Leptin receptor gene influences pig gut microbiota both through feed intake and independently Authors: Rafael Suárez-Mesa, Houda Laghouaouta, Jordi Estellé, Ahmad Reza Seradj, Roger Ros-Freixedes, Ramona N Pena, Joan Estany Published: 2025 ABSTRACT AbstractBackgroundAdiposity results from a complex interplay among host genetics, diet and gut microbiota. Major genes are valuable models for unravelling how host genetics influence microbiota. The defective recessive T allele of the rs709596309 (C > T) polymorphism in the porcine leptin receptor gene (LEPR) leads to hyperphagia and obesity. Here, we report the first results on the impact of LEPR on microbiota composition in pigs. To this end, we compared saliva and faecal microbiota of 46 TT and 48 C− (22 CC and 26 CT) Duroc pigs from the same line and correlated the observed microbial changes with feed intake, faecal volatile fatty acid composition and lean growth.ResultsThe TT genotype induced a shift in the faecal but not the saliva microbiota, accompanied by a substitution of faecal isovaleric acid for butyric acid. Fourteen genera were identified as microbial candidates influenced by LEPR in the gut. Of these, six genera were prevalent across all pigs, including three Oscillospiraceae genera (UCG-005, NK4A214, and UGC-002) and Treponema, which were less abundant in TT pigs, as well as Subdoligranulum and Coprococcus, which were more abundant. Oscillospiraceae UCG-005 experienced the greatest decrease with the TT genotype (-3.20%), while Subdoligranulum (+ 1.60%) showed the greatest increase. The most abundant genera in TT pigs were positively correlated with butyric acid and negatively correlated with isovaleric acid, and vice versa. The association of these LEPR-associated genera with body weight and fatness aligned with their relationship to faecal butyric and isovaleric acids. Variation in feed intake strongly explained their abundance, such that after accounting for feed intake, LEPR was associated only with Oscillospiraceae UCG-005 and four occasional bacteria (Frisingicoccus, Lachnospiraceae UCG-009, Lachnospiraceae XPB1014 and Megasphera). These occasional genera were found in 56.4% to 77.7% of the pigs, with only Megasphera exhibiting higher abundance in TT pigs.ConclusionsDisfunction of LEPR led to a shift in the gut microbiota, favouring the enrichment of starch-degrading rather than protein-degrading genera. Our results underscore the importance of feed intake as a microbiota-altering factor, as well as the potential of Oscillospiraceae UCG-005 as a LEPR-specific microbial biomarker associated with obesity.Supplementary InformationThe online version contains supplementary material available at 10.1186/s42523-025-00508-7.Keywords: Adiposity, Fat, Genetics, Intestinal microbiota, Growth, Obesity, Porcine Introduction The root cause of fat accumulation is an excess of caloric intake over expenditure, and this depends on both the genetics of the individual and the environment. The intestinal tract is the major barrier that connects the host´s internal environment with the external environment and, as such, it plays a key role in nutrient absorption. Gut microbiota is an essential component of the intestinal tract, since, as part of the digestion process, gut microorganisms produce fermentation metabolites, mostly short-chain fatty acids, that then can be absorbed and used by the host. Since pioneering works about two decades ago [1–3], numerous studies in mice and humans have shown that there is a complex interplay between host genetics, diet and gut microbiota in relation to adiposity [4–6]. The use of individuals carrying major genetic variants related to metabolism, particularly when subjected to common environmental conditions, can help unravel the specific influence of host genetics on shifts in gut microbial composition. A few host genes have been examined for their impact on gut microbiota composition [4, 7], with the genes encoding leptin and the leptin receptor being among the most studied due to their direct relationship with hyperphagia and obesity. The gut microbiota of homozygous obese mice with the non-functional mutation ob in the leptin gene has been shown to be less abundant in Bacteroidota (formerly known as Bacterioidetes) and more abundant in Firmicutes [2]. A similar shift in microbiota composition was observed in homozygous obese mice carrying the db mutation in the leptin receptor gene [6, 8]. In the case of db/db mice, however, the microbiota alteration is not solely due to hyperphagia but also to a direct regulatory function of leptin signalling [9]. Although several studies in mice and humans confirmed a positive relationship of obesity with the Firmicutes-to-Bacteroidota ratio in the gut microbiota [10], others have found no association [11] or even an opposite trend [12]. Moreover, refined diets can produce similar effects regardless of the presence of obesity [13]. Consequently, these phyla have been questioned as reliable biomarkers for obesity [14], and instead, shifts at lower taxonomic levels are being sought as potential targets [15, 16]. Pigs can serve as a model for human-related research due to the similarity in their digestive physiology and, to some extent, intestinal microbiotas [17]. Eliminating excess fat is a common objective in meat species, especially in pigs, where selective breeding primarily focuses on increasing lean content and feed efficiency. As a result, most alleles with major effects on increased fat content have been eliminated from commercial lines. The recessive T allele of the rs709596309 (C > T) single nucleotide polymorphism in the leptin receptor gene (LEPR) [18], which mirrors the effects observed in db/db mice, is still present in some Duroc lines used in pork quality markets. Previous research in these lines has shown that homozygous pigs for the T allele exhibited increased fatness and circulating leptin [19–21] as well as a more favourable energy balance [22]. Experimental studies in pigs on the impact of the host genome on gut microbiota in relation to adiposity have focused on microbial comparisons between fat and lean breeds, either directly from their faeces [23–25] or by transplanting them into germ-free mice [26–30]. However, studies examining the effects of single mutations have been lacking. Mice transplanted with the faecal microbiota from fatty breeds generally exhibited a higher Firmicutes–to-Bacteroidota ratio and replicated the phenotype of their donors. Nevertheless, results were less consistent when comparisons were made directly in the faeces or at lower taxonomic levels. Despite this, evidence supports genus-level differences in the faeces of pigs with divergent backfat thickness [31]. This study aimed to investigate how the TT genotype for the rs709596309 (C > T) polymorphism in the leptin receptor gene affects microbiota composition and nutrient utilization in pigs. To do this, we compared the microbial composition of saliva and faeces between TT and C− (CC and CT) Duroc pigs, which were characterised for feed intake, faecal volatile fatty acid composition