{"paper_id":"19cfa6e2-a3eb-44b8-9939-617ed9013683","body_text":"Leveraging Joint-Specific Phenotypes for Genome-Wide Association Studies and SNP Heritability Estimation of Equine Osteochondrosis and Fetlock Osteochondral Fragments | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Leveraging Joint-Specific Phenotypes for Genome-Wide Association Studies and SNP Heritability Estimation of Equine Osteochondrosis and Fetlock Osteochondral Fragments Bram Van Mol, Steven Janssens, Roel Meyermans, Léa Chapard, Maarten Oosterlinck, and 2 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8832466/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 8 You are reading this latest preprint version Abstract Background Equine osteochondrosis (OC) and fetlock osteochondral fragments are highly prevalent developmental joint disorders caused by failure of the growing skeleton, often leading to intra-articular fragments that require arthroscopic surgery and impose substantial economic and animal welfare burdens. To reduce prevalence, phenotype-based selection has had limited success, prompting interest in genomic approaches. Progress in genetic research has been hindered by difficulties in assembling large study populations, inconsistent phenotype definitions, and the disease’s dynamic nature. This study aimed to perform joint-specific case–control genome‐wide association studies (GWAS) and estimate single nucleotide polymorphism (SNP) based heritabilities for OC and fetlock osteochondral fragments in horses. Seven base phenotypes with two to four levels of strictness were defined using a novel semi-quantitative scoring system and paired with high-density genotypes. Results Across 23 GWAS on 716 horses, 214 suggestive SNPs were identified, including 50 that reached genome-wide significance. Based on predefined criteria, eight genomic regions of interest were identified, revealing 11 novel candidate genes: LDB2 for total OC; SCUBE3 , MAPK13 and MAPK14 for hock OC; and FGF6 , FGF23 , IL17A , IL17F , TRAM2 , BMP5 , and PPP2R5C for palmaro-/plantaroproximal osteochondral fragments of the proximal phalanx (POF). No candidate genes were detected for distal intermediate ridge of the tibia (DIRT) OC. For stifle OC, fetlock OC, and dorsoproximal fragments of the proximal phalanx (DOF), suggestive SNPs did not define genomic regions of interest. SNP-based heritability estimates ranged from 26.2–27.1% for total OC; 53.9–61.0% for hock OC; 49.4–61.2% for DIRT OC; 1.5–14.2% for stifle OC; 0.0-13.7% for fetlock OC; 10.0-54.8% for POF; and 5.7–39.6% for DOF. Conclusion This study represents the first GWAS on DOF in horses and identifies novel candidate genes involved in angiogenesis and bone development for total OC, hock OC and POF. The results demonstrate the importance of stringent, joint-specific phenotypes, as even minor differences in phenotype definition substantially affected heritability estimates and the detection of genomic associations and candidate genes. The high heritability of hock and DIRT OC indicates considerable potential for rapid genetic improvement, compared to the lower heritability observed for total OC, fetlock OC, and stifle OC. Equine Osteochondrosis Fetlock osteochondral fragments horse Genome-wide association study SNP-based heritability Joint-specific phenotypes Figures Figure 1 Figure 2 Background Osteochondrosis (OC) and fetlock osteochondral fragments are common osteoarticular disorders in horses, characterized by failure of the growing skeleton. OC is identified radiographically by abnormal bone contours, with or without adjacent radiopaque fragments, at specific predilection sites. Fetlock osteochondral fragments, particularly dorsoproximal osteochondral fragments (DOF) and palmaro-/plantaroproximal osteochondral fragments (POF) of the proximal phalanx, are identified by radiopaque fragments at their respective sites. The etiopathogenesis of OC, DOF, and POF differs and involves both environmental factors [ 1 ], such as the prenatal environment, biomechanical trauma, growth, nutrition, and weaning, and genetic factors [ 2 ]. OC and fetlock osteochondral fragments increase the risk of joint effusion, reduced performance, and eventually lameness [ 3 – 6 ]. The high prevalence rates of these disorders [ 7 – 12 ], combined with significant economic and welfare costs due to preventive and curative surgeries, has prompted studbooks to implement strict phenotype-based selection criteria [ 2 , 13 ]. However, these efforts have been undermined by the dynamic nature of lesion development and healing, surgical fragment removal without radiological evidence, and environmental factors masking true genetic predisposition [ 2 ]. These interfering factors may cause inaccurate breeding decisions, hindering prevalence reduction. Consequently, breeding organizations have invested in the detection of single nucleotide polymorphisms (SNPs) and quantitative trait loci (QTL) associated with OC and fetlock osteochondral fragments, as well as in the development of genomic breeding values for these disorders. Across breeds, heritability estimates for certain types of osteochondral lesions can be as high as 0.52, indicating substantial potential for genetic improvement through selective breeding [ 14 , 15 ]. However, heritability varies significantly among lesion types, with some showing very low estimates, suggesting that environmental factors may have a stronger influence than genetic ones. Consequently, targeted management practices are likely to be more effective in reducing lesion prevalence than selection alone. To support effective breeding decisions and guide selection strategies, accurate joint-specific heritability estimates for specific osteochondral disorder (sub)types are needed [ 2 , 16 ]. A major challenge in the transition towards using genomic information is the accurate identification of SNPs, QTLs, and genes associated with susceptibility to OC, DOF, and POF. Genetic analyses across ten equine breeds have identified QTLs associated with OC and POF on 27 autosomes and one sex chromosome, leading to the identification of 71 candidate genes [ 2 ]. In contrast, no QTLs for DOF had been reported prior to this study. The large number and variability of identified QTLs and candidate genes can be attributed, at least in part, to the lack of uniform phenotyping protocols and phenotype definitions across studies. In addition, fragmented data across breeds and limited sample sizes continue to hinder progress. To address these challenges and provide news insights into the genetics underlying OC and fetlock osteochondral fragments, we conducted joint-specific case–control genome‐wide association studies (GWAS) and estimated SNP-based heritabilities using data from 716 horses. We leveraged joint-specific case-control phenotypes derived from a newly developed semi-quantitative scoring system paired with 670K SNP-chip genotypes. Seven base phenotype definitions (total OC, hock OC, distal intermediate ridge of the tibia OC, stifle OC, fetlock OC, DOF and POF) across two to four different strictness levels were used to enhance the specificity of the genetic analyses. Importantly, this represents the first published GWAS of DOF in horses. Our objectives were to: 1) identify significantly associated SNPs; 2) identify novel candidate genes; 3) estimate SNP‐based heritability; and 4) evaluate if and how GWAS results and heritability estimates vary depending on the phenotypes studied. Methods Animals Blood samples, radiographic data and pedigree data were collected from 763 horses admitted at the Faculty of Veterinary Medicine Ghent University in Belgium for radiographic examination (stallion inspection or pre-purchase examination) or arthroscopy (osteochondral fragment removal) from 2012 to 2023. The sample included 469 males and 293 females registered across 12 Warmblood studbooks, with one additional male being an English Thoroughbred. The majority of horses (84.27%) were registered in Belgian Studbooks: Belgian Warmblood Horse, Zangersheide and Belgian Sport Horse (Additional file 1: Supplementary Table S1). At the time of radiographic screening, the horses had a minimum age of 216 days and a maximum age of 6132 days, with a mean age of 1100.6 days (standard deviation (SD) = 663.1, interquartile range (IQR) 775.5 - 1182.0). Radiography Horses admitted for radiographic examination were sedated with 0.1 ml/100 kg detomidine (Detogesic ® ) and 0.1 ml/100 kg butorphanol (Torbugesic ® ) administered intravenously. The hooves were trimmed and cleaned, and the sulci of the frog were packed with modelling compound (Play-Doh ® ). The radiographic protocol included at least the following projections. Lateromedial, dorso(55°)proximal-palmarodistal oblique and dorso(65°)proximal-palmarodistal oblique projections of both front feet. Lateromedial projection of all four fetlock joints. Dorsoplantar, lateromedial, dorso(45°)lateral-plantaromedial oblique and plantaro(45°)lateral-dorsomedial oblique projections of both hocks. Caudal(60°)lateral-dorsomedial oblique projection of both stifles. Complementary projections were performed if a radiographic abnormality was suspected on one of these basic projections or if the client preferred a more extensive radiographic examination. For horses admitted for arthroscopy, the radiographic examination performed by the referring veterinarian was evaluated. If