A total of 14 fresh hams from Duroc × (Landrace × Large White) male pigs were utilized. The pieces came from the experimental animals used in the research of Pérez-Ciria et al. (2022), in which the impact of castration type and different feeds was evaluated on productive performances and carcass traits. All the experimental procedures used in that research followed the ethical committee requirements of the University of Zaragoza (ref. PI29/18). Concretely, the hams were chosen at random from pigs receiving the high-energy diet; seven belonged to SCM and seven belonged to IM. The surgical castration was practiced in the SCM group the first week of life and the immunization against GnRF in the IM group was carried out with Improvac® (Zoetis Belgium S.A., Louvain-la-Neuve, Belgium) using three doses, with approx. 25, 58 and 79 kg of body weight (56, 101 and 122 ± 3 d of age, respectively). The general management and feeding at the farm were the same for all of them. All pigs were slaughtered in the abattoir (Teruel, Spain) at the same day, with 142 ± 11.8 kg of body weight (199 ± 3 d of age). There, the left ham from each carcass was taken, trimmed and individually weighed (Table 1).

Upon arrival at the ham-curing facilities, hams were classified according to the weight. Then, the residual blood was removed by a bleeding-massaging machine that presses the femoral artery. The six phases of the dry-curing process were the following:

1. Salting; each ham was introduced in a salting tumbler and 2.5 g of nitrifying salt (a mixture of sodium chloride, maltodextrin, sodium ascorbate and potassium nitrate) per kg of meat mass were applied. Then hams were placed in stackable bins, coated with common salt and kept at 0-2°C and 75-90% of relative humidity (RH) for 0.8 days per kg of meat mass.
2. Washing with water and molded.
3. Resting; hams were hung in racks with hangers and stored from 3.5 to 5°C and from 80-82 to 72-77% of RH for 90 days.
4. Drying; the temperature was gradually increased from 8 to 21°C and the RH reduced from 70-75 to 68-73% for 136 days. Finally, lard was applied manually to the muscular part of the hams to prevent the entry of microorganisms and to avoid over-drying.
5. Maturing; the temperature continued increasing from 25 to 28°C and the RH was maintained at 70-75% for 79 days.
6. Aging; hams stayed in a natural dryer until reaching 32°C for 256 days.
7. The individual weight of all pieces was recorded after salting, resting, drying, maturing and aging.

Once the dry-curing process ended (19 months later), hams were manually boned, sectioned in three parts and individually vacuum packaged. The proximal part of each ham (the opposite part to the hoof) was chosen to carry out the laboratorial analyses and was stored at 4°C until then. One month later, one slice of the sectioned surface of each piece was removed with a slicer (Sammic S.L., Azkoitia, Gipuzkoa, Spain) to determine the color measurements in the piece and another slice was cut to carry out the image analyses. After muscle dissection, the biceps femoris muscle (170 ± 20 g) was destined to measure texture. This muscle was minced with a chopper (Moulinette chopper dpa1, Moulinex®, Groupe SEB Iberica S.A., Barcelona, Spain) to analyze the chemical composition, fatty acid (FA) profile of intramuscular fat (IMF) and volatile compounds. Finally, samples of subcutaneous fat (approximately 100 ± 20 g) were taken to determinate boar taint compounds.

Color was evaluated on subcutaneous fat and on the muscles quadriceps femoris and biceps femoris using a spectrophotometer (CM-2600d, Konica Minolta Holdings, Inc., Osaka, Japan), previously calibrated, with illuminant D65 and observer angle of 10º, in CIELAB color space (CIE, 1986). The mean of three random readings of each section was used to obtain lightness (L*), redness (a*), yellowness (b*), chroma and hue angle (h_ab).

One photograph of each slice was taken following Ripoll et al. (2019a). All images were transferred to a computer and no image editing was applied other than the cropping of the image. Subcutaneous fat thickness was measured at three points (Fig. 1): at the midpoint of the quadriceps femoris muscle, between the quadriceps femoris muscle and the biceps femoris muscle, and at the right side of the biceps femoris muscle. Marbling was estimated in the biceps femoris muscle following the methodology described by Mendizabal et al. (2005). The program ImageJ v1.48 (National Institutes of Health, USA) was used to determine subcutaneous fat thickness and marbling.