and lean growth after exposure to the same diet and environmental conditions. Materials and Methods Experimental design, animals and performance phenotypes The pigs used in this research were 94 barrows (46 TT and 48 C−, of which 26 were CT and 22 CC), all born on the same farm from 37 contemporary litters of the same Duroc line [32]. Each litter was born to a different dam. The TT pigs were from 11 sires and 23 dams and the C− pigs from 13 sires and 27 dams. Among these, 8 sires and 13 dams produced offspring of both genotypes. All pigs were genotyped for the LEPR rs709596309 C > T variant as described by [22]. At around 65 days of age (standard deviation, SD: 2), pigs were delivered in one batch to the IRTA swine test station in Monells (Department of Climate Action, Food and Rural Agenda, Govern of Catalonia, Spain) for performance testing. There, the barrows were housed in alternating pens of two pigs (1 m2/pig) with the same genotype (TT or C−) and reared under the same experimental conditions until 202 days of age. To mitigate potential maternal effects and competition at feeding, live body weight at entry was equalised across genotypes (0.9 kg of difference) and within pen (0.3 kg of difference, on average). During the fattening period, pigs had ad libitum access to commercial growing (until 90 days of age; 10.0 MJ/kg of net energy, 15.0% crude protein) and finishing (from 90 days of age to slaughter; 10.2 MJ/kg of net energy, 13.7% crude protein) diets (Table S1; Esporc, Riudarenes, Girona, Spain). No antibiotics were given during the trial. All pigs were individually weighed at 155 days of age (on-test) and at 202 days of age (off-test), at which time their backfat and loin thickness at 5 cm off the midline at the position of the last rib were also measured with a portable ultrasonic scanner (Piglog 105; Frontmatec, Kolding, Denmark). Feed intake was recorded daily on a pen basis and summed up throughout the period. The average daily gain and feed intake of the pen were calculated thereafter. Three random samples of the finishing feed were taken for further analyses. All pigs were slaughtered at 203 days of age in the same abattoir, where carcass weight, carcass backfat and loin thickness, and carcass lean content were recorded using an automatic carcass grading system (AutoFOM, SFK-Technology, Herlev, Denmark). Carcass backfat and loin thickness were measured at 6 cm off the midline between the third and fourth last ribs, and carcass lean content was estimated using the corresponding prediction equation approved for Spain (Commission Implementing Decision (EU) 2020/113 of 23 January 2020). A full description of the experiment, along with the results on the impact of the LEPR gene on pig performance and quality traits, is detailed in [21]. A summary of the performance traits used in this research for each LEPR genotype is presented in Table 1. Table 1Performance of pigs used in the experiment by LEPR genotypeLEPR genotypeTT (n = 46)C− (n = 48)Trait aMeanSD bMeanSD bOn-test Age at on-test, days154.43.1155.41.4 Live body weight, kg92.99.492.07.9 Backfat thickness, mm15.73.513.92.6 Loin thickness, mm43.44.442.04.1Off-test Age at off-test, days201.43.1202.41.4 Live body weight, kg136.711.2131.011.9 Backfat thickness, mm26.73.823.84.2 Loin thickness, mm45.33.946.14.8 Daily weight gain on test, g/d935127825130 Daily feed intake on test, g/d38483393456323 Feed efficiency on test, g/kg41414424245511Carcass traits Age at slaughter, days202.43.1203.41.4 Carcass weight, kg102.39.598.19.4 Carcass backfat thickness, mm39.17.032.16.5 Carcass loin thickness, mm37.87.942.06.1 Lean content, %34.67.241.46.3a Faeces were sampled 4 days before off-test age. Daily feed intake is the average per day from on-test to off-test on a pen basis. Adapted from Suárez-Mesa et al. [21]b SD: Standard deviation Performance of pigs used in the experiment by LEPR genotype a Faeces were sampled 4 days before off-test age. Daily feed intake is the average per day from on-test to off-test on a pen basis. Adapted from Suárez-Mesa et al. [21] b SD: Standard deviation Saliva and faeces sampling procedures A sample of saliva and rectal content of each pig was collected 4 days before the end of the trial (198 days of age). Saliva samples were taken by buccal swabs between 12:00 and 14:00 with the Performagene kit (DNA Genotek Inc., Ottawa, Canada). Since the pigs had no fixed feeding times, sampling was carried out independently of feeding. Saliva was collected for 30 s by moving a sterile cotton sponge and mopping the saliva on the pig’s cheek pouch and under the tongue. The swab was placed in the tube supplied by the manufacturer and shaken vigorously to thoroughly mix the sample. Immediately after collecting the saliva, a stool sample of around 200 g was obtained directly from the rectal ampulla. Both the saliva and stool samples were snap-frozen in liquid nitrogen and then stored in plastic bags at − 80 °C until needed. Chemical composition of faeces and digestibility Chemical analyses were conducted on feed and faecal samples following the AOAC methods [33]. The dry matter was determined by oven drying at 105 °C until a constant weight was achieved (AOAC ref. 934.01) and the ash content was measured by incinerating the sample in a muffle furnace at 550°C for four hours (AOAC ref. 942.05). Crude fat was determined by extracting lipids using hexane as organic solvent (AOAC ref. 920.39), while crude protein was calculated by determining the nitrogen content through the combustion method (AOAC ref. 990.03) and multiplying the nitrogen value by a conversion factor of 6.25. The apparent total tract digestibility of each nutrient was estimated using acid-insoluble ash as an internal marker [34] and calculated as 100 minus the quotient between the ratio of the content of each nutrient in faeces and feed and the ratio of the marker content in faeces and feed [35]. Determination of volatile fatty acids in the faeces Frozen faecal samples were thawed at 4 °C overnight. Two grams of faeces were added to tubes containing 10 mL of a deproteinising solution, which consisted of 20 mL/L ortho-phosphoric acid and 2 g/L of 4-methylvaleric acid (used as internal marker) in distilled water. The tubes were centrifuged at 17,400 × g for 10 min at 4 °C (Eppendorf 5425R, Eppendorf AG, Hamburg, Germany). The resulting supernatant was filtered through a 0.50 μm filter (Millipore Co., Bedford, MA) and pipetted into 2.0 mL gas chromatography vials (Supelco, Cat. No. 27265, Bellefonte, PA). Quantitative determination of volatile fatty acids (VFA; AA: acetic acid; PA: propionic acid; BA: butyric acid; VA; valeric acid; IBA: isobutyric acid; and IVA: isovaleric acid) was performed by injecting a 2 µL-sample from the above vials into a gas chromatograph (7890 A, Agilent Technologies, Palo Alto, CA, USA) equipped with a DB-FFAP capillary column (30 m × 0.25 mm ID × 0.25 μm, Agilent Technologies, USA) and a flame ionisation detector using helium (99.99% purity [C50], Carburos Metálicos, Spain) as the carrier gas at 1 mL/min. The gas chromatograph was