the radiographic screening was incomplete or if there was doubt about radiographic abnormalities on their projections, supplementary radiographic projections were performed. Phenotyping As part of the equine hospital routine, all the radiographic examinations were evaluated by at least one specialist and one resident of the European College of Veterinary Diagnostic Imaging. Subsequently, all radiographic images and accompanying reports were reevaluated and scored by the first author. To classify radiographic findings indicative of osteochondrosis, a new five-grade scoring system was developed. This system scored the bone contours of specific anatomical locations as follows: 0 - normal, 1 - flattened, 2 - irregular, 3 - notch, and 4 - notch with adjacent osteochondral fragment(s). The common predilection sites of OC were evaluated using this system, including the lateral trochlear ridge of the femur, the distal intermediate ridge of the tibia, the lateral trochlear ridge of the talus, the medial malleolus of the tibia, and the dorsal part of the sagittal ridge of the third metacarpal/metatarsal bone [7, 17, 18]. The most dorsoproximal part of the sagittal ridge of the third metacarpal/metatarsal bone was not taken into account during scoring because there is a lot of radiographic variation (irregularity, indentation, radiolucency, notch) in this region [19] which does not histologically resemble OC lesions [20]. Additionally, uncommon locations of OC such as the medial trochlear ridge of the femur, medial trochlear ridge of the talus and the articular surface of the patella using the same system were scored [7]. Fetlock osteochondral fragments were scored as either 0 (absent) or 1 (present) and were subdivided into three categories: DOF, POF, and other osteochondral fragments in the fetlock joint, such as ununited palmar/plantar eminences, synovial pad fragments, and fractures of the proximal sesamoid bones. Genotyping All 763 horses were genotyped using the Axiom Equine Genotyping Array (670,796 SNPs) in three different batches: 184 were genotyped at SEGALAB (Laboratório de Sanidade Animal e Segurança Alimentar SA), 359 at CeGen-ISCIII (Spanish National Centre of Genotyping), and 220 at the Array and Analysis facility Uppsala University. For 396 horses, SNP positions were remapped from the EquCab 2.0 reference genome to EquCab 3.0 using ThermoFisher Scientific’s axiom analysis suite software. PLINK 1.9 [21, 22] and R [23] were used to merge the genotypes and to identify the 535,330 common SNP positions consisting of those remapped from EquCab 2.0 and those originally mapped to EquCab 3.0. Quality control Quality control was performed using PLINK 1.9 [21, 22] following a protocol based on Anderson et al. [24]. The per-individual quality control resulted in the removal of 2 horses with discordant sex information and the removal of 1 horse with outlying heterozygosity rate. Additionally, a parentage test was performed in which the genomic relatedness was compared with the pedigree relatedness resulting in the removal of 44 horses. The principal component analysis could not identify large differences in ancestry between breeds (Additional file 2: Supplementary Figure S1), therefore the remaining 716 horses were considered as one population. The per-marker quality control excluded: (1) 22,070 SNPs located on sex chromosomes, (2) 212 SNPs with a call rate <95%, (3) 3,817 SNPs with significant deviation from the Hardy–Weinberg equilibrium (P <0.0001) and (4) 74,920 SNPs with low minor allele frequency (<1%) resulting in 434,311 remaining SNPs. The remaining SNPs were pruned for linkage disequilibrium at a squared correlation coefficient (R 2 ) level of 0.5 thereby retaining 177,029 SNPs. Analysis For the case/control genome-wide association analysis we tested two statistical approaches: a linear regression model and a mixed linear model that incorporated a sparse genetic relationship matrix as a random effect to account for population structure and relatedness. Despite the theoretical advantages of the mixed linear model, its implementation had minimal impact on the results. Consequently, we selected the linear regression model as the final analytical approach due to its computational efficiency and comparable performance. Sex was included as a fixed effect, and to control for residual population structure, the first six principal components of the genotypic data were additionally included as fixed effects. These six principal components, determined by a scree test, accounted for 46.9% of the total genetic variation. Suggestive signals were identified using a threshold of -log10(1/177,029) (P < 5.6×10 -6 ) [25], while genome-wide significant signals were identified using the Bonferroni-corrected threshold of -log10(0.05/177,029) (P < 2.8×10 −7 ) [26, 27]. GWA analysis and visualisation were performed in R using genome-wide complex trait analysis (GCTA) [28, 29] and the qqman package [30]. To define genomic regions of interest, we applied the following criteria. GWAS case–control groups were required to include at least 40 cases and more than 40 controls. In addition, at least three distinct SNPs within a 3 Mb region had to reach at least suggestive significance across the different phenotype-definition strictness levels. Genes within each identified genomic region, including an additional 0.5 Mb range upstream and downstream, were identified using Ensemble genome browser BioMart tool [31, 32]. Case and control selection was conducted on a joint-specific basis, using progressively stricter phenotype definitions to increase specificity (A to D). For total OC, the basis definition (A) for cases was the presence of a bone contour notch, with or without adjacent osteochondral fragment(s) (score ≥ 3), at the common predilection site(s) of OC of the hock and/or stifle and/or fetlock. Controls were defined by the presence of a normal bone contour (score 0) at the same site(s). The additional strictness level (B) required both cases and controls to have no (score 0) osteochondral fragments in the fetlock joint and a normal bone contour (score 0) at uncommon OC locations. For OC of the hock, stifle, and fetlock separately, the basis definition (A) for cases was the presence of a bone contour notch, with or without adjacent osteochondral fragment(s) (score ≥ 3), at the common predilection site(s) of OC within the respective joint. Controls were defined by the presence of a normal bone contour (score 0) at the same site(s). The additional strictness levels, applied cumulatively and required for both cases and controls, were: (B) a normal bone contour (score 0) at other common OC predilection sites; (C) the absence of osteochondral fragments (score 0) in the fetlock joint and a normal bone contour (score 0) at uncommon OC locations. For OC of the distal intermediate ridge of the tibia (DIRT OC), which is the most common location of hock OC [17], a specific GWAS was performed. The basis definition (A) required cases to have a bone contour notch, with or without adjacent osteochondral fragments (score ≥ 3), at the distal intermediate ridge. Controls were required to have a normal bone contour (score 0) at the same site. The additional strictness levels, applied cumulatively and to both cases and controls, were: (B) a normal bone contour (score 0) at the lateral trochlear ridge of the talus and the medial malleolus of the tibia (other common predilection sites of hock OC); (C) a normal bone contour (score 0) at other common OC predilection sites; (D) the absence of osteochondral fragments (score 0) in the fetlock joint and a normal bone contour (score 0) at uncommon OC locations. For both POF and DOF, the basis definition (A) classified cases by the presence of the respective condition in at least one fetlock joint, while controls were defined by its absence in all fetlock joints. The additional strictness levels, applied cumulatively and required for both cases and controls, were: (B) the absence of the other condition (POF or DOF) and other osteochondral fragments in the fetlock joint; (C) a normal bone contour (score 0) at the common location for fetlock OC; and (D) a normal bone contour (score 0) at other common and uncommon OC locations. SNP-based heritability for the different phenotype definitions was estimated using the restricted maximum likelihood (REML) approach in GCTA [28]. A genomic relationship matrix (GRM) was constructed from autosomal SNPs, and case-control analyses were performed under progressively stricter phenotype definitions to evaluate their impact on heritability estimates. The same covariates from the GWAS analysis (sex and the first six principal components) were included. Heritability was calculated based on the genetic and residual variance estimated from the model. Results Manhattan plots for the GWAS of total OC, hock OC, DIRT OC and POF are presented in Figure 1, with a complete overview of Manhattan and and Quantile-Quantile plots for all phenotypes provided in Additional file 3: Supplementary Figures S2-S15. Across the different phenotype definitions, we identified 214 SNPs with suggestive significance, 50 of which reached genome-wide significance. Genome-wide significant association signals were identified for total OC (score ≥ 3 at the fetlock and/or hock and/or stifle OC) on Equus caballus chromosome (ECA) 3; for hock OC (score ≥ 3 at the distal