The measure of maximum stress was performed following Honikel (1998)held in Helsinki in 1992, a group of scientists with many years of experience in the field of meat quality assessment convened in February 1993 for the first time, and subsequently in 1994 and 1995, in Kulmbach at the German Federal Centre for Meat Research under the auspices of the OECD research project Management of Biological Resources. Three specific areas were discussed in order to develop internationally accepted reference methods: water-holding capacity tenderness colour of meat. In the autumn of 1997 the methods were brought into their final form at the Meat Industry Research Institute of New Zealand (MIRINZ. Each sample was cut in prism-shaped pieces with a 100 mm² (10 × 10 mm) cross-section with the fiber direction parallel to a long dimension of at least 30 mm. A total of 8-10 prisms per sample were sheared perpendicular to the fiber orientation using a Warner-Bratzler device, with a cross-head speed of 2.5 mm/s, attached to an Instron universal testing machine (Model 5543, Instron Ltd, Buckinghamshire, UK) attached to a computer. Maximum stress was the load at maximum peak shear force per unit of cross-section (Ripoll et al., 2019b).

Moisture, ash, protein and IMF were analyzed following the procedures described in BOE (1979). Moisture was determined using an oven (Memmert UFE500, Schwabach, Germany) at 102°C during 48 h and ash by a muffle (Model 10-PR/400, Forns Hobersal S.L., Caldes de Montbui, Barcelona) at 550°C during 7 h. Protein was analyzed utilizing a 2300 Kjeltec Analyzer Unit (Foss Tecator, Höganäs, Sweden) and IMF by an ANKOM^XT15 Extraction System (ANKOM Techonology, Macedon, NY) after hydrolysis (ANKOM^HCL Hydrolysis System).

Sodium chloride was determined following Matissek et al. (1998). A total of 3 g of each sample and 50 mL of milli-Q water were agitated in a shaker-incubator (Rotabit, J.P. Selecta S.A., Abrera, Barcelona) at 190 rpm during 30 min using a magnet. Finally, after the addition of 2 mL of nitric acid, samples were analyzed in a titrator (SM Titrino 702, Metrohm Hispania, Madrid, Spain).

Potassium nitrate and sodium nitrite were also analyzed following the official methods of analysis of meat products described in BOE (1979, 1981, 1982). In the case of potassium nitrate, 4 g of each sample were weighted in an Erlenmeyer flask of 250 mL and 150 mL of ethyl alcohol were added. Samples were agitated in a thermostatic bath (Bunsen BTG, Bunsen, Humanes de Madrid, Madrid) during 1 h. Once cooled, 5 mL of each of the Carrez reagents I and II, prepared with zinc acetate dihydrate and potassium hexacyanoferrate (II) trihydrate, respectively, were added, and milli-Q water were also added to level the flask at 250 mL. The content of this flask was filtered in a flask of 100 mL until its level. The filtrated was discarded and the remaining part was put in another flask of 250 mL, which was placed in a heating plate (Combiplac, J.P. Selecta S.A., Abrera, Barcelona) to evaporate ethyl alcohol, until achieving a volume of 50 mL. Then, this volume was transferred to the flask of 100 mL and milli-Q water was added to level it and it was flipped. Later, a total of 10 mL was transferred to a 50 mL flask and 1 mL of brucine-sulfanilic acid and 10 mL of sulphuric acid were added (color reaction) and it was left to rest 10 min in the dark. This flask was made up to 40 mL with milli-Q water and left to rest 15 min in the dark. Later, it was cooled and levelled. Lastly, a spectrophotometer (Shimadzu UV-1700 Pharmaspec, Kyoto, Japan) was used to determine potassium nitrate content at 410 nm. The procedure to determine sodium nitrite was similar, except for the reagent used for the color reaction, which was prepared mixing equal parts of two solutions. The first solution contained 1.50 g of sulfanilic acid, 50 mL of acetic acid and approx. 200 mL of milli-Q water to make up to 250 mL. The second solution contained 0.075 g of 1-naphthylamine, 50 mL of acetic acid and approximately 200 mL of milli-Q water to make up to 250 mL.

Collagen and water activity (a_w) were determined by near-infrared spectroscopy (measuring range: 850-1100 nm). Each sample was put in a circular small cup of 8.8 mm of depth and 134 mm of diameter that was introduced in the FoodScan™2 equipment (FOSS Iberia S.A., Barcelona).