operated at 150 °C at the injector, and 250 °C at the detector, with variable oven temperatures. The column temperature was initially 90 °C, then increased to 150 °C at 15 °C/min, to 170 °C at 5 °C/min, and finally 240 °C at 20 °C/min and kept at this temperature for 5 min (total time of 16.5 min). The quantification was carried out through area normalisation by adding into each sample a mix of the target VFA as internal standard (Sigma-Aldrich, St. Louis, MO, USA). Total VFA content was calculated as the sum of the six individual VFA and the proportion of each VFA as the percentage of each individual VFA relative to total VFA. Microbiota DNA extraction, 16 S rRNA sequencing and data processing Microbiota DNA was extracted from saliva and faecal samples using the HigherPurity™ Soil DNA Isolation Kit (Canvax Reagents SL, Valladolid, Spain), in accordance with the manufacturer’s instructions, with the following modification: frozen samples (0.1 g) were disrupted in a bead homogeniser (BeadMill 4, ThermoFisher) with two borosilicate 2-mm beads and a lysis buffer, and incubated at 95 °C for five minutes. This step was repeated twice. The sample tubes were then spun at 10,000 g for 30 s, and the supernatant was transferred to a fresh tube. Finally, the DNA was eluted from the column in a volume of 100 µL. DNA concentration and quality were assessed using spectrometry (Nanodrop-100) and fluorometry with the dsDNA HS Assay kit (Qubit 4, Invitrogen, Carlsbad, CA). The hypervariable region V3-V4 of the 16 S rRNA was amplified using primers previously described in [36] and sequenced with paired-end reads with the Illumina MiSeq platform (Illumina, San Diego, CA, USA) at the Centre for Genomic Regulation (CRG, Barcelona, Spain). Raw paired-end sequence reads were processed into Amplicon Sequence Variants (ASVs) using the Divisive Amplicon Denoising Algorithm 2 (DADA2) pipeline [37] in R [38] by following the recommendations of the authors in the tutorial. Briefly, quality control of raw sequence reads was performed to remove low-quality bases and adapter sequences were trimmed. ASVs were created and denoised, while chimeric sequences and singletons were removed. The final ASV table contained a total of 28,983 ASVs from 187 samples (93 of saliva and 94 of faeces). The taxonomic annotation of ASVs was conducted at the genus level using the SILVA v.138 database [39]. After data processing, each sample had an average of 65,075 sequencing reads. There was no difference in sequencing reads between genotypes, both in saliva (64,214, for TT, and 64,638, for C-, p-value = 0.90) and faeces (65,658, for TT, and 65,749, for C−, p-value = 0.98). Microbiota data analysis The microbiota composition of saliva and faeces was analysed separately. For diversity analyses, samples were rarefied with phyloseq v.1.38.0 [40] to an even sampling depth (21,984 for saliva and 18,335 for faeces), which corresponded to the minimum number of counts. α-diversity was estimated using the Shannon index, while β-diversity was assessed through Bray-Curtis dissimilarities. Using the betadisper function from the vegan_2.6-2 package [41], the Bray-Curtis distance from each sample to the centroid of its genotype was calculated to test for homogeneity of variance within groups. Microbiota dissimilarity between groups (saliva vs. faeces, and TT vs. C−) was assessed using a permutational multivariate analysis of variance (PERMANOVA) on Bray-Curtis distances with the adonis2 function in phyloseq. Estimation of the effect of LEPR The effect of the LEPR genotype (TT vs. C−) on each individual trait (Shannon index, chemical composition, digestibility and VFA), as well as on microbial abundance at the phylum, family and genus levels was tested using an F-test based on a model including the genotype. The sire and dam effect were not included as random factors in the final model because they were not significant for the investigated traits based on the Wald test. The effect of feed intake on microbiota composition was estimated by regressing microbial abundance on the average daily feed intake during the test. When appropriate, the LEPR-specific effect was determined by adjusting for feed intake variation through the inclusion of the average daily feed intake during the test as a covariate in the model. Microbiota abundance of each microbial taxa was calculated as the percentage of the taxa counts relative to the total counts in the sample. Due to its compositional nature [42], microbial abundance and the other traits expressed as percentages were transformed using the logit transformation, and the results were then back-transformed to the original scale for presentation. Missing genera were imputed to 10− 6 to replace zeros, a value that is below the minimum abundance observed across the genera. Since daily feed intake was recorded per pen, adjustments for feed intake were based on the average of the individual values from the two pigs in each pen. To minimize the risk of false positives, differences in microbial abundance between LEPR genotypes were tested using the conservative Bonferroni multiple testing correction, with adjusted p-values at the phylum, family and genus levels set at 4·10− 3, 10− 3 and 5· 10− 4, respectively. The critical adjusted p-value at the genus level was relaxed to 8· 10− 4 for genera associated with VFA. In addition, for significant genera, the likelihood ratio chi-square was used to test for unequal distribution of zeros between genotypes. All previous analyses, including the estimation of Pearson correlations and linear regression coefficients between genus abundance and recorded individual traits, were conducted using JMP Pro 17 (SAS Institute Inc., Cary, NC, USA). Results Composition of microbiota in saliva and faeces The composition of the microbiota composition in saliva and faeces at the phylum and family levels is presented in Fig. 1. A total of 325 genera (belonging to 117 families across 19 phyla) were identified in saliva, compared to only 228 genera (from 79 families across 17 phyla) in faeces. The two most abundant families in saliva were Yersiniaceae (39.0%) and Moraxellaceae (6.8%), from the Proteobacteria phylum, while, in faeces, the most abundant families were Lachnospiraceae (21.2%) and Oscillospiraceae (13.9%), from the Firmicutes phylum. The microbiota composition in saliva and faeces differed, both in terms of α-diversity (Fig. 2) and β-diversity Fig. S1.A). Rarefied data revealed that the faecal microbiota had a higher α-diversity (Shannon index, p-value < 0.001) but lower dispersion compared to the saliva microbiota (Fig. 2). Genera not present in at least 50% of the saliva samples or 50% of the faecal samples were filtered out to further reduce the risk of detecting false-positive associations of microbiota with LEPR and investigated traits. Thus, microbial abundance analyses were performed using 64 genera (from 36 families and 6 phyla) in saliva samples and 109 genera (from 36 families and 12 phyla) in faecal samples. The results remained unchanged when genera present in at least 50% of the samples for a