intermediate ridge of the tibia and/or lateral trochlear ridge of the talus and/or medial malleolus of the tibia) on ECA 20 and 24; for DIRT OC (score ≥ 3 at the distal intermediate ridge of the tibia) on ECA 17 and 29; for stifle OC (score ≥ 3 at the lateral trochlear ridge of the femur) on ECA 30; for fetlock OC (score ≥ 3 at the dorsal part of the sagittal ridge of the third metacarpal/metatarsal bone) on ECA 17, 22 and 27; for POF (score = 1 for POF in at least one of the fetlock joints) on ECA 1, 2, 4, 5, 6, 7, 10, 14, 15, 18, 19, 20, 23, 24 and 25; and for DOF (score = 1 for DOF in at least one of the fetlock joints) on ECA 10, 15 and 18. All SNPs with suggestive and genome-wide significance and their corresponding significance levels for each phenotype are listed in Additional file 4: Supplementary Table S2. Table 1 provides an overview of the genomic regions that met our criteria, listing the gene that contains the SNP with the lowest P-value as well as other candidate genes within each region. For each region, we reported the number of SNPs reaching suggestive and genome-wide significance, along with the lowest observed P-value. Across the identified genomic regions, including an additional 0.5 Mb range upstream and downstream, we detected 201 protein coding genes (Additional file 5: Supplementary Table S3). To nominate candidate genes, we annotated the gene ontology and identified the following genes associated with metabolic pathways involved in ossification and angiogenesis: LDB2 , FGF6 , FGF23 , SCUBE3 , MAPK13 , MAPK14 , IL17A , IL17F , TRAM2, BMP5 and PPP2R5C . Table 1. Genomic regions and candidate genes associated with osteochondrosis phenotypes in horses based on a case/control GWAS using a linear regression model. Regions are defined as genomic intervals in which at least three SNPs within a 3 Mb window show suggestive significance across the varying phenotype definition strictness levels. Trait refers to the phenotype definition (see supplementary figure referenced in parentheses for details on phenotype definitions and number of horses). ‘N SNPs’ indicates the number of SNPs with suggestive significance (P < 5.6×10 -6 ), with the number reaching genome-wide significance (P < 2.8×10 −7 ) in parentheses. The column “–log₁₀( P )” shows the smallest P -value among SNPs in the region. Candidate genes include those located within the identified region plus 0.5 Mb upstream and downstream. ECA Region (Mb) Trait N SNP -log 10 (P) Candidate Genes 3 108.28-108.50 T (S2) 3 (1) 6.77 LDB2 (LIM domain binding 2) 6 32.94-33.92 POF (S12) 7 (1) 6.62 FGF23 (fibroblast growth factor 23); FGF6 (fibroblast growth factor 6) 7 5.80-8.03 POF (S12) 3 (1) 6.70 20 35.58-38.87 H (S4) 4 (2) 7.28 SCUCBE3 (signal peptide, CUB domain and EGF like domain containing 3); MAPK13 (mitogen-activated protein kinase 13); MAPK14 (mitogen-activated protein kinase 14) 20 50.18-60.96 POF (S12) 16 (7) 8.52 IL17A (interleukin 17A); IL17F (interleukin 17F); TRAM2 (translocation associated membrane protein 2); BMP5 (bone morphogeneticprotein 5) 24 43.66-44.61 POF (S12) 5 (3) 13.26 PPP2R5C (protein phosphatase 2 regulatory subunit B'gamma) 24 43.66-45.03 H (S4) 3 (2) 6.89 29 14.03-16.56 DIRT (S6) 3 (1) 7.24 DOF, Dorsoproximal osteochondral fragments of the proximal phalanx; ECA, Equus caballus chromosome; F, Fetlock osteochondrosis; H, Hock osteochondrosis; DIRT, Osteochondrosis at the distal intermediate ridge of the tibia; Mb, Megabase; POF, Palmaro-/ plantaroproximal osteochondral fragments of the proximal phalanx;; S, Stifle osteochondrosis; SNP, Single nucleotide polymorphism; T, Fetlock and/or hock and/or stifle osteochondrosis. SNP-based heritability estimates, adjusted for sex and population structure, are presented in Additional file 6: Supplementary Table S4 and illustrated in Figure 2. For total OC, heritability was moderate, ranging from 26.2% (standard error (SE) = 11.9%) at strictness level A to 27.1% (SE = 16.8%) at level B. Higher heritability estimates were observed for hock OC and DIRT OC across strictness levels, with values ranging from 53.9% (SE = 11.2%) to 61.0% (SE = 16.2%) and 49.4% (SE = 11.6%) to 61.2% (SE = 22.4%), respectively. In contrast, stifle and fetlock OC had low heritability estimates, ranging from 1.5% (SE = 23.7%) to 14.2% (SE = 10.2%) and 0.0% (SE = 24.8%) to 13.7% (SE = 11.7%), respectively. Heritability estimates for POF and DOF showed a wide range depending on the phenotype definition. For POF, the heritability estimates ranged from 10.0% (SE = 22.7%) at strictness level D to 54.8% (SE = 14.0%) at strictness level B, whereas for DOF they ranged from 5.7% (SE = 10.2%) at strictness level A to 39.6% (SE = 27.8%) at strictness level D. Discussion In this study, we present joint-specific case–control GWAS and SNP heritability estimation for OC and fetlock osteochondral fragments in a cohort of 716 horses. The population is almost entirely composed of Warmblood horses (99.9%), with only one English Thoroughbred included, and 84.27% are registered in Belgian studbooks. Therefore, the findings are most relevant for Warmblood horse populations, particularly those registered in Belgium. To improve the specificity of the genetic analyses, we applied progressively refined phenotype definitions based on a newly developed semi-quantitative scoring system paired with high-density 670K SNP-chip genotypes. Increasing phenotype definitions strictness improved the specificity of case and control classifications, thereby reducing potential confounding. However, this refinement came at the cost of a reduced sample size, which in turn lowered statistical power of the GWAS. The smaller sample size introduced greater variability and potential unreliability, particularly under the most restrictive phenotype definitions. The reduced statistical power was reflected in the Manhattan and QQ plots (Additional file 3: Supplementary Figures S2 -S15), which showed signs of genomic inflation, false-positive signals, and spurious associations under the strictest phenotype definitions. To mitigate these challenges, we applied the predefined criteria for identifying genomic regions of interest, balancing the advantages of improved phenotype specificity against the limitations of reduced sample size. Across seven base phenotype definitions (total OC, hock OC, distal intermediate ridge of the tibia OC, stifle OC, fetlock OC, DOF and POF), each evaluated at two to four strictness levels, we identified 214 genome-wide suggestive SNPs, 50 of which reached genome-wide significance and 8 genomic regions of interest that met our predefined criteria. From these regions, we nominated 11 candidate genes based on gene ontology annotations. To compare our findings with previous research, we used the review by Van Mol et al. [ 2 ], which provides an overview of the genomic coordinates and phenotype definitions of QTLs previously reported for equine OC and osteochondral fetlock fragments. This allowed us to evaluate whether the identified genomic regions in our study overlapped with previously reported QTLs and whether any of the nominated candidate genes had been described in prior literature. Given that our analyses were based on the EquCab 3.0 reference genome, while most prior studies used EquCab 2.0, we remapped our genomic coordinates to EquCab 2.0 using the UCSC LiftOver tool [ 33 ] to enable direct comparison. Since the review did not include gene expression studies, we additionally screened the literature for genes with a significantly altered expression in early OC lesions [ 34 – 41 ]. We focused on early OC lesions, as gene expression changes observed in later stages may reflect secondary degenerative responses rather than primary pathogenic processes. Based on this comparison, all 11 candidate genes identified in our study appear to be novel. This result is remarkable but not entirely surprising considering the differences in study design. Previous studies generally relied on broader phenotype definitions, whereas the present study employed more stringent, joint-specific phenotypes. This may have revealed loci that remained undetected when lesions are grouped. Furthermore, advances in functional annotation may have facilitated the identification of genes that were previously overlooked. The inability to confirm previously reported candidate genes emphasizes the complex genetic architecture underlying these conditions and demonstrates the need for larger studies based on standardized joint-specific phenotypes to establish reproducible associations. In the following sections, we describe the identified genomic regions and compare them with regions previously reported in association with equine OC or fetlock osteochondral fragments. We discuss the biological plausibility of the newly nominated candidate genes in relation to the pathogenesis of these conditions. Finally, we examine how GWAS results and heritability estimates varied depending on the phenotype definitions and assess the implications of these findings for future genetic research. Total osteochondrosis In association with the total OC phenotype (Fig. 1 A), we identified a genomic region on ECA 3 spanning 108.28–108.50 Mb (corresponding to 106.45-106.67 Mb in EquCab 2.0). Although this exact region has not been reported in prior studies, Drabbe et al. [ 42 ] identified a nearby QTL at 115.91-115.92 Mb (EquCab 3.0) in Belgian