The contents of α-tocopherol, γ-tocopherol, δ-tocopherol, retinol and cholesterol were determined following the methods described by Bertolín et al. (2018) using ultra-high performance liquid chromatography (ACQUITY UPLC H-Class liquid chromatograph (Waters, Milford, MA, USA) equipped with a silica-based bonded phase column (Acquity UPLC HSS T3, 1.8 μm × 2.1 mm × 150 mm column; Waters, USA), an absorbance detector (Acquity UPLC Photodiode Array PDA eλ Detector; Waters, USA) and a fluorescence detector (2475 Multi λ Fluorescence Detector; Waters, USA)). To determine lipid oxidation, the content of malondialdehyde was analyzed, following Bertolín et al. (2019) using ultra-high performance liquid chromatography coupled to a fluorescence detector.

Firstly, all samples were lyophilized. Then the FA extraction and methylation was carried out following Lee et al. (2012). A Bruker Scion 436-GC gas chromatograph (Bruker, Billerica, MA, USA) equipped with SP-2560 capillary column (100 m × 0.25 mm ID × 0.20 µm film thickness; Supelco, Saint Louis, MO, USA) was used for FA determination. The identification of the FAs was done using certified reference materials (GLC-401, GLC-463, GLC-532, GLC-538, GLC-642 and GLC-643, Nu-Chek Prep Inc., Elysian, MN, USA). The FAs were quantified based on the guidelines described in ISO 12966-4 (2015) as mg of FA/100 mg of total FAs (% of total FAs). The percentages of total saturated FAs (SFA), monounsaturated FAs (MUFA), polyunsaturated FAs (PUFA), n-3 and n-6 and the ratios PUFA/SFA and n-6/n-3 were calculated from individual FA percentages.

Static headspace technique by using a Turbomatrix HS16 sampler (PerkinElmer, Massachusetts, USA) was used to analyze the volatile profile. A total of 4 g of each homogenized sample were placed in vials of 20 mL that were hermetically closed. The samples were thermostatized at 130ºC for 20 min and 1 min of pressurization time. The injection was carried out over 12 s at 25 psi and an inlet temperature of 220ºC. A Clarus 500 gas chromatograph coupled with a mass spectrometer (Perkin Elmer, MA, USA) equipped with a DB-Wax capillary column (60 m × 0.25 mm ID × 0.25 µm film thickness; Agilent Technologies, California, USA) was used to separate and identify the extracted compounds. A flow of 1 mL/min of helium was used as carrier gas. The oven temperature was 45ºC held for 2 min, 45-200ºC at a rate 4ºC/min, and finally to 225ºC at 10ºC/min, and held for 5 min. The mass spectrometer used the electron impact mode with an ionization potential of 70 eV and an ion source temperature of 200ºC. The interface temperature was 220ºC. The mass spectrometer scanned in full scan mode (35-300 m/z). A TurboMass vers. 5.4.2 Workstation was used for the gas chromatograph-mass spectrometer system. Tentative identification of the volatile components was achieved by comparison of the mass spectra with mass spectral data from the Nist MS Search Program 2.0 library and by comparison of previously reported Retention Index with those calculated using a n-alkane (C7-C25) series under the same analysis conditions according to the Van Den Dool & Kratz´s (1963) equation. The relative percentage was expressed as a mass fraction of the total peaks area and fluorobenzene was used as internal standard. The percentages of total aldehydes, ketones, hydrocarbons, alcohols, sulfur compounds, furans and acids were calculated from individual volatile compound percentages.

Androstenone, skatole and indole concentrations in the subcutaneous fat samples were measured by high-performance liquid chromatography as described Pérez-Ciria et al. (2021), using the same laboratory equipment. The concentrations were expressed as µg/g of liquid fat.

All statistical analyses were performed using SAS vers. 9.4 (SAS Institute Inc., Cary, NC, USA). Data were analyzed using the GLM procedure. The model included the type of castration (surgical vs. immunological) as fixed effect. Fresh ham weight and final dry-cured ham weight were included as covariates, when significant (p<0.05), to analyze ham weight losses and the rest of the variables studied, respectively. Androstenone concentration was not statistically analyzed since all values in both types of male pigs were below the detection level of the equipment used (0.20 µg/g of fat); consequently, a descriptive analysis was carried out with this variable.

