a partir de muestras de tejido adi­ poso subcutáneo, procedentes de 38 partidas de cerdos de tres campañas de montanera. Las muestras correspondieron con las categorías de alimentación contempladas en la Nor­ ma de Calidad: 251 fueron de bellota, 164 de recebo, 197 de campo y 137 de cebo. Los métodos contrastados han sido ácidos grasos, NIR, alfa y gamma tocoferol, sensor químico, triacilglicéridos, isótopos estables y neofitadieno. Los distin­ tos laboratorios recibieron las muestras sin información so­ bre el tipo de alimentación y remitieron sus predicciones respecto a las categorías previamente mencionadas. Los resultados obtenidos mostraron un elevado porcentaje de aciertos de los métodos cuando la alimentación se corres­ ponde con las categorías extremas (bellota y cebo), sin em­ bargo existía una dificultad en clasificar correctamente las muestras de cerdos de alimentación mixta con bellota y pien­ so (recebo) y confusión de algunos métodos cuando los cer­ dos son cebados con piensos que incluyen grasas vegetales con altos niveles de ácido oleico. Aunque la consideración de sólo dos categorías (cebo/campo y recebo/bellota) eleva el nivel de acierto por encima del 90% en algunos métodos, la combinación de dos o tres técnicas para muestras proble­ máticas permitiría diferenciar las cuatro categorías con el mismo porcentaje de éxito.


INTRODUCTION
For centuries, the Iberian pig has maximized the natural resources of the Dehesa ecosystem especially by montanera, a feeding system in which the Iberian pigs eat acorns and grass from these Mediterranean holm oak and cork woodlands from November to March.This perfect acclimatization has led the conservation of the Dehesa which is threatened by several determinants arising from industrial development.The products derived from these animals have high quality because this traditional feeding type lends special physicochemical characteristics to them.In fact, the high costs of the extensive production and Dehesa management are economically justified by the high price of their derivatives.
In the last 30 years the increased demand for Iberian pork products has led to a rise in production which is nonparallel with the natural resources available.The Dehesa forest cannot support the density of animals that the market demands and the Iberian pigs are feeding with grains and compound feeds.As a result, the derivatives have different characteristics than those that come from
From a legal point of view, the Quality Standard currently in force (BOE, 2007) established four categories in function of their final fattening phase feeding type: bellota or montanera, recebo, cebo de campo and cebo (Figure 1).The Quality Standard also binds a specific labeling for each category.This new Quality Standard was proposed with the purpose of amending deficiencies; however, it removed the analytical method and did not replace it with any other.Nowadays, an objective analytical method is not applied and the responsibility for discerning among the different feeding systems is left to the farm inspection during the montanera season and the estimations of the acorn production and the increase in animal weight.
In addition to the proposal of fatty acid quantification as an analytical control method, some authors suggested the quantification of any other components or the analyses of physicochemical characteristics to distinguish among the different feeding types, especially effective for the cebo category.
In 2008, a Project funded by the "Instituto Nacional de Investigación Agraria y Alimentaria" (INIA) and the MAGRAMA which contrasted the validity of the different analytical methods developed by the mentioned research groups was organized.The objective of the current study consisted of presenting the prediction results of each analytical method to assign the animals according to the feeding type received during the final fattening stage, using back fat samples of Iberian pigs collected in three different montanera seasons.
pigs that carried out their final fattening according to montanera and the prices of the products are lower.This kind of intensive production was 80% of the total in 2011 [Electronic Record of the Iberian pig, "Ministerio de Agricultura, Alimentación y Medio Ambiente" (MAGRAMA)].
Since Iberian pigs were fed with products different from those of the Dehesa, different characteristics were observed both in meat and fat content.In particular, marbling, iodine index, melting temperature or fatty acid composition of backfat were used to distinguish among the different feeding types (Flores et al., 1988).In the late 80's, it was necessary to establish an analytical method that discriminates among products of animals feeding through a montanera system and those feeding with compound feeds.From the montanera season 1988/89 and following, the determination of palmitic, stearic, oleic and linoleic backfat fatty acids was chosen among the mentioned methods as the reference analytical method in the contracts of sale between breeders and producers (BOE, 1998).
However, backfat fatty acid composition is easily modified by adding particular components to the compound feeds; therefore, the composition of these four fatty acids is very similar to those feeding by montanera system and the power of the analytical method to distinguish among the categories decreases.
Despite this, the first Quality Standard Spanish legislation (BOE, 2001) established backfat fatty acid quantification as a complementary method to the farm inspection in order to label the products depending on the feeding type of the pig batches in their final fattening.Later, a protocol regulating the  bellota pigs.In the third montanera season the number of groups was increased and diversified even though the number of samples analyzed in each group was lower.