Warmbloods, using the same phenotype definition. Additionally, Naccache et al. [ 14 ] identified another nearby QTL at 105.40-105.94 Mb (EquCab 2.0) for total osteochondrosis dissecans (OCD) in Hanoverian Warmbloods, although, their phenotype definition differed from ours, as they included only horses with both an abnormal bone contour and an adjacent fragment as cases and considered two additional predilection sites. Notably, the genome-wide significant SNP on ECA 3 in our initial analysis (Fig. 1 A) disappeared when stricter criteria were applied to both cases and controls (Fig. 1 B). On ECA 3 the SNP with the lowest P-value associated with total OC is located within LIM domain binding 2 ( LDB2 ), which has an important role in mediating transcription through the formation of higher-order transcription complexes and is expressed in endothelial cells. One of the genes regulated by LDB2 is delta-like ligand 4 ( DLL4 ), a gene important for sprouting angiogenesis and vascular remodeling [ 43 ]. Given that the initial stage of OC involves the focal failure of epiphyseal growth cartilage canal vessels, and that the final outcome depends on whether this area of ischemic chondronecrosis can heal [ 44 – 55 ], LDB2 represents a promising candidate gene for further investigation. No other candidate genes were detected in the associated genomic region on ECA 3. Hock osteochondrosis For hock OC, our analysis (Fig. 1 D) revealed two novel genomic regions of interest: one on ECA 20 (35.58–38.87 Mb) and another on ECA 24 (43.66–45.03 Mb0). To our knowledge, this is the first study in which SNPs on ECA 20 and ECA 24 have been associated with any form of OC in horses. Notably, one of the two SNP with suggestive significance on ECA 3 in Fig. 1 C, located at 107.64 Mb (corresponding to 105.82 Mb in EquCab 2.0) falls within a previously reported QTL for hock OC spanning 100.39-107.92 Mb (EquCab 2.0) [ 56 , 57 ]. However, this signal was no longer evident under the stricter phenotype definition applied in Fig. 1 D. Importantly, neither of these studies used the same phenotype definition as in our study [ 56 , 57 ]. Orr et al. [ 56 ] included only horses with OCD (abnormal bone contour + adjacent fragment) as cases and considered four additional predilection sites, while Teyssedre et al. [ 57 ] did not specify the precise anatomical locations affected by osteochondrosis in the hock. Interestingly, Lykkjen et al. [ 58 ], who used a phenotype definition similar to that in Fig. 1 C, identified a significant SNP on ECA 3 at 113.50 Mb (EquCab 2.0). In contrast to these and other GWAS that have identified QTLs for hock OC [ 14 , 58 , 59 ], our study is the first to use a phenotype definition that includes only the most common predilection sites of hock OC and excludes animals with concurrent stifle and/or fetlock OC (Fig. 1 D), thereby reducing potential confounding. The most significantly associated SNPs with hock OC were located on ECA 20 and ECA 24, within the ribosomal protein S10 ( RPS10 ) and CDC42 binding protein kinase beta ( CDC42BPB ) genes, respectively. Neither gene appear to have a direct connection with osteochondrosis, and no candidate genes were identified within the associated region on ECA 24. However, within the associated genomic region of ECA 20, several candidate genes were identified, such as signal peptide, CUB domain and EGF like domain containing 3 ( SCUBE3) and mitogen-activated protein kinase 13 and 14 ( MAPK13 and MAPK14 ). SCUBE3 encodes a signal peptide that promotes bone morphogenesis, playing an important role in bone morphology and metabolism [ 60 ]. GWAS have identified SNPs of SCUBE3 among height associated loci in both pigs [ 61 ] and humans [ 62 ]. SCUBE3 knockout mice have viable phenotypes but show impaired BMP-mediated chondrogenesis and osteogenesis, which, among other defects, leads to abnormal endochondral bone development. Moreover, mutations in SCUBE3 have been linked to various human skeletal disorders [ 60 , 63 ]. MAPK13 and MAPK14, members of the p38 MAPK family, are important for cellular stress responses, tissue development, and homeostasis. The MAPK pathway regulates important processes in skeletogenesis and bone maintenance, including chondrocyte and osteoblast differentiation, extracellular matrix deposition, and mineralization. MAPK13, with selective tissue expression, modulates inflammatory responses and insulin secretion [ 64 ]. Disruption of the insulin balance has been associated with increased occurrence of OC [ 65 – 69 ]. Conversely, MAPK14, which is ubiquitously expressed, influences a broader range of processes, including inflammatory responses, core cellular processes, and the Wnt signaling pathway [ 64 ]. Wnt signaling is altered in OC lesions and may be associated with disease pathogenesis [ 38 , 70 , 71 ]. MAPK14 predominates in growth-plate chondrocytes, particularly in the pre-hypertrophic zone, where it regulates hypertrophic chondrocyte differentiation. MAPK14 is also highly expressed and essential for osteoclastogenesis, and deletions of MAPK14 have been associated with skeletal abnormalities in mice [ 64 , 72 ]. The MAPK pathway mediates the effects of multiple growth and transcription factors influencing bone cell function and differentiation [ 64 ] of which some are found to be differently expressed in cartilage of OC lesions: TGF-β [ 34 , 73 ], SOX9 [ 40 ] and RUNX2 [ 36 ]. In our opinion, the overlap between the aforementioned processes regulated by the MAPK pathway and their links to OC, identifies MAPK13 and MAPK14 as compelling candidate genes for further research. Distal intermediate ridge of the tibia osteochondrosis To enhance signal detection, we refined the phenotype of hock OC even further by focusing exclusively on OC of the distal intermediate ridge of the tibia (Additional file 3: Supplementary Figure S6 ). This revealed a region on ECA 29 (Fig. 1 E, F) spanning 14.03–16.56 Mb (corresponding to 12.96–15.49 Mb in EquCab 2.0), not detected in the broader hock OC phenotype definition (Fig. 1 C, D). Naccache et al. [ 14 ] reported a nearby QTL for hock OC (16.07–16.94 Mb) and total OC (16.74–16.94 Mb in EquCab 2.0), though without specifying the exact anatomical sites affected by OC. Notably, the SNPs with the lowest P-value on ECA 1 at 34.59 Mb (Fig. 1 E) and ECA 29 at 14.03 Mb (Fig. 1 F) were also the lowest P-value SNPs on their respective chromosomes in the hock OC GWAS (Fig. 1 C, D), but at lower significance levels. While the SNP on ECA 1 did not meet the established criteria for a significant signal, the SNP on ECA 29 did, achieving at least suggestive significance in all four phenotype definitions, with its highest significance reaching -log 10 (P) = 7.24 (Fig. 1 F). This SNP lies in a noncoding region, with no candidate genes in the associated genomic region. Palmaroproximal and plantaroproximal osteochondral fragments The GWAS for POF (Fig. 1 G, H, I) identified four genomic regions of interest on: ECA 6 (32.94–33.92 Mb), ECA 7 (5.80–8.03 Mb), ECA 20 (50.18–60.96 Mb), and ECA 24 (43.66–44.61 Mb). To our knowledge, this is the first study associating SNPs on ECA 6, 20 and 24 with POF [ 2 ]. On ECA 7, Lykkjen et al. [ 74 ] reported a QTL for POF at 69.9–80.5 Mb (EquCab 2.0), which does not overlap with the region identified in this study (remapped to 5.33–7.55 Mb EquCab 2.0). The most significant SNP associated with POF on ECA 7 lies at 8.03 Mb, within a noncoding region, with no candidate genes in the associated genomic region. On ECA 6, the most significant associated SNP also lies in a noncoding region, with fibroblast growth factors 6 and 23 ( FGF6 and FGF23 ) identified as candidate genes within the associated genomic region. FGFs are spatiotemporally expressed during skeletal development, regulating cellular functions and extracellular matrix proteins metabolism [ 75 – 77 ]. FGF23 regulates bone mineralization, parathyroid hormone secretion, and phosphate/vitamin D balance. Its dysregulation is associated with skeletal diseases characterized by defective bone mineralization, such as hypophosphatemic rickets and hyperphosphatemic calcinosis [ 76 – 80 ]. FGF6 contributes to muscle regeneration and bone remodeling [ 76 , 81 ]. On ECA 20, the most significant associated SNP is located within KHDRBS2 (KH RNA binding domain containing, signal transduction associated 2), which has no evident link to osteochondral disorders. Within the associated genomic region, interleukin 17A ( IL17A ), interleukin 17F ( IL17F ), translocation associated membrane protein 2 (TRAM2) , and bone morphogenetic protein 5 ( BMP5 ) were identified as candidate genes. L17A and IL17F encode cytokines involved in bone repair and cartilage matrix turnover and have been implicated in the pathogenesis of spondylarthritis [ 82 – 84 ]. TRAM2 encodes a component of the endoplasmic reticulum translocon, a protein channel that regulates collagen I translocation and calcium transport, thereby supporting bone formation [ 85 – 87 ]. BMP5 encodes a bone morphogenetic protein important for bone formation, growth, remodeling, and repair and has been linked to various genetic skeletal disorders, including the short-ear mouse phenotype [ 88 – 90 ]. On ECA 24, the QTL associated with POF (43.66–44.61 Mb) overlapped with the QTL associated with hock OC (43.66–45.03 Mb). Within this QTL four closely located SNPs at 43.66, 44.08, 44.24, and 44.45 MB