Normality of the residuals was checked with Shapiro-Wilk’s test using the UNIVARIATE procedure. In cases in which normality was not achieved, variables were transformed with or or Napierian logarithm or or before statistical analysis if it was possible. When normality could not be found with data transformation, Mann-Whitney U-test was carried out to analyze these variables. Homogeneity of variances was checked with Levene’s test. When homoscedasticity was not achieved, Welch’s test was applied.

The pig was the experimental unit. Data are presented in tables as original mean ± standard deviation. A p-value <0.05 was considered as a significant difference.

As shown in Table 1, throughout dry-curing process (composed by the phases of salting, resting, drying, maturing and aging), the type of castration did not influence (p>0.05) on ham weight losses. Thus, the final weight of the end product (dry-cured ham) from IM was similar (p=0.328) to that from SCM.

There were no differences (p>0.05) on color traits of subcutaneous fat between SCM and IM (Table 2). However, hams from IM presented lower (p<0.05) value in both muscles studied than those from SCM. Besides, IM also showed lower a* values in the quadriceps femoris (p=0.044) and b* in the biceps femoris (p=0.045) than SCM.

Hams from IM had similar (p>0.05) subcutaneous fat thickness and marbling than those from SCM (Table 3). Likewise, the type of castration had no effect (p=0.283) on maximum stress (67.57 ± 12.62 vs. 60.07 ± 12.38 in SCM and IM, respectively; data not shown in Table 3).

The moisture, ash, protein and IMF of hams were not affected (p>0.05) by the type of castration (Table 4). Regarding salt contents, sodium chloride and sodium nitrite concentrations were also similar (p>0.05), but IM presented lower (p=0.042) potassium nitrate content than SCM. No influence (p>0.05) of the type of castration was observed on collagen content, a_w, and the contents of tocopherols, cholesterol and malondialdehyde. However, hams from IM presented lower (p=0.039) retinol content than those from SCM.

Table 5 shows the impact of the type of castration (surgical vs. immunological) of male pigs on the FA profile of the biceps femoris muscle of dry-cured hams. No significant differences (p>0.05) were observed between SCM and IM in total SFA, MUFA, PUFA, n-3 and n-6 proportions and neither in PUFA/SFA and n-6/n-3 ratios.

A total of 39 volatile compounds were identified in the biceps femoris muscle of dry-cured hams (Table 6), including the following groups: aldehydes (9), ketones (7), hydrocarbons (6), alcohols (11), sulfur compounds (1), furans (2) and acids (3).

The type of castration had no influence (p=0.618) on total aldehydes percentage. However, within this group, hams from IM presented higher (p=0.006) proportion of heptanal, nonanal and (E)-2-octenal than those from SCM. No impact (p=0.708) of the type of castration was observed on the percentage of total ketones. It is worth noting that inside this group, hams from IM showed lower (p=0.009) proportion of 2,3-butanedione and higher (p=0.006) of 2-octanone. The percentage of total hydrocarbons was also similar (p=0.183) in both experimental groups regardless of the type of castration. Within this group, hams from IM had lower percentage of α-pinene (p=0.027) and limonene (p=0.012) and greater of methylbenzene (p=0.024) than those from SCM. Immunocastration of male pigs reduced (p=0.019) the percentage of total alcohols in comparison to surgical castration, mainly due to the lower proportion of 1-octen-3-ol (p=0.009) and 1-octanol (p=0.008), although hams from IM presented higher (p=0.006) 2-methyl-1-propanol percentage than those from SCM. Likewise, hams from IM showed lower (p=0.039) percentage of total sulfur compounds than those from SCM owing to the lower (p=0.039) proportion of dimethyl disulfide. The type of castration had no effect (p=0.183) on the percentage of total furans, although IM presented lower (p=0.015) proportion of 2-ethylfuran than SCM. Finally, hams from IM did show greater (p=0.012) percentage of total acids than those from SCM, because of the higher proportions of 2-methylpropanoic (p=0.006), butanoic (p=0.009) and hexanoic (p=0.024) acids.

Table 7 shows the impact of the type of castration of male pigs on the boar taint compounds. All hams, both from SCM and IM groups, presented low androstenone concentrations, below the detection level of the equipment (0.20 µg/g) (data not shown). However, IM hams had higher skatole (p=0.004) and indole (p=0.029) concentrations than those from SCM.