given LEPR genotype (8 in saliva and 10 in faeces) were included in the analyses. Microbiota composition of saliva and faeces. Taxa are coloured by family and families are ordered from highest to lowest abundance in faeces. The phylum is indicated before each family Comparison of α-diversity between LEPR genotypes in saliva and faeces. The box plots show the median (centre line), interquartile range (box limits) for Shannon’s index The effect of LEPR on pig microbiota The LEPR genotype showed no association with oral microbiota diversity, as evidenced by the α-diversity (Shannon index, p-value = 0.99, Fig. 2), β-dissimilarity (p-value = 0.61, Fig. S1.B) and β-dispersion (p-value = 0.45, Fig. S2.A) metrics. Furthermore, no significant effect of LEPR was found on the relative abundance of taxa in the salivary microbiota at the phylum and family levels (Fig. S3). The LEPR genotype also showed no significant association with faecal α-diversity (Shannon index, p-value = 0.32; Fig. 2) or β-diversity dispersion (p-value = 0.94; Fig. S2.B). However, it did influence Bray-Curtis β-dissimilarity (p-value < 0.0001; Fig. S1.C), indicating that LEPR contributed to shaping the composition of the faecal microbiota. Six phyla in the faecal microbiota differed between LEPR genotypes (Fig. 3), although each comprised no more than one to four genera. In fact, of the 109 genera used in the faeces, 83 belonged to Firmicutes and 10 to Bacteroidota. As expected, Firmicutes were more abundant in TT pigs (86.2% vs. 81.9%, p-value = 0.04) and Bacteroidota in C− pigs (11.4% vs. 9.0%, p-value = 0.20). However, no significant differences were found between LEPR genotypes after Bonferroni correction. Due to the high individual variability of Bacteroidota, the Firmicutes–to-Bacteroidota ratio did not differ between LEPR genotypes (22.9, for TT, and 16.4, for C−, p-value = 0.53). Six families showed significant differences between LEPR genotypes (Fig. 3), but only four had significant genera (Fig. 4) - three bacterial families and one archaeal family. Further research at the genus level within the three significant bacterial families revealed that the TT pigs had a lower faecal relative abundance of Oscillospiraceae (UCG-005, NK4A214, and UGC-002 genera) and Spirochaetaceae (Treponema genus) families, and a higher abundance of the Veillonellaceae family (Dialister and Megasphera genera) compared to C − pigs. Major changes in faecal microbiota occurred in Oscillospiraceae UCG-005 (4.82%, in TT, vs. 8.02%, in C−), Treponema (0.31%, in TT, vs. 1.39%, in C−), Oscillospiraceae NK4A214 (1.72%, in TT, vs. 3.09%, in C−) and Oscillospiraceae UCG-002 (1.21%, in TT, vs. 2.42%, in C−), which decreased by more than 1% in TT pigs. The positive shift of TT pigs on Dialister (0.06%, in TT, vs. < 0.01%, in C−) and Megasphera (0.04%, in TT, vs. < 0.01%, in C−) was smaller percentage-wise, but larger in relative terms. These two bacteria, in addition to being less abundant, were absent in a sizeable proportion of individuals, particularly in C− pigs (10.6%, in TT, and 26.6%, in C−, for Megasphera, and 9.6%, in TT, vs. 26.6%, in C−, for Dialister). Nevertheless, when considering only the pigs that harboured these bacteria, TT pigs continued to show a higher proportion of Megasphera (0.75% vs. 0.25%, p-value < 0.01) and Dialister (0.82% vs. 0.25%, p-value < 0.01). Differential abundance of microbiota phyla and families between LEPR genotypes in faeces. Each phylum is represented by a distinct colour. Bars represent the difference between the TT and C− genotypes, expressed as a percentage. Underlined phyla and families were significantly different between LEPR genotypes at the Bonferroni-adjusted significance level Differential abundance of microbiota genera between LEPR genotypes in faeces. Each family is represented by a distinct colour. The difference in abundance between genotypes is presented as the abundance in the TT genotype minus the abundance in the C− genotype, expressed as a percentage. Only significant genera that belong to families with significant differences between LEPR genotypes are depicted. Both genera and families showed significant differences at the Bonferroni-adjusted level of significance Since LEPR has long been associated with feed intake (Table 1), and feed intake can influence faecal microbiota composition, we investigated the relationship between the faecal abundance of the six bacteria mentioned above and feed intake. Through regression analysis, we showed that the abundance of all six bacteria varied in line with the observed effect of LEPR on feed intake (Fig. 5). Thus, to determine whether LEPR specifically affects microbiota composition independently of feed intake, we adjusted the effect of LEPR on bacterial abundance for the amount of feed consumed during the test. When feed intake was included as a covariate in the model, LEPR remained significantly (p-value < 0.01) associated with Oscillospiraceae UCG-005, Megasphera and Treponema, but not with NK4A214 (p-value = 0.06), Oscillospiraceae UCG-002 (p-value = 0.13) and Dialister (p-value = 0.32). Conversely, with LEPR in the model, feed intake was no longer significantly associated with Oscillospiraceae UCG-005 (p-value = 0.76) or Megasphera (p-value = 0.20). Taken together, these results indicate that Oscillospiraceae UCG-002, Oscillospiraceae NK4A214 and Dialister respond primarily to variation in feed intake, whereas Oscillospiraceae UCG-005 and Megasphera are more specifically associated with LEPR. Relationship between genus abundance and feed intake. Linear regression of the logit-transformed relative abundance of bacterial genera (Y) on daily feed intake (DFI). Depicted bacteria, along with their respective families, exhibited significantly different abundances between LEPR genotypes at the Bonferroni-adjusted level. Genotypes are indicated in red (TT) and blue (C−) The Methanobacteriaceae archaeal family also differed across LEPR genotypes (Fig. 3), as the Methanobrevibacter genus, its only representative, was less abundant in TT pigs (Fig. 4; <0.01% vs. 0.09%, for TT and C− pigs, respectively). This was because TT pigs had a higher proportion of individuals without Methanobrevibacter (19.2% vs. 3.2%, p-value < 0.05). The effect of LEPR on Methanobrevibacter disappeared when only pigs with the archaeon present were considered (p-value = 0.55). In any case, the effect of LEPR was no longer significant after adjusting for feed intake (p-value = 0.06). The effect of LEPR on digestibility and VFA No differences were detected between LEPR genotypes for the chemical composition of faeces, except for dry matter content (Table 2), which was lower in TT pigs (20.8%, for TT, vs. 22.4%, for C−, p-value = 0.03). The LEPR genotypes also did not differ in the digestibility of protein and fat content (Table 2) or in the total production of VFA (Table 3); however, they did affect VFA composition (Table 3). Specifically, TT pigs had higher BA (11.6% vs. 10.7%, p-value = 0.04) and lower IVA (1.1% vs. 2.1%, p-value < 0.0001) compared to C − pigs. The difference in IVA between