reached genome wide significance, with their association with POF increasing as the phenotype definition became more stringent (Additional file 4: Supplementary Table S2 ). The most significant associated SNP (− log10(P) = 13.26) at 44.08 MB lies within protein phosphatase 2 regulatory subunit B'gamma ( PPP2R5C ), which has been implicated in overgrowth syndromes in humans [ 91 ]. No other candidate genes were found in the associated genomic region. Similar to the other candidate genes ( FGF6, FGF23, IL17A, IL17F, TRAM2 , and BMP5) , PPP2R5C can be linked to bone development and remodelling processes. Given that POF is proposed to result from weakened bone and pathological avulsion fractures [ 46 , 92 ], all these genes represent plausible candidate genes contributing to its pathogenesis. Stifle osteochondrosis, fetlock osteochondrosis, and dorsoproximal osteochondral fragments In the GWAS for stifle OC (Additional file 3: Supplementary Figure S8), fetlock OC (Additional file 3: Supplementary Figure S10), and DOF (Additional file 3: Supplementary Figure S14), none of the SNPs met the established criteria to be classified as a significant signal. To our knowledge, this is the first published GWAS conducted for DOF. The absence of significant associations likely reflects the complex, polygenic nature of these traits, characterized by numerous small-effect loci that require larger sample sizes for detection. For DOF, as with POF, fragment removal leaving no radiological trace introduces the risk of false-negative diagnoses, adding to the challenge [ 2 ]. Furthermore, as reviewed by Van Mol et al. [ 1 ], osteochondral disorders are influenced by various environmental factors, which may overshadow genetic predisposition for some phenotypes and mask genetic associations if not adequately accounted for. Single nucleotide polymorphism based heritability estimates The higher heritability estimates for hock OC and DIRT OC compared to other OC types are consistent with findings from previous studies [ 57 , 93 , 94 ]. A limitation of our approach is that SEs increase as the phenotype definition become stricter (levels A -> D), due to reduced numbers of cases and controls. In the literature, SEs ranged between 0.11–0.15 for OC at different anatomical locations [ 57 , 94 ], which is comparable to our SEs at strictness level A across all phenotype definitions ranging between 0.10–0.12. At stricter levels (B - D), SEs ranges between 0.12–0.28, reducing the reliability of these estimates, and highlighting the need for larger genotype datasets to achieve more precise estimates. The wide variability observed in the heritability of DOF and POF may also reflect potential misclassification of cases as controls, as previously discussed. The higher heritability of hock and DIRT OC suggests that selective breeding programs could more effectively reduce their incidence, making these lesions potentially higher-priority targets for breeding programs than total OC, stifle OC, or fetlock OC. However, caution is warranted given the still relatively high SEs of the estimates. In contrast, the low heritability estimates for stifle and fetlock OC indicate that additive genetic variance explains a smaller proportion of phenotypic variance, implying that environmental factors and/or genotype-environment interactions may play a more substantial role in disease risk. This finding aligns with the absence of significant GWAS signals for stifle and fetlock OC. Implications and future research The semi-quantitative scoring system presented in this study could serve as a standardized phenotyping protocol for radiographic screenings of horses presented for studbook admission. Consistent international implementation across studbooks would generate high-quality, comparable phenotypes across populations, which are essential for reliable genetic analyses and, ultimately, genetic selection. Pooling data across studbooks to form multi-population reference datasets will improve the cost–benefit ratio of investing in genomic evaluations for OC and fetlock osteochondral fragments. Such datasets would also provide a solid foundation for integrating genomic information into breeding programs An advantage of GWAS is its SNP-based framework, which enables the estimation of SNP-based heritabilities. Incorporating SNP genotyping into selection protocols would further expand datasets and, by reducing SEs, improve the precision of heritability estimates. These estimates will help identifying which conditions are meaningful targets for selection. Notably, horses affected by stifle and/or fetlock OC, which according to the current study show relatively low heritability, are currently excluded from breeding by studbooks. If larger datasets in future research confirm these low heritabilities, this would suggest that excluding horses based on such lesions results in a considerable loss of genetic diversity while offering only limited benefit in reducing disease prevalence. A practical implementation would proceed in several steps. First, harmonize and apply the scoring system across participating studbooks. Second, collect standardized phenotypes and representative genotypes to establish a multi-population reference panel. Third, estimate SNP-based heritabilities and genomic breeding values. Fourth, after appropriate validation and cost–benefit analyses, integrate genomic information into breeding protocols. In parallel, future research should prioritize targeted follow-up of the nominated loci. Functional validation of the candidate genes, both from this study and future studies, will be essential to confirm causal relationships. Ultimately, harmonized phenotyping, larger multi-population datasets, and integrated genomic approaches will result in evidence-based breeding strategies that reduce disease prevalence while preserving genetic diversity. Conclusion Across 23 case–control GWAS of equine osteochondrosis and fetlock osteochondral fragments in a population of 716 horses, we identified 214 suggestive SNPs, 50 of which reached genome-wide significance. Using predefined criteria, 8 genomic regions of interest were identified, leading to the discovery of 11 novel candidate genes. For total OC, a region on ECA 3 was identified containing the candidate gene LDB2 , which is a regulator of angiogenesis. For hock OC, novel loci on ECA 20 and ECA 24 were identified harbouring candidate genes SCUBE3 , MAPK13 , and MAPK14 , which are important for skeletogenesis. Further refinement to OC at the distal intermediate ridge revealed a genomic region of interest on ECA 29. For POF, a total of four genomic regions on ECA 6, 7, 20 and 24 were identified, pointing to FGF6 , FGF23 , IL17A , IL17F , TRAM2 , BMP5 , and PPP2R5C as plausible candidate genes due to their link with bone development and remodelling. In contrast, for stifle OC, fetlock OC, and DOF, none of the SNPs met the established significance criteria, reflecting the likely complex polygenic nature and possible overshadowing environmental factors for these traits. Importantly, this study represents the first published GWAS conducted for DOF. Notably, previously reported candidate genes could not be reaffirmed in this study, potentially due to differences in study populations and/or phenotype definitions. SNP-based heritability estimates varied considerably among phenotypes, with moderate estimates for total OC (26.2–27.1%), high estimates for hock and DIRT OC (49.4–61.2%), low estimates for stifle and fetlock OC (0.0-14.2%), and highly variable estimates for POF and DOF (5.7–54.8%). Our findings demonstrate the importance of stringent lesion-specific phenotypes, as even slight differences in phenotype definition strictness within the same population can substantially affect the detection of genomic associations, identification of candidate genes and estimation of heritability. Declarations Ethics approval and consent to participate All protocols were approved by the Ethical Committee of the Faculty of Veterinary Medicine, Ghent University, Belgium (EC 2021/060). Informed consent was obtained from the horse owners prior to the inclusion of their horses in the study. Consent for publication Not applicable. Availability of data and materials The data that support the findings of this study are available from the Department of Large Animal Surgery, Anaesthesia and Orthopaedics, Faculty of Veterinary Medicine, Ghent University, and the Center for Animal Breeding and Genetics, Department of Biosystems, KU Leuven, but restrictions apply to the availability of these data, which were used under license for the current study and are therefore not publicly available. Data are, however, available from the authors upon reasonable request and with permission from the Department of Large Animal Surgery, Anaesthesia and Orthopaedics, Faculty of Veterinary Medicine, Ghent University, and the Center for Animal Breeding and Genetics, Department of Biosystems, KU Leuven. Competing interests The authors declare that they have no competing interests. Funding Bram Van Mol received a PhD fellowship in fundamental research from The Research Foundation – Flanders (FWO-FR 11B3921N). The genotyping was funded by Paardenpunt Vlaanderen as part of the ‘PaardenGenomica 2020’ project. Authors' contributions BVM collected, analysed, and interpreted the data and wrote the manuscript. SJ, RM, and LC contributed to the study design and critically reviewed the statistical analyses. SJ, RM, LC and NB critically reviewed the manuscript. All authors read and approved the final manuscript. Acknowledgements The authors would like to thank the Belgian Warmblood (BWP) studbook, Paardenpunt Vlaanderen and the Belgian Agency for Agriculture and Fisheries for their cooperation. References Van Mol B, Oosterlinck M, Janssens S, Buys N, Pille F. Environmental factors of equine osteochondrosis and fetlock osteochondral fragments: A scoping review – Part 1. Vet J. 2024;308:106249. doi: https://doi.org/10.1016/j.tvjl.2024.106249. Van Mol B, Janssens S, Oosterlinck M, Pille F, Buys N. Genetic factors of equine osteochondrosis and fetlock osteochondral fragments: A scoping review - Part 2. Vet J. 2024;308:106258. doi: https://doi.org/10.1016/j.tvjl.2024.106258. Declercq J, Martens A, Maes D, Boussauw B, Forsyth R, Boening KJ. Dorsoproximal proximal phalanx osteochondral fragmentation in 117 Warmblood horses. Vet Comp Orthop Traumatol. 