Analytical methods
The samples were analyzed with different analytical methods which were carried out by different research groups.They received the samples without any additional information about the feeding type.After, the research groups sent the results and predictions to the project coordinator.The methods employed in the current study and the research groups who participated in the project were: 1. Quantification of methyl esters of palmitic, stearic, oleic and linoleic fatty acids (FA) by gas chromatography (Flores et al., 1988;Osorio et al., 1991;Ordóñez et al., 1996).Instituto Tecnológico Agroalimentario, Junta de Extremadura.2. Near infrared spectroscopy (NIR) (De Pedro et al., 1995;GarcíaOlmo et al., 2009).Escuela Superior Ingenieros Agrónomos, Universidad de Córdoba.3. Quantification of alpha and gamma tocopherol (TOC) by highperformance liquid chromatography (Rey et al., 1998;Rey et al., 2006).Facultad de Veterinarias, Universidad Complutense de Madrid.4. Chemometric analysis of chemical spectral fingerprint of volatile compounds by chemical

Animals
Pigs were classified into different groups according to the four Quality Standard Spanish legislation categories (bellota, recebo, campo and cebo).Bellota and recebo pigs eat acorns during the montanera season in which acorns are mature.This period starts in November and finishes in March.All animals were slaughtered over three consecutive montanera seasons (2008/2009, 2009/2010 and 2010/2011) despite the fact that campo and cebo animals can be fattened throughout the year.Samples of back fat (300 g approx.)were taken close to the rump after slaughter.
Tables 1, 2 and 3 show a brief description of the pigs controlled in the three seasons.Each batch corresponds to a group of pigs with identical management and feeding regime, slaughtered on the same date at the same slaughterhouse.All pigs were Iberian unless specified otherwise in the cited tables.The high fat compound feeds have vegetable fats with high levels of oleic fatty acids.Table 4 shows the total number of analyzed samples according to the Quality Standard categories.
The study of the first montanera season focused on the analysis of the animals fed on acorns from different geographic locations, while the second montanera season was especially designed to consider the recebo animals and the batches of pigs fed with high fat compound feeds in order to obtain fatty acid profiles similar to those found in details the degree of accuracy in three categories (cebo/campo, recebo and bellota) since most of the methods in the first season did not differentiate among these categories.The quantification of neophytadiene is the exception because the base of this method is the use of grass intake as an indicator of extensive production versus cebo pigs (Tejeda et al., 1999).Table 6 also shows the prediction of four categories performed by the other three methods (TOC, SQ and TRIG) The results of the first season show a good general prediction for the bellota batches (B081 and B082).The prediction level of the cebo batch (CE81), in which the pigs did not exploit other resources like grass and did not consume high fat compound feed, achieved 100% in most of the methods (except a 93% with the SQ).
The influence of the high fat or not high fat compound feed was very relevant in the campo pigs.Samples from animals feeding with high fat compound feeds were correctly classified (C08 02).However, when the campo pigs were fed using compound feeds with a high content in oleic fatty acid (CA081) there was no method with more than 50% of success, even four of them classified all the samples as recebo or bellota.Neophytadiene (92% success) correctly classified the samples because this method detected a lower intake of grass than the usual in the categories of recebo and bellota.

Statistical analyses
The results are shown as the percentage of correct predictions of the analytical methods in each batch and/or in each Quality Standard category.The total of successes within the Quality Standard category is calculated from the total number of individuals.