LEPR genotypes persisted after adjusting for feed intake (p-value = 0.03), but not for dry matter (p-value = 0.26) or BA (p-value = 0.87). Table 2Chemical composition of faeces and digestibility by LEPR genotypeTraitLEPR genotype ap-valueTTC−Dry matter, %20.8 (19.8; 21.8)22.4 (21.4; 23.4)0.03Crude protein, %19.5 (18.8; 20.1)18.7 (18.1; 19.4)0.11Crude fat, %7.2 (6.1; 8.4)6.5 (5.7; 7.5)0.35Digestibility, %Dry Matter95.7 (95.5; 95.9)95.5 (95.3; 95.7)0.11Protein73.8 [72.0; 75.7)75.4 (3.5; 77.1)0.25Fat79.5 (75.4; 83.2)81.5 (78.1; 84.4)0.45a Means (95% confidence interval) Chemical composition of faeces and digestibility by LEPR genotype a Means (95% confidence interval) Table 3Volatile fatty acid content and composition in faeces by LEPR genotypeTraitaLEPR genotypeTT C- p-valueVolatile, Mmol/g129.5 (120.8; 138.1)131.0 (122.6; 139.5)0.80Short-Chain, % Acetic acid64.1 (64.0; 64.2)63.7 (62.9; 64.5)0.52 Propionic acid19.5 (19.1; 20.0)19.3 (18.9; 19.8)0.55 Butyric acid11.6 (11.0; 12.3)10.7 (10.2; 11.3)0.04 Valeric acid1.5 (1.4; 1.6)1.6 (1.4; 1.7)0.41Branched-chain, % Isobutyric acid1.8 (1.6; 2.0)1.8 (1.7; 2.0)0.63 Isovaleric acid1.1 (0.9; 1.3)2.1 (1.8; 2.6)<0.0001a Means (95% confidence interval) Volatile fatty acid content and composition in faeces by LEPR genotype a Means (95% confidence interval) Faecal microbiota and VFA The seven genera shown in Fig. 4 displayed moderate correlations with BA (ranging from − 0.30, for Treponema, to 0.36, for Megasphera), as well as with IVA (ranging from 0.22, for Treponema, to 0.38, for Oscillospiraceae NK4A214), with the exception of Dialister and Megasphaera, which did not significantly correlate with IVA (Fig. 6). Sixty other genera were associated with either BA or IVA (p-value < 0.05), with Lachnospiraceae ND3007 and Christensenellaceae R-7 showing the strongest correlation with BA (0.60) and with IVA (0.40), respectively. Of these, seven were also significantly associated with LEPR at the Bonferroni-adjusted level (p-value < 8· 10− 4), even though their families were not. Notably, five of these genera belong to the Lachnospiraceae family: Coprococcus (1.96%, in TT, vs. 1.22%, in C−), Lachnospiraceae NK3A20 (0.04%, in TT, vs. < 0.01%, in C−), Frisingicoccus (< 0.01%, in TT, vs. 0.23%, in C−), Lachnospiraceae UCG-009 (< 0.01%, in TT, vs. 0.01%, in C−) and Lachnospiraceae XPB1014 (< 0.01%, in TT, vs. 0.05%, in C−). The other two genera were Subdoligranulum (2.95%, in TT, vs. 1.35%, in C−), from the Ruminococcaceae family, and dgA-11 gut group (< 0.01%, in TT, vs. 0.05%, in C−), from the Rikellkenaceae family. Only three of the Lachnospiraceae genera remained associated with LEPR after adjusting for feed intake (Frisingicoccus, UCG-009 and XPB1014, p-value < 0.05), Of these, only Lachnospiraceae XPB1014 retained the association when the analysis was limited to pigs that harboured the bacterium. The association of LEPR with Frisingicoccus and Lachnospiraceae UCG-009 was due to a higher proportion of TT pigs lacking these bacteria (19.2%, in TT, vs. 3.2%, in C−, for Frisingicoccus, and 30.9%, in TT, vs. 12.8%, in C−, for Lachnospiraceae UCG-009, p-value < 0.05). Remarkably, the abundance of all three bacteria was negatively related to BA and positively related to IVA (Fig. 6). In contrast, feed intake was the main driver of the increased abundance of the other two Lachnospiraceae genera (Coprococcus and NK3A20) and Subdoligranulum in TT pigs (p-value < 0.05). These genera showed and inverse relationship with BA and IVA, being positively associated with BA and negatively with IVA. Neither LEPR (p-value = 0.17) nor feed intake (p-value = 0.25) significantly influenced the dgA-11 gut group when analysed together in the same model. As shown in Fig. 6, the genera that increased BA were associated with faeces having higher moister, fat and protein content, while those that increased IVA showed the opposite pattern. Heatmap correlation between genus abundance and chemical composition in faeces. Pearson correlations of genus abundance (right side) with volatile fatty acid content (VFA), individual volatile fatty acids, dry matter, crude fat and crude protein and apparent digestibility (bottom side). Correlations were calculated on the logit-transformed values of traits expressed in percentage. Significant correlations (p-value < 0.05) had a value higher (lower) than 0.20 (-0.20). Positive correlations are represented in blue, while negative correlations are represented in red Relationship between faecal microbiota and fatness In line with the effects of LEPR, the abundance of the 14 genera previously identified as LEPR-sensitive (Fig. 6) showed a stronger association with fat content than with lean mass (Fig. S4). The correlation of these genera with body weight, fat and lean ranged from − 0.45, for Lachnospiraceae XPB1014 with carcass backfat thickness, to 0.44, for Subdoligranulum with backfat thickness at off-test. The relationship between microbial abundance and fatness mirrored the effect of the genotype TT on microbiota. Thus, the LEPR-specific bacteria Oscillospiraceae UCG-005, Treponema and Lachnospiraceae XPB1014 were negatively correlated with fatness, while Megasphera was positively correlated (Fig. 7), even when the analysis was limited to pigs harbouring the bacteria (p-value < 0.01, for Lachnospiraceae XPB1014, and p-value = 0.02, for Megasphera). In contrast, the negative association between Frisingicoccus and backfat thickness vanished when pigs without Frisingicoccus were excluded (p-value = 0.16). The greater proportion of TT pigs lacking Lachnospiraceae UCG-009 was insufficient to establish an association with fatness (p-value = 0.21). The relationship of feed intake-driven bacteria with fatness paralleled their association with feed intake: positive for Subdoligranulum and Coprococcus and negative for Oscillospiraceae NKA214 and Oscillospiraceae UCG-002 (Fig. S5). The positive relationship of Dialister and Lachnospiraceae NK3A20 with fatness was due to their higher prevalence in more voracious TT pigs (p-values of 0.25 and 0.84, respectively, when only pigs carrying these bacteria were considered). Although Methanobrevibacter and dgA-11 gut group were not affected by LEPR and feed intake (p-value < 0.05) when both were included in the model, they still showed a negative relationship with fatness (Fig. S4). Relationship of LEPR-specific genera with fatness. Linear regression of backfat thickness (Y) at off-test (4 days after faecal sampling) on the logit-transformed relative bacterial abundance (X). The faecal abundance of the represented genera was influenced by LEPR regardless of feed intake. First row: prevalent bacteria; second row: occasional bacteria that were associated with backfat thickness (p-value < 0.05) in pigs harbouring these bacteria; and third row: occasional bacteria that were not associated with backfat thickness (p-value > 0.05) in pigs harbouring these bacteria. Genotypes are indicated in red (TT) and blue (C−) Discussion We have previously shown that TT pigs consume more feed and have higher fat levels compared to their C– counterparts [21]. Here, we show that TT pigs also produce moister faeces, indicating that their increased faecal output is likely linked to a faster feed passage rate through the digestive tract. This change was not enough to reveal a decline in digestive efficiency or total VFA, as suggested by previous reports [43], but sufficient to detect a shift in VFA composition, in line with other studies showing that VFA are more sensitive to feed intake than to digestibility [44, 45]. LEPR-driven hyperphagia led to a shift from IVA to BA. Since BA (unbranched-VFA) stimulate colonic smooth muscle contraction and IVA (branched-VFA) induce relaxation [46], the higher BA and lower IVA observed in TT pigs may have further enhanced their intestinal motility. Both BA and IVA are metabolites from microbial fermentation, with BA derived from carbohydrates and IVA from proteins, specifically leucine [47, 48]. Fermentation in the gastrointestinal tract is not neutral with respect to feed intake, as a higher BA to IVA ratio favours feed consumption [49, 50]. With TT pigs having a higher ratio, this may have also helped boost their appetite. Branched-VFA (IVA) are produced when unabsorbed proteins (leucine) in the small intestine are fermented by colonic microorganisms. Microbial activity in the gut reflects the trade-offs between dietary carbohydrate and protein fermentation [51]. The pigs used here were fed ad libitum a high-leucine diet (0.77% SID leucine; [21]), which may have favoured protein-degrading microbial populations and likely contributed to accentuating the differences in IVA between LEPR genotypes. The fact that LEPR affected faecal but not oral microbiota indicates that LEPR causes a displacement of microbiota during digestion. Consistent with findings in humans [52], mice [2], and pigs [23, 53–56], Firmicutes was the most abundant phylum. The dominating abundance of Firmicutes aligns with other reports showing that their presence increases with free access to feed [57, 58], age [55, 59], and proximity to the distal end of the intestinal tract [24, 60, 61]. As a result, LEPR exerted the greatest influence on certain prevalent genera within Firmicutes, although not in the anticipated direction favoring fatness. Specifically, TT pigs exhibited a lower abundance of the Oscillospiraceae UCG-005, UCG-002 and NK4A214, which were only partially compensated by increases in Subdoligranulum and Coprococcus from the Ruminococcaceae and Lachnospiraceae families, respectively. This finding reveals that the redistribution of genera within Firmicutes also contributes to its relationship with obesity. The Oscillospiraceae family, especially Oscillospiraceae UCG-002, has been associated with VA and IVA [43, 62, 63], while Subdoligranulum [64] and Coprococcus [65] with BA. This supports the hypothesis that TT pigs develop a gut environment that favours carbohydrate-fermenting rather than protein-fermenting microbiota. Treponema, which was also less abundant in TT pigs, was the only prevalent non-Firmicutes genus altered by LEPR. The genus Treponema includes a broad diversity of parasitic and commensal species. Previous findings in pigs indicate that Treponema can be linked to an increased ability to digest carbohydrates [66, 67] as well as proteins [56], as suggested by the data here. In line with this, there is evidence confirming that Treponema contributes to feed efficiency in pigs [67], although not all species exhibit this effect [68]. Eight occasional genera (not present in all pigs) were affected by LEPR. Of these, six were Firmicutes, including four Lachnospiraceae (NK3A20, UCG-009, XPB1014 and Frisingicoccus) and two Veillonaceae (Dialister and Megasphera), which together accounted for a very small share of the faecal microbiota (0.14%, in TT, and 0.28%, in C− pigs). Bacteria from these families help break down starch and other sugars, placing them among the producers of BA [69–71]. Only NK3A20, Dialister and Megasphera were positively related to BA, which could explain their higher prevalence in TT pigs. In contrast, the other three Lachnospiraceae genera (UCG-009, XPB1014 and Frisingicoccus), along with the dgA-11 group, the only Bacteroidota genera, and the archaeon Methanobrevibacter, were positively correlated with IVA. These five genera were more abundant in C− pigs, suggesting they may play a role in promoting protein degradation. Consistent with this, Methanobrevibacter has been found to be more abundant in the faeces of pigs with high feed efficiency [68]. Interestingly, the recessive nature of the T allele, previously demonstrated for fat-related traits [20] as well as for body weight and associated traits [22], is likewise evident in the microbiota composition (Table S2), thereby further supporting the role of the leptin receptor gene in modulating the abundance of these genera. A question worth exploring is whether the effect of LEPR on microbiota composition is merely a side-effect of feed intake. We examined this hypothesis by adjusting microbial abundance for feed intake. Although we could not monitor other feeding patterns, Zhang et al. [72] showed that feeding frequency does not influence the intestinal microbiota. Six of the LEPR-susceptible genera in TT pigs shifted due to increased feed intake, with some being enhanced (Subdoligranulum, Coprococcus, Lachnospiraceae NK3A20 and Dialister) and others inhibited (Oscillospiraceae UCG-002, Oscillospiraceae NK4A214 and Treponema). The correlation pattern between feed intake and VFA (+ 0.51, for BA, -0.48, for VA, -0.46, for IBA, and − 0.53, for IVA, p-value < 0.01), closely mirrors that of Subdoligranulum, Coprococcus, Oscillospiraceae UCG-002, Oscillospiraceae NK4A214 and Treponema with VFA (Fig. 6). The rise of Subdoligranulum and Coprococcus and the fall of Oscillospiraceae UCG-002, Oscillospiraceae NK4A214 and Treponema with feed intake highlights their role in balancing carbohydrate and protein fermentation. It also warns for potential misinterpretations in microbiota comparisons between treatments with different feed intake. For instance, it cannot be ruled out that feed intake is the underlying cause of the relationship between Subdoligranulum and fat mass [73]. Dalby et al. [13] already pointed out that diet and obesity can be confounded in relation to gut microbiota composition. We identified five genera in which LEPR had a direct effect on their relative abundance (Oscillospiraceae UCG-005, Lachnospiraceae UCG-009 and Lachnospiraceae XPB1014, Frisingicoccus, and Megasphera). These genera are the best candidates for delving deeper into the effects of LEPR on fatness, aside from increased feed intake. The probability of correctly assigning the TT genotype using these bacteria as markers was 78.7% and 89%, for pigs and pens, respectively. No pen effect was found for any of them, so we can infer that they were not affected by horizontal transmission after cohousing. Except Megasphera, which utilises lactate as substrate [74], the rest were less abundant in TT pigs and produced less BA and more AA than average microbiota (Fig. 6). Among