2009;22(1):1-6. doi: 10.3415/vcot-08-02-0016. Brink P, Dolvik NI, Tverdal A. Lameness and effusion of the tarsocrural joints after arthroscopy of osteochondritis dissecans in horses. Vet Rec. 2009;165(24):709-12. van Weeren PR. Chapter 89 - Osteochondritis Dissecans. In: Auer JA, Stick JA, Kümmerle JM, Prange T, editors. Equine Surgery. 5th ed. Philadelphia, PA: Elsevier Saunders; 2019. p. 1509-28. Verwilghen DR, Janssens S, Busoni V, Pille F, Johnston C, Serteyn D. Do developmental orthopaedic disorders influence future jumping performances in Warmblood stallions? Equine Vet J. 2013;45(5):578-81. doi: 10.1111/evj.12027. van Grevenhof EM, Ducro BJ, van Weeren PR, van Tartwijk J, van den Belt AJ, Bijma P. Prevalence of various radiographic manifestations of osteochondrosis and their correlations between and within joints in Dutch Warmblood horses. Equine Vet J. 2009;41(1):11-6. doi: 10.2746/042516408x334794. Arnan P, Hertsch B. Röntgenologische Untersuchung zur Erfassung der Osteochondrosis dissecans in Fessel, Sprung-und Kniegelenk in Vergleich von Fohlen zum Zweijährigen. Proceedings Göttinger Pferdetage. 2004;4:115-24. Stock KF, Hamann H, Distl O. Prevalence of osseous fragments in distal and proximal interphalangeal, metacarpo- and metatarsophalangeal and tarsocrural joints of Hanoverian Warmblood horses. J Vet Med A Physiol Pathol Clin Med. 2005;52(8):388-94. doi: 10.1111/j.1439-0442.2005.00753.x. Vos NJ. Incidence of osteochondrosis (dissecans) in Dutch warmblood horses presented for pre-purchase examination. Irish Veterinary Journal. 2008;61(1):33. doi: 10.1186/2046-0481-61-1-33. Hilla D, Distl O. Heritabilities and genetic correlations between fetlock, hock and stifle osteochondrosis and fetlock osteochondral fragments in Hanoverian Warmblood horses. Journal of Animal Breeding and Genetics. 2014;131(1):71-81. doi: 10.1111/jbg.12062. Jonsson L, Dalin G, Egenvall A, Nasholm A, Roepstorff L, Philipsson J. Equine hospital data as a source for study of prevalence and heritability of osteochondrosis and palmar/plantar osseous fragments of Swedish Warmblood horses. Equine Vet J. 2011;43(6):695-700. doi: 10.1111/j.2042-3306.2010.00354.x. Ortved KF. Surgical Management of Osteochondrosis in Foals. Vet Clin N Am, Equine Pract. 2017;33(2):379-96. doi: https://doi.org/10.1016/j.cveq.2017.03.010. Naccache F, Metzger J, Distl O. Genetic risk factors for osteochondrosis in various horse breeds. Equine Vet J. 2018;50(5):556-63. doi: 10.1111/evj.12824. McCoy AM, Norton EM, Kemper AM, Beeson SK, Mickelson JR, McCue ME. SNP-based heritability and genetic architecture of tarsal osteochondrosis in North American Standardbred horses. Anim Genet. 2019;50(1):78-81. doi: 10.1111/age.12738. Russell J, Matika O, Russell T, Reardon RJM. Heritability and prevalence of selected osteochondrosis lesions in yearling Thoroughbred horses. Equine Vet J. 2017;49(3):282-7. doi: https://doi.org/10.1111/evj.12613. McIlwraith CW, Foerner JJ, Davis DM. Osteochondritis dissecans of the tarsocrural joint: results of treatment with arthroscopic surgery. Equine Vet J. 1991;23(3):155-62. doi: 10.1111/j.2042-3306.1991.tb02746.x. McIlwraith CW. Inferences from referred clinical cases of osteochondritis dissecans. Equine Vet J. 1993;25(S16):27-30. doi: https://doi.org/10.1111/j.2042-3306.1993.tb04850.x. Hauspie S, Martens A, Declercq J, Busoni V, Vanderperren K, van Bree H, et al. Radiographic features of the dorsal condylar sagittal ridge of the third metacarpal and metatarsal bones in young Warmblood stallions. Veterinary and Comparative Orthopaedics and Traumatology. 2010;23(6):411-6. doi: 10.3415/vcot-09-12-0130. Hauspie S, Forsyth R, Vanderperren K, Declercq J, Martens A, Saunders JH. The histological appearance of the proximal aspect of the dorsal condylar sagittal ridge of the third metacarpal and metatarsal bone in young warmblood horses: normal appearance and correlation with detected radiographic variations. Anat Histol Embryol. 2013;42(3):232-8. doi: 10.1111/ahe.12006. Chang CC, Chow CC, Tellier LC, Vattikuti S, Purcell SM, Lee JJ. Second-generation PLINK: rising to the challenge of larger and richer datasets. Gigascience. 2015;4:7. doi: 10.1186/s13742-015-0047-8. Purcell S, Neale B, Todd-Brown K, Thomas L, Ferreira MA, Bender D, et al. PLINK: a tool set for whole-genome association and population-based linkage analyses. American journal of human genetics. 2007;81(3):559-75. doi: 10.1086/519795. R Core Team R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria. 2023. Anderson CA, Pettersson FH, Clarke GM, Cardon LR, Morris AP, Zondervan KT. Data quality control in genetic case-control association studies. Nature Protocols. 2010;5(9):1564-73. doi: 10.1038/nprot.2010.116. Lander E, Kruglyak L. Genetic dissection of complex traits: guidelines for interpreting and reporting linkage results. Nature Genetics. 1995;11(3):241-7. doi: 10.1038/ng1195-241. Dudbridge F, Gusnanto A. Estimation of significance thresholds for genomewide association scans. Genetic Epidemiology. 2008;32(3):227-34. doi: https://doi.org/10.1002/gepi.20297. Li MX, Yeung JM, Cherny SS, Sham PC. Evaluating the effective numbers of independent tests and significant p-value thresholds in commercial genotyping arrays and public imputation reference datasets. Hum Genet. 2012;131(5):747-56. doi: 10.1007/s00439-011-1118-2. Yang J, Lee SH, Goddard ME, Visscher PM. GCTA: a tool for genome-wide complex trait analysis. The American Journal of Human Genetics. 2011;88(1):76-82. Jiang L, Zheng Z, Qi T, Kemper KE, Wray NR, Visscher PM, et al. A resource-efficient tool for mixed model association analysis of large-scale data. Nat Genet. 2019;51(12):1749-55. doi: 10.1038/s41588-019-0530-8. Turner S. qqman: an R package for visualizing GWAS results using Q-Q and manhattan plots. Journal of Open Source Software. 2018;3:731. doi: 10.21105/joss.00731. Harrison PW, Amode MR, Austine-Orimoloye O, Azov Andrey G, Barba M, Barnes I, et al. Ensembl 2024. Nucleic Acids Research. 2023;52(D1):D891-D9. doi: 10.1093/nar/gkad1049. Kinsella RJ, Kähäri A, Haider S, Zamora J, Proctor G, Spudich G, et al. Ensembl BioMarts: a hub for data retrieval across taxonomic space. Database (Oxford). 2011;2011:bar030. doi: 10.1093/database/bar030. Hinrichs AS, Karolchik D, Baertsch R, Barber GP, Bejerano G, Clawson H, et al. The UCSC Genome Browser Database: update 2006. Nucleic Acids Research. 2006;34(suppl_1):D590-D8. doi: 10.1093/nar/gkj144. Henson FM, Schofield PN, Jeffcott LB. Expression of transforming growth factor-beta 1 in normal and dyschondroplastic articular growth cartilage of the young horse. Equine Vet J. 1997;29(6):434-9. doi: 10.1111/j.2042-3306.1997.tb03155.x. Semevolos SA, Strassheim ML, Haupt JL, Nixon AJ. Expression patterns of hedgehog signaling peptides in naturally acquired equine osteochondrosis. J Orthop Res. 2005;23(5):1152-9. doi: 10.1016/j.orthres.2005.01.024. Mirams M, Tatarczuch L, Ahmed YA, Pagel CN, Jeffcott LB, Davies HMS, et al. Altered Gene Expression in Early Osteochondrosis Lesions. Journal of Orthopaedic Research. 2009;27(4):452-7. doi: 10.1002/jor.20761. Riddick TL, Duesterdieck-Zellmer K, Semevolos SA. Gene and protein expression of cartilage canal and osteochondral junction chondrocytes and full-thickness cartilage in early equine osteochondrosis. Vet J. 2012;194(3):319-25. doi: https://doi.org/10.1016/j.tvjl.2012.04.023. Kinsley MA, Semevolos SA, Duesterdieck-Zellmer KF. Wnt/-catenin signaling of cartilage canal and osteochondral junction chondrocytes and full thickness cartilage in early equine osteochondrosis. Journal of Orthopaedic Research. 2015;33(10):1433-8. doi: 10.1002/jor.22846. Mirams M, Ayodele BA, Tatarczuch L, Henson FM, Pagel CN, Mackie EJ. Identification of novel osteochondrosis Associated genes. Journal of Orthopaedic Research. 2016;34(3):404-11. doi: 10.1002/jor.23033. Grissom SK, Semevolos SA, Duesterdieck-Zellmer K. Role of cartilage and bone matrix regulation in early equine osteochondrosis. Bone Reports. 2023;18. doi: 10.1016/j.bonr.2023.101653. Austbo L, Roed KH, Dolvik NI, Skretting G. Identification of differentially expressed genes associated with osteochondrosis in standardbred horses using rna arbitrarily primed PCR. Animal Biotechnology. 2010;21(2):135-9. doi: 10.1080/10495391003608316. Drabbe A, Janssens S, Blott S, Ducro BJ, Fontanel M, Francois L, et al. Genome-Wide Association Analyses of Osteochondrosis in Belgian Warmbloods Reveal Candidate Genes Associated With Chondrocyte Development. Journal of Equine Veterinary Science. 