2008/09 and 2009/10 seasons
The prediction results of the analytical methods carried out during the two first seasons are shown in Tables 5 and 6.These results are expressed as the percentage of correct predictions in each controlled pig batch and each feeding type category.in spite of the fact that neophytadiene did not differentiate between campo and recebo categories in some batches.
In the second season, the feeding type of some batches was more complex because different kinds of recebo and batches with high fat compound feeds were added.Table 6 shows a degree of accuracy in the prediction lower than the previous season.The cebo batch CE091 was the only one fed with a not high fat compound feed that had a high degree of accuracy in the prediction.On the other hand, many samples of the bellota batches (B091 and B092) the methods classified most of the samples as bellota.R082 had an increase in weight of 40 kg due to the exclusive acorn intake before slaughter without an extra contribution of compound feeds, and the classification was uneven; several methods classified this batch correctly as bellota but NIR and NEOP classified most of the samples as campo (Table 6).
None of the methods achieved a high percentage of correct prediction for the total of eight batches; with a global percentage of correct prediction ranging between 74% (isotopes) and 82% (neophytadiene), The analytical methods classified a high number of samples of the recebo batch R091, with a weight gain in montanera according to the Quality Standard, as campo or bellota, with a majority of wrong prediction in the bellota category.However, in the R092 batch, a special recebo type with some supply of compound feed every day during the montanera period, the methods cataloged most of the samples in the bellota category.
The global percentages of correct predictions decreased in this season more than 20 points compared to the previous season (ranging between 42% in NIR and 62% in tocopherol or chemical sensor) and the variety of correct prediction among batches for each analytical method was also maintained.

2010/11 season
The main objective of the study in this season was the analyses of the accuracy degree of the analytical methods in the classification of the samples in each category.The diversity of the samples in the previous seasons was a "training" to monitor the accuracy of the predictions of each analytical method.In this season, the number and the diversity of batches were increased despite the number of samples in each batch being lower than the previous seasons.The were improperly classified as recebo by several methods, even the B091 batch with a high weight gain during the montanera period.
The results obtained for the campo and recebo categories confirmed those of the previous season.The campo batches with high fat compound feeds (CA091 and CA093) had a degree of accuracy lower than the CA092 batch which was fed with not high fat compound feeds; despite the number of wrong classification being high in the three batches, most of them were classified as recebo and bellota.The pig origin and the formulation of high fat compound feed of the CA091 batch was similar to the CA081 batch, however the degree of