them, Oscillospiraceae UCG-005 is the best positioned, as it is not only prevalent but also the genus with the greatest difference between LEPR genotypes. Other studies have also found that Oscillospiraceae UCG-005 is negatively associated with BA [75, 76] and, like some Lachnospiraceae, with leaner types [77, 78]. Since these genera also showed a positive association pattern with IVA, we can hypothesise that the switch from IVA to BA observed in TT pigs obey to a leptin-mediated signal shift that affected the abundance of these bacteria [79]. In this regard, evidence indicates that leptin receptor signalling directly regulates microbiota composition by upregulating the expression of genes encoding antimicrobial peptides in the intestinal epithelium [9], thereby potentially enhancing host defence against specific gut pathogens [80]. Rajala et al. [9] further demonstrated that this regulatory effect extends beyond the intestine and occurs independently of feed intake. Unfortunately, the functional roles of the LEPR-specific genera remain poorly characterized. Nonetheless, Oscillospiraceae UCG-005 has been associated with immune response [81] and healthy status [82], while Lachnospiraceae UCG-009 and Frisingicoccus have been linked to lower faecal water content [83, 84], with Frisingicoccus also associated with increased lipolysis [85]. Overall, this suggests that their reduced abundance in TT pigs may indicate increased susceptibility to diarrheal symptoms and systemic inflammation. On the other hand, gut microbiota modulates leptin sensitivity [86, 87] via gut-derived VFA. Besides serving as an energy source, VFA act as signalling molecules by binding to free-fatty acid receptors in the gut epithelium, which are involved in energy metabolism [14, 88]. Although the role of gut-derived VFA in leptin signalling remains unclear [89], there are reports linking BA to leptin expression [90] and AA and IVA to anorectic neuropeptide signalling [91, 92]. Bacteria linked to the LEPR genotype were moderately associated with fatness and body weight. To date, research in pigs investigating the relationship between microbiota and fatness has focussed on microbiota comparisons between fat and lean breeds [23, 25–30], or, at most, between individuals with extreme phenotypes within a breed [31]. It is documented that, in pigs, gut bacterial abundance exhibits heritable variation [54, 55], which can respond to selection [93]. However, aside from a few studies in humans [94], the effect of LEPR on microbiota composition has been investigated almost exclusively in rodents [4, 79]. Here, we present the first study in pigs connecting host genetics via LEPR with microbiota and fatness. In line with Yang et al. [7], who demonstrated that a deletion in the gene encoding N-acetyl-galactosaminyl-transferase reduces the abundance of Erysipelotrichaceae in a mosaic pig population, we found that the recessive T allele of the LEPR rs709596309 polymorphism lowers the levels of Oscillospiraceae, thus confirming the influence of host genetics on microbiota composition. Genera from this family have been proposed as potential sources of probiotics, owing to their association with low body fat content and various health benefits for humans [95]. Other genera also appear to play a role in the relationship of microbiota and fatness, but the results at the enterotype [55, 96], family or genus [31, 97] level are not consistent. In this regard, it is noteworthy that the correlated response to selection for an enterotype that enhanced post-weaning growth in at least eight of the LEPR-sensitive genera identified here, including Oscillospiraceae UCG-005, aligned with the trend observed for the TT genotype [93]. Consistent with our findings, this enterotype was enriched in functions related to starch degradation rather than protein degradation. Of the eight genera, Oscillospiraceae UCG-005 was the only one that did not covary with feed intake, exhibiting also one of the strongest negative responses. This highlights the importance of monitoring feed intake when interpreting microbiota results and supports the potential of Oscillospiraceae UCG-005 as a LEPR-specific and obesity microbial biomarker. Conclusions This is the first study in pigs to show that LEPR influences microbiota composition in faeces. Pigs carrying the TT genotype of rs709596309 (C > T) in LEPR display higher faecal BA and lower IVA concentrations, consistent with a greater abundance of starch-degrading over protein-degrading genera. We identified 14 candidate genera associated with this change, but only Oscillospiraceae UCG-005 and 4 occasional bacteria (3 Lachnospiraceae and Megasphaera) did not respond to feed intake. Candidate genera more abundant in TT pigs were positively associated with body weight and fatness, and vice versa. Our results highlight the impact of host genetics on microbiota composition, both through changes in feed intake and independently, as well as the potential of Oscillospiraceae UCG-005 as a LEPR-specific microbial marker. A low level of faecal Oscillospiraceae UCG-005 may be indicative of an increased risk of obesity. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1: Table S1. Composition and nutrients of the diets fed during the fattening period Supplementary Material 2: Table S2. Relative abundance of differential microbial genera across LEPR genotypes Supplementary Material 3: Figure S1. β-diversity dissimilarity in microbiota composition. Non-metric multidimensional scaling (NDMS) plots for microbiota composition based on pair-wise Bray–Curtis distances between (A) saliva and faeces and between LEPR genotypes in (B) saliva and (C) faeces. NMDS1 and NMDS2 refer to the first two axes from NMDS. Microbiota composition in faecal samples (in red) was significantly different from saliva samples (A, p-value<0.001). Microbiota composition in the TT genotype (in red) was significantly different from the C– genotype (in blue) in faeces (C, p-value<0.001) but not in saliva (B, p-value=0.61). p-values were determined by the PERMANOVA test via the adonis2 function from the vegan_2.6-2 package. Supplementary Material 4: Figure S2. β- diversity dispersion in saliva and faeces by LEPR genotype. The box plots show the median (centre line), interquartile range (box limits) for the average Bray-Curtis distance of each sample to the centroid of its genotype in (A) saliva and (B) faeces. Genotypes are indicated in red (TT) and blue (C–). Dispersion of β-diversity did not differ between genotypes either in saliva (A, p-value=0.45) or faeces (B, p-value=0.94). Supplementary Material 5: Figure S3. Abundance of microbiota phyla and families between LEPR genotypes in saliva. Each phylum is represented by a distinct colour. Bars represent the difference between the TT and C– genotypes, expressed as a percentage. No difference reached the Bonferroni-adjusted significance level. Supplementary Material 6: Figure S4. Heatmap correlation between genus abundance in faeces and growth and fatness. Pearson correlations of genus abundance (right side) with live body weight and backfat