2022;111. doi: 10.1016/j.jevs.2022.103870. Choi HJ, Rho SS, Choi DH, Kwon YG. LDB2 regulates the expression of DLL4 through the formation of oligomeric complexes in endothelial cells. BMB Rep. 2018;51(1):21-6. doi: 10.5483/bmbrep.2018.51.1.140. Carlson CS, Cullins LD, Meuten DJ. Osteochondrosis of the articular-epiphyseal cartilage complex in young horses: evidence for a defect in cartilage canal blood supply. Vet Pathol. 1995;32(6):641-7. doi: 10.1177/030098589503200605. Jeffcott L. Developmental diseases affecting growing horses. 2nd European Workshop on Equine Nutrition. ENESAD, Dijon, FRANCE2004. p. 243-55. Olstad K, Ekman S, Carlson CS. An Update on the Pathogenesis of Osteochondrosis. Vet Pathol. 2015;52(5):785-802. doi: 10.1177/0300985815588778. Ytrehus B, Carlson CS, Ekman S. Etiology and Pathogenesis of Osteochondrosis. Vet Pathol. 2007;44(4):429-48. doi: 10.1354/vp.44-4-429. Olstad K, Hendrickson EHS, Carlson CS, Ekman S, Dolvik NI. Transection of vessels in epiphyseal cartilage canals leads to osteochondrosis and osteochondrosis dissecans in the femoro-patellar joint of foals; a potential model of juvenile osteochondritis dissecans. Osteoarthritis and Cartilage. 2013;21(5):730-8. doi: 10.1016/j.joca.2013.02.005. Dik KJ, Enzerink E, van Weeren PR. Radiographic development of osteochondral abnormalities, in the hock and stifle of Dutch Warmblood foals, from age 1 to 11 months. Equine Vet J Suppl. 1999(31):9-15. doi: 10.1111/j.2042-3306.1999.tb05308.x. Santschi EM, Prichard MA, Whitman JL, Batten CA, Strathman TA, Canada NC, et al. Stifle radiography in Thoroughbreds from 6 to 18 months of age. Equine Vet Educ. 2020;32:78-84. doi: 10.1111/eve.13194. GrØndahl AM. The incidence of osteochondrosis in the tibiotarsal joint of norwegian standardbred trotters. Journal of Equine Veterinary Science. 1991;11(5):272-4. doi: https://doi.org/10.1016/S0737-0806(06)81313-4. Carlsten J, Sandgren B, Dalin G. Development of osteochondrosis in the tarsocrural joint and osteochondral fragments in the fetlock joints of Standardbred trotters. I. A radiological survey. Equine Vet J. 1993;25(S16):42-7. doi: https://doi.org/10.1111/j.2042-3306.1993.tb04853.x. Baccarin RYA, Pereira MA, Roncati NV, Bergamaschi RRC, Hagen SCF. Development of osteochondrosis in Lusitano foals: A radiographic study. Canadian Veterinary Journal-Revue Veterinaire Canadienne. 2012;53(10):1079-84. Jacquet S, Robert C, Valette JP, Denoix JM. Evolution of radiological findings detected in the limbs of 321 young horses between the ages of 6 and 18months. Vet J. 2013;197(1):58-64. doi: https://doi.org/10.1016/j.tvjl.2013.03.042. Stromberg B. A review of the salient features of osteochondrosis in the horse. Equine Vet J. 1979;11(4):211-4. doi: 10.1111/j.2042-3306.1979.tb01346.x. Orr N, Hill EW, Gu J, Govindarajan P, Conroy J, van Grevenhof EM, et al. Genome-wide association study of osteochondrosis in the tarsocrural joint of Dutch Warmblood horses identifies susceptibility loci on chromosomes 3 and 10. Anim Genet. 2013;44(4):408-12. doi: 10.1111/age.12016. Teyssedre S, Dupuis MC, Guerin G, Schibler L, Denoix JM, Elsen JM, et al. Genome-wide association studies for osteochondrosis in French Trotter horses. Journal of Animal Science. 2012;90(1):45-53. doi: 10.2527/jas.2011-4031. Lykkjen S, Dolvik NI, McCue ME, Rendahl AK, Mickelson JR, Roed KH. Genome-wide association analysis of osteochondrosis of the tibiotarsal joint in Norwegian Standardbred trotters. Anim Genet. 2010;41:111-20. doi: 10.1111/j.1365-2052.2010.02117.x. McCoy AM, Beeson SK, Splan RK, Lykkjen S, Ralston SL, Mickelson JR, et al. Identification and validation of risk loci for osteochondrosis in standardbreds. Bmc Genomics. 2016;17. doi: 10.1186/s12864-016-2385-z. Lin Y-C, Sahoo BK, Gau S-S, Yang R-B. The biology of SCUBE. Journal of Biomedical Science. 2023;30(1):33. doi: 10.1186/s12929-023-00925-3. Wang L, Zhang L, Yan H, Liu X, Li N, Liang J, et al. Genome-wide association studies identify the loci for 5 exterior traits in a Large White × Minzhu pig population. PLoS One. 2014;9(8):e103766. doi: 10.1371/journal.pone.0103766. Gudbjartsson DF, Walters GB, Thorleifsson G, Stefansson H, Halldorsson BV, Zusmanovich P, et al. Many sequence variants affecting diversity of adult human height. Nat Genet. 2008;40(5):609-15. doi: 10.1038/ng.122. Lin Y-C, Niceta M, Muto V, Vona B, Pagnamenta AT, Maroofian R, et al. SCUBE3 loss-of-function causes a recognizable recessive developmental disorder due to defective bone morphogenetic protein signaling. The American Journal of Human Genetics. 2021;108(1):115-33. doi: https://doi.org/10.1016/j.ajhg.2020.11.015. Thouverey C, Caverzasio J. Focus on the p38 MAPK signaling pathway in bone development and maintenance. Bonekey Rep. 2015;4:711. doi: 10.1038/bonekey.2015.80. Ralston SL. Hyperglycemia/hyperinsulinemia after feeding a meal of grain to young horses with osteochondritis dissecans (OCD) lesions. Pferdeheilkunde Equine Medicine. 1996;12:320-2. Pagan J, Geor R, Caddel S, Pryor P, Hoekstra K. The relationship between glycemic response and the incidence of OCD in Thoroughbred weanlings: a field study. Proc Am Ass equine Practnrs. 2001;47. Harris P, Staniar W, Ellis AD. Effect of exercise and diet on the incidence of DOD. The Growing Horse: Nutrition and Prevention of Growth Disorders Wageningen Academic Publishers, Wageningen, The Netherlands 2nd European Conference on Equine Nutrition Dijon, France2004. p. 273-90. Staniar WB. Linking dietary energy and skeletal development in the horse. Revista Brasileira de Zootecnia. 2010;39:138-44. Henson FM, Davenport C, Butler L, Moran I, Shingleton WD, Jeffcott LB, et al. Effects of insulin and insulin-like growth factors I and II on the growth of equine fetal and neonatal chondrocytes. Equine Vet J. 1997;29(6):441-7. doi: https://doi.org/10.1111/j.2042-3306.1997.tb03156.x. Serteyn D, Mendoza García L, Sandersen C, Caudron I, piquemal D, Chavatte-Palmer P, et al. High plasma concentrations of sclerostin, an inhibitor of the Wnt signaling pathway in young horses affected by osteochondrosis. Open journal of orthepedics. 2014. doi: 10.4236/ojo.2014.412051. Marchant EA, Semevolos SA. Differential Protein Expression of the Marginal Transitional Zone in Foals with Osteochondrosis. Journal of Equine Veterinary Science. 2022;116:104055. doi: https://doi.org/10.1016/j.jevs.2022.104055. Han J, Wu J, Silke J. An overview of mammalian p38 mitogen-activated protein kinases, central regulators of cell stress and receptor signaling. F1000Res. 2020;9. doi: 10.12688/f1000research.22092.1. Semevolos SA, Nixon AJ, Brower-Toland BD. Changes in molecular expression of aggrecan and collagen types I, II, and X, insulin-like growth factor-I, and transforming growth factor-β1 in articular cartilage obtained from horses with naturally acquired osteochondrosis. Am J Vet Res. 2001;62(7):1088-94. doi: 10.2460/ajvr.2001.62.1088. Lykkjen S, Dolvik NI, McCue ME, Rendahl AK, Mickelson JR, Roed KH. Equine developmental orthopaedic diseases - a genome-wide association study of first phalanx plantar osteochondral fragments in Standardbred trotters. Anim Genet. 2013;44(6):766-9. doi: 10.1111/age.12064. Fayad C, Legeai-Mallet L. FGF Signaling: A Key Pathway During Skeletal Development. In: Rossi A, Zaucke F, editors. The Extracellular Matrix in Genetic Skeletal Disorders. Cham: Springer Nature Switzerland; 2024. p. 247-85. Xie Y, Su N, Yang J, Tan Q, Huang S, Jin M, et al. FGF/FGFR signaling in health and disease. Signal Transduction and Targeted Therapy. 2020;5(1):181. doi: 10.1038/s41392-020-00222-7. Ornitz DM, Marie PJ. Fibroblast growth factor signaling in skeletal development and disease. Genes Dev. 2015;29(14):1463-86. doi: 10.1101/gad.266551.115. Bär L, Stournaras C, Lang F, Föller M. Regulation of fibroblast growth factor 23 (FGF23) in health and disease. FEBS Letters. 2019;593(15):1879-900. doi: https://doi.org/10.1002/1873-3468.13494. Su N, Jin M, Chen L. Role of FGF/FGFR signaling in skeletal development and homeostasis: learning from mouse models. Bone Research. 2014;2(1):14003. doi: 10.1038/boneres.2014.3. Chonchol M, Greene T, Zhang Y, Hoofnagle AN, Cheung AK. Low Vitamin D and High Fibroblast Growth Factor 23 Serum Levels Associate with Infectious and Cardiac Deaths in the HEMO Study. J Am Soc Nephrol. 2016;27(1):227-37. doi: 10.1681/asn.2014101009. Bosetti M, Leigheb M, Brooks RA, Boccafoschi F, Cannas MF. Regulation of osteoblast and osteoclast functions by FGF-6. Journal of Cellular Physiology. 2010;225(2):466-71. doi: https://doi.org/10.1002/jcp.22225. Huangfu L, Li R, Huang Y, Wang S. The IL-17 family in diseases: from bench to bedside. Signal Transduction and Targeted Therapy. 2023;8(1):402. doi: 10.1038/s41392-023-01620-3. Tanigawa S, Aida Y, Kawato T, Honda K, Nakayama G, Motohashi M, et al. Interleukin-17F affects cartilage matrix turnover by increasing the expression of collagenases and stromelysin-1 and by decreasing the expression of their inhibitors and extracellular matrix components in chondrocytes. Cytokine. 2011;56(2):376-86. doi: https://doi.org/10.1016/j.cyto.2011.08.015. Jo S, Wang SE, Lee YL, Kang S, Lee B, Han J, et al. IL-17A induces osteoblast differentiation by activating JAK2/STAT3 in ankylosing spondylitis. Arthritis Research & Therapy. 