DISCUSSION
The experimental design of the current study and the results obtained are completely original and they are relevant in several senses: the analytical methods are tested under the same conditions because they all use the same samples; the number of analyzed samples is large and individualized; the four categories of the Quality Standard are well represented; and, finally, there is great diversity within each category with respect to feed management.results are shown in Table 7 according to category of feeding instead of batches and also include the percentage of errors and their distribution over categories when only three categories are considered, grouping campo and cebo into one category.
The excellent grade of accuracy in the bellota batch exceeding 80% and 90% and the problems in conducting a correct for the recebo category (ranging from 7% to 68%) were confirmed.In fact, the methods tended to classify recebo as bellota when the batches had a high weight gain in montanera and as campo for those with a lower net weight gain.
The prediction of the analytical methods for the cebo category was disparate.Some of the analytical methods are susceptible to changes in fatty acid composition; therefore the intake of the high fat compound feeds complicates the prediction.In addition to this, if the animals feed grass in small fences, some methods can set the samples as recebo or even bellota.However, when  8 also contains the percentage of correct classification when the methods consider on the one hand cebo and campo (between 20% and 68%) and on the other hand recebo and bellota (between 45% and 77%).Thus, cebo and campo are closely linked like the other two categories with acorn in their feeding regime.Campo requirements, unlike those of cebo 1), call for making use of the territory (maximum of 15 pigs per ha) and extends the slaughter age up to 12 months.The usual availability of pastures in the extensive system during the seasons of autumn winter (Montanera season) deposits traces in the subcutaneous fat.This is also possible in the spring season.Some methods are able to detect such traces and to distinguish between both categories with 68% success.The batches selected were slaughtered in order to collect the samples throughout the Montanera period in which all the campo batches and also some of cebo batches had the opportunity to consume pastures and even a few acorns.The Iberian pigs slaughtered in periods with no or scarce pastures should be more difficult to distinguish, although there are no results The batches of Iberian pigs did not belong to groups of animals organized according to a classical experimental design.They were selected from pigs fattened in commercial farms, except the batches of the Valdesequera experimental line.The main requirement to be included in the trial was the strict observance of the conditions for the feeding categories defined by the Quality Standard.However, the rule only establishes minimum conditions and tolerates a wide range concerning the amount of acorn consumed or the amount and composition of compound feeds.The design of the selected batches of pigs had intended to include all the diversity.
Previous works showed the competence of the analytical methods to discriminate between Iberian pigs fed in Montanera and with compound feeds.The present study confirms those results because the techniques distinguish the batches with high consumption of acorn from others fed only with commercial fodder or with low acorn ingestion.The acorn intake provides special characteristics that allow for distinguishing between the joint categories recebo/bellota from cebo/campo.As Table 8 shows, taking into account only two categories in the third season, the percentage of success in the classification of some methods is close to Therefore, an analysis with the tocopherol method would be adequate to discriminate samples of subcutaneous fat in two categories (cebo/campo y recebo/bellota) with 98% success.Adding the fatty acid quantification, isotopes and neophytadiene make it possible to distinguish among the four categories with 98% accuracy.
As pointed out previously, all the analyses and the prediction exercises took place with individual samples.The number of erroneous classifications within batches never exceeds 25% of the animals, which means that a unique analysis of the fat samples in each batch (as required by the official protocol: BOE, 2007) would be even more successful.

CONCLUSION
Several analytical methods have demonstrated their capacity to distinguish between feeding regimes based on the consumption of acorn and pastures from regimes based on compound feeds.The animal traceability on farms should be and additional tool for this task.The differentiation results between the categories of recebo and bellota applying only one method, although hopeful, are not definite.A certain classification of animals with a mixed regime of acorn and compound feed demands more research taking the issue into account in the compound feed supply.An analysis of a high number of batches, more variability with respect to the feeding system and the consideration of the batch as an experimental unit could facilitate a solution to the problem.Finally, the application of two or three analytical methods increases the to confirm this hypothesis.Similar characteristics of compound feeds for cebo and campo animals would be an impediment to differentiate between them.
The predictions results are better for recebo and bellota.Analytical methods accurately separate these categories from each other when the consumption of acorn is replaced with compound feeds, although the amount of acorn consumed and the quantity and the provided time of the compound feeds for the recebo category are essential to establish clear limits between them.A high quantity of acorn and the compound feeds containing vegetable fats with high levels of oleic fatty acids result in wrong predictions of recebo as bellota.At the same time, the bellota category from montanera with a low quantity of acorn would be predicted as recebo.
The obtained results point out the fact that the application of only one method to achieve an accurate classification of the samples into four categories is not possible.Nevertheless, a combination of several methods could be more effective.Table 8 shows the predictions made by the rest of the methods on 32% of samples with wrong classifications between cebo and campo with respect to tocopherol; this technique has more accurate global results.Isotopes would be right in 19% of them and neophytadiene in 11%.Therefore, a combination of the three methods reaches a 98% success rate.The same procedure with 23% of wrong classifications of recebo and bellota results in 14% of correct classification with respect to fatty acids and 7% for isotopes.The combination of these three techniques increases the accuracy up to 98% again.

Figure 1
Figure 1 Schematic representation of the feeding categories included in the Quality Standard Spanish legislation (BOE, 2007).

Table 1 Description and number of samples of the batch analyzed in the season of 2008/09 Category Code Location Description N
Very extensive management in mountain geographyWeight gain with acorn very high (> 46 kg) J.M. GARCÍA CASCO, M. MUñOZ AND E. GONZÁLEZ