and loin thickness at 155 days and 202 days of age (4 days after faecal sampling) and with carcass weight, carcass backfat and loin thickness, and lean percentage (5 days after faecal sampling) (bottom side). Correlations were calculated on the logit-transformed values of traits expressed in percentage. Significant correlations (p-value<0.05) had a value higher (lower) than 0.20 (-0.20). Positive correlations are represented in blue, while negative correlations are represented in red. Supplementary Material 7: Figure S5. Relationship of genera affected by feed intake with fatness. Linear regression of backfat thickness at off-test (4 days after faecal sampling) on the logit-transformed relative bacterial abundance (X). The faecal abundance of the represented genera was influenced by the effect of LEPR on feed intake. First row: prevalent bacteria that were positively associated with fatness; second row: prevalent bacteria that were negatively associated with fatness; and third row: occasional bacteria that were not associated with backfat thickness (p-value > 0.05) in pigs harbouring these bacteria Genotypes are indicated in red (TT) and blue (C–). Acknowledgements We acknowledge Josep Reixach of Selección Batallé for his valuable support in the design and development of the experiment and sampling, and Pilar Sopeña of the Animal Breeding and Genetics Laboratory, Universitat de Lleida, for her assistance with laboratory analyses. Abbreviations Acetic acid Amplicon sequence variants Butyric acid Divisive amplicon denoising algorithm Daily feed intake Isobutyric acid Isovaleric acid Leptin receptor gene Propionic acid Standard deviation Standardized Ileal Digestible Valeric acid Volatile fatty acids Author contributions Conceptualization and design of the experiment: JE; Data and sample collection: RSM and JE; Laboratory analyses: RSM, ARS and RNP; Data curation and bioinformatics: RSM, HL, JE (INRAE) and RRF; Data analysis: RSM, HL and JE, Funding: JE and RRF; Writing original draft: RSM, HL and JE. All authors contributed to the interpretation of data for the manuscript and review and approved the final manuscript. Funding This research is part of the Grant PID2021-125689OB-I00, funded by MCIN/AEI/ 10.13039/501100011033 and by “ERDF A way of making Europe”, by the “European Union. Houda Laghouaouta was funded by the AGAUR Agency of the Catalan Government for a 3-month visit at the INRAE GABI laboratory (Jouy-en-Josas, France) under the supervision of Jordi Estellé. Data availability The sequencing data for this study are available from NCBI Bioproject accession number PRJNA1258869 (https://www.ncbi.nlm.nih.gov/bioproject/PRJNA1258869). The data sets generated and/or analysed during this study are included in the published article and in the Supplementary information files, and are available at the Research Data Repository of the Catalan Open Research Area (CORA-RDR) (10.34810/data2031). Declarations Ethics approval and consent to participate Pigs used in the study were raised and slaughtered following applicable regulations and good practice guidelines on the protection of animals kept for farming purposes, during transport and slaughter. In accordance with European Directive 2010/63/EU and Spanish Royal Decree 53/2013, no invasive procedures or treatments were performed on the animals in this study. The experimental protocol was reviewed by the Ethical Committee on Animal Experimentation of the University of Lleida (CEEA 04–06/21). Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Supplementary Materials Supplementary Material 1: Table S1. Composition and nutrients of the diets fed during the fattening period Supplementary Material 2: Table S2. Relative abundance of differential microbial genera across LEPR genotypes Supplementary Material 3: Figure S1. β-diversity dissimilarity in microbiota composition. Non-metric multidimensional scaling (NDMS) plots for microbiota composition based on pair-wise Bray–Curtis distances between (A) saliva and faeces and between LEPR genotypes in (B) saliva and (C) faeces. NMDS1 and NMDS2 refer to the first two axes from NMDS. Microbiota composition in faecal samples (in red) was significantly different from saliva samples (A, p-value<0.001). Microbiota composition in the TT genotype (in red) was significantly different from the C– genotype (in blue) in faeces (C, p-value<0.001) but not in saliva (B, p-value=0.61). p-values were determined by the PERMANOVA test via the adonis2 function from the vegan_2.6-2 package. Supplementary Material 4: Figure S2. β- diversity dispersion in saliva and faeces by LEPR genotype. The box plots show the median (centre line), interquartile range (box limits) for the average Bray-Curtis distance of each sample to the centroid of its genotype in (A) saliva and (B) faeces. Genotypes are indicated in red (TT) and blue (C–). Dispersion of β-diversity did not differ between genotypes either in saliva (A, p-value=0.45) or faeces (B, p-value=0.94). Supplementary Material 5: Figure S3. Abundance of microbiota phyla and families between LEPR genotypes in saliva. Each phylum is represented by a distinct colour. Bars represent the difference between the TT and C– genotypes, expressed as a percentage. No difference reached the Bonferroni-adjusted significance level. Supplementary Material 6: Figure S4. Heatmap correlation between genus abundance in faeces and growth and fatness. Pearson correlations of genus abundance (right side) with live body weight and backfat and loin thickness at 155 days and 202 days of age (4 days after faecal sampling) and with carcass weight, carcass backfat and loin thickness, and lean percentage (5 days after faecal sampling) (bottom side). Correlations were calculated on the logit-transformed values of traits expressed in percentage. Significant correlations (p-value<0.05) had a value higher (lower) than 0.20 (-0.20). Positive correlations are represented in blue, while negative correlations are represented in red. Supplementary Material 7: Figure S5. Relationship of genera affected by feed intake with fatness. Linear regression of backfat thickness at off-test (4 days after faecal sampling) on the logit-transformed relative bacterial abundance (X). The faecal abundance of the represented genera was influenced by the effect of LEPR on feed intake. First row: prevalent bacteria that were positively associated with fatness; second row: prevalent bacteria that were negatively associated with fatness; and third row: occasional bacteria that were not associated with backfat thickness (p-value > 0.05) in pigs harbouring these bacteria Genotypes are indicated in red (TT) and blue (C–). Data Availability Statement The sequencing data for this study are available from NCBI Bioproject accession number PRJNA1258869 (https://www.ncbi.nlm.nih.gov/bioproject/PRJNA1258869). The data sets generated and/or analysed during this study are included in the published article and in the Supplementary information files, and are available at the Research Data Repository of the Catalan Open Research Area (CORA-RDR) (10.34810/data2031).