2018;20(1):115. doi: 10.1186/s13075-018-1582-3. Pregizer S, Barski A, Gersbach CA, García AJ, Frenkel B. Identification of novel Runx2 targets in osteoblasts: Cell type-specific BMP-dependent regulation of Tram2. Journal of Cellular Biochemistry. 2007;102(6):1458-71. doi: https://doi.org/10.1002/jcb.21366. Zhang J, Ji Y, Jiang S, Shi M, Cai W, Miron RJ, et al. Calcium–Collagen Coupling is Vital for Biomineralization Schedule. Advanced Science. 2021;8(15):2100363. doi: https://doi.org/10.1002/advs.202100363. Johnson AE, van Waes MA. The Translocon: A Dynamic Gateway at the ER Membrane. Annual Review of Cell and Developmental Biology. 1999;15(Volume 15, 1999):799-842. doi: https://doi.org/10.1146/annurev.cellbio.15.1.799. King JA, Marker PC, Seung KJ, Kingsley DM. BMP5 and the Molecular, Skeletal, and Soft-Tissue Alterations in short ear Mice. Developmental Biology. 1994;166(1):112-22. doi: https://doi.org/10.1006/dbio.1994.1300. Storm EE, Kingsley DM. Joint patterning defects caused by single and double mutations in members of the bone morphogenetic protein (BMP) family. Development. 1996;122(12):3969-79. doi: 10.1242/dev.122.12.3969. Salazar VS, Gamer LW, Rosen V. BMP signalling in skeletal development, disease and repair. Nature Reviews Endocrinology. 2016;12(4):203-21. doi: 10.1038/nrendo.2016.12. Loveday C, Tatton-Brown K, Clarke M, Westwood I, Renwick A, Ramsay E, et al. Mutations in the PP2A regulatory subunit B family genes PPP2R5B, PPP2R5C and PPP2R5D cause human overgrowth. Human Molecular Genetics. 2015;24(17):4775-9. doi: 10.1093/hmg/ddv182. McCoy AM, Toth F, Dolvik NI, Ekman S, Ellermann J, Olstad K, et al. Articular osteochondrosis: a comparison of naturally-occurring human and animal disease. Osteoarthritis and Cartilage. 2013;21(11):1638-47. doi: https://doi.org/10.1016/j.joca.2013.08.011. van Grevenhof EM, Schurink A, Ducro BJ, van Weeren PR, van Tartwijk J, Bijma P, et al. Genetic variables of various manifestations of osteochondrosis and their correlations between and within joints in Dutch warmblood horses. Journal of Animal Science. 2009;87(6):1906-12. doi: 10.2527/jas.2008-1199. Zimmermann E, Distl O. SNP-Based Heritability of Osteochondrosis Dissecans in Hanoverian Warmblood Horses. Animals. 2023;13(9). doi: 10.3390/ani13091462. Additional Declarations No competing interests reported. Supplementary Files AdditionalFile1SupplementaryTableS1.docx AdditionalFile6SupplementaryTableS4.docx AdditionalFile2SupplementaryFigureS1.docx AdditionalFile5SupplementaryTableS3.docx AdditionalFile4SupplementaryTableS2.docx AdditionalFile3SupplementaryFiguresS2S15.docx Cite Share Download PDF Status: Under Review Version 1 posted Reviews received at journal 19 Apr, 2026 Reviewers agreed at journal 23 Mar, 2026 Reviewers agreed at journal 18 Mar, 2026 Reviewers invited by journal 16 Mar, 2026 Editor assigned by journal 23 Feb, 2026 Editor invited by journal 20 Feb, 2026 Submission checks completed at journal 19 Feb, 2026 First submitted to journal 19 Feb, 2026 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {\"props\":{\"pageProps\":{\"initialData\":{\"identity\":\"rs-8832466\",\"acceptedTermsAndConditions\":true,\"allowDirectSubmit\":false,\"archivedVersions\":[],\"articleType\":\"Research Article\",\"associatedPublications\":[],\"authors\":[{\"id\":608196441,\"identity\":\"209352e1-d980-4668-9095-e8bd225b3398\",\"order_by\":0,\"name\":\"Bram Van Mol\",\"email\":\"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA4klEQVRIiWNgGAWjYBACAwSTufEBhMFDtBbGZgOIBhK0tEkQpcVc+vDTDYxtdnm67Y1t1Ty/bPLsGXgPPsCnxbIvzewGY1tysdmZg223efvSinkY+JIN8GkxOMNgdoPhDHPithuJQC09hxN7GHjMJPBrYf8G1FIP1lLM2/MfpMX8B34tPEBbKg6DtTDz/DgAtgWfDgbLHp6yGwkVxxO3nTnYLDm3ITmx5zBfMl6HmfOwb7vxwaA6cdvx5oMf3vyxS2xv7z34Aa81IJAAYzC2AQlmgupRwB/SlI+CUTAKRsHIAAChuk11Sb9UDQAAAABJRU5ErkJggg==\",\"orcid\":\"\",\"institution\":\"Ghent University\",\"correspondingAuthor\":true,\"prefix\":\"\",\"firstName\":\"Bram\",\"middleName\":\"Van\",\"lastName\":\"Mol\",\"suffix\":\"\"},{\"id\":608196442,\"identity\":\"97e9810d-0b50-41d6-b9b9-8f9602261807\",\"order_by\":1,\"name\":\"Steven Janssens\",\"email\":\"\",\"orcid\":\"\",\"institution\":\"KU Leuven\",\"correspondingAuthor\":false,\"prefix\":\"\",\"firstName\":\"Steven\",\"middleName\":\"\",\"lastName\":\"Janssens\",\"suffix\":\"\"},{\"id\":608196443,\"identity\":\"6ed69b9d-c6af-4afa-a9da-7134ec2e7bae\",\"order_by\":2,\"name\":\"Roel Meyermans\",\"email\":\"\",\"orcid\":\"\",\"institution\":\"KU Leuven\",\"correspondingAuthor\":false,\"prefix\":\"\",\"firstName\":\"Roel\",\"middleName\":\"\",\"lastName\":\"Meyermans\",\"suffix\":\"\"},{\"id\":608196444,\"identity\":\"c14ece37-9b8f-42ad-88bc-8e69561eb878\",\"order_by\":3,\"name\":\"Léa Chapard\",\"email\":\"\",\"orcid\":\"\",\"institution\":\"Université Paris-Saclay\",\"correspondingAuthor\":false,\"prefix\":\"\",\"firstName\":\"Léa\",\"middleName\":\"\",\"lastName\":\"Chapard\",\"suffix\":\"\"},{\"id\":608196445,\"identity\":\"6b83534b-c6a5-4662-9320-8272545f2167\",\"order_by\":4,\"name\":\"Maarten Oosterlinck\",\"email\":\"\",\"orcid\":\"\",\"institution\":\"Ghent University\",\"correspondingAuthor\":false,\"prefix\":\"\",\"firstName\":\"Maarten\",\"middleName\":\"\",\"lastName\":\"Oosterlinck\",\"suffix\":\"\"},{\"id\":608196446,\"identity\":\"586c5c6f-e506-4652-93d8-584331e6aaed\",\"order_by\":5,\"name\":\"Frederik Pille\",\"email\":\"\",\"orcid\":\"\",\"institution\":\"Ghent University\",\"correspondingAuthor\":false,\"prefix\":\"\",\"firstName\":\"Frederik\",\"middleName\":\"\",\"lastName\":\"Pille\",\"suffix\":\"\"},{\"id\":608196447,\"identity\":\"3763b9c1-f36b-445e-ac33-8700fa442171\",\"order_by\":6,\"name\":\"Nadine Buys\",\"email\":\"\",\"orcid\":\"\",\"institution\":\"KU Leuven\",\"correspondingAuthor\":false,\"prefix\":\"\",\"firstName\":\"Nadine\",\"middleName\":\"\",\"lastName\":\"Buys\",\"suffix\":\"\"}],\"badges\":[],\"createdAt\":\"2026-02-09 16:10:34\",\"currentVersionCode\":1,\"declarations\":\"\",\"doi\":\"10.21203/rs.3.rs-8832466/v1\",\"doiUrl\":\"https://doi.org/10.21203/rs.3.rs-8832466/v1\",\"draftVersion\":[],\"editorialEvents\":[],\"editorialNote\":\"\",\"failedWorkflow\":false,\"files\":[{\"id\":105035640,\"identity\":\"049f0010-3ad7-4e03-94b1-b8df36cf4293\",\"added_by\":\"auto\",\"created_at\":\"2026-03-20 07:26:23\",\"extension\":\"png\",\"order_by\":1,\"title\":\"Figure 1\",\"display\":\"\",\"copyAsset\":false,\"role\":\"figure\",\"size\":477187,\"visible\":true,\"origin\":\"\",\"legend\":\"\\u003cp\\u003eManhattan plots from GWAS. \\u003cstrong\\u003eA, B)\\u003c/strong\\u003e GWAS for total OC. \\u003cstrong\\u003eA)\\u003c/strong\\u003e Cases had OC (score ≥ 3) at one or more of the following sites: distal intermediate ridge of the tibia, lateral trochlear ridge of the talus, medial malleolus of the tibia, lateral trochlear ridge of the femur, dorsal part of the sagittal ridge of the third metacarpal/metatarsal bone. Controls had a normal bone contour (score 0) at these locations (307 cases and 325 controls). \\u003cstrong\\u003eB)\\u003c/strong\\u003e Both groups were additionally required to be free from DOF, POF and other osteochondral fragments in the fetlock joint (score 0), and to present a normal bone contour (score 0) at the medial trochlear ridge of the femur, medial trochlear ridge of the talus and articular surface of the patella (174 cases and 236 controls). \\u003cstrong\\u003eC, D)\\u003c/strong\\u003e GWAS for hock OC. \\u003cstrong\\u003eC)\\u003c/strong\\u003e Cases had OC (score ≥ 3) at the distal intermediate ridge of the tibia and/or lateral trochlear ridge of the talus and/or medial malleolus of the tibia, while controls had a normal bone contour (score 0) at these sites (187 cases and 478 controls). \\u003cstrong\\u003eD)\\u003c/strong\\u003e For both groups, a normal bone contour (score 0) was additionally required at the lateral trochlear ridge of the femur and the dorsal part of the sagittal ridge of the third metacarpal/metatarsal bone (121 cases and 325 controls). \\u003cstrong\\u003eE, F)\\u003c/strong\\u003e GWAS for DIRT OC.\\u003cstrong\\u003e E) \\u003c/strong\\u003eCases had OC (score ≥ 3) at the distal intermediate ridge of the tibia, and controls had a normal bone contour (score 0) at this site. For both groups, a normal bone contour (score0) was additionally required at the lateral trochlear ridge of the talus, the medial malleolus of the tibia (146 cases and 478 controls). \\u003cstrong\\u003eF) \\u003c/strong\\u003eFor both groups, a normal bone contour (score 0) was additionally required at the lateral trochlear ridge of the femur and the dorsal part of the sagittal ridge of the third metacarpal/metatarsal bone (100 cases and 325 controls). \\u003cstrong\\u003eG, H, I)\\u003c/strong\\u003e GWAS for POF. \\u003cstrong\\u003eG) \\u003c/strong\\u003eCases had POF in at least one fetlock joint, and controls had no POF (76 cases and 619 controls). \\u003cstrong\\u003eH)\\u003c/strong\\u003e Both groups were additionally required to be free from DOF and other osteochondral fragments in the fetlock joint (47 cases and 486 controls). \\u003cstrong\\u003eI) \\u003c/strong\\u003eBoth groups were additionally required to be free from fetlock OC by presenting a normal bone contour (score 0) at the dorsal part of the sagittal ridge of the third metacarpal/metatarsal bone (40 cases and 394 controls).\\u003c/p\\u003e\",\"description\":\"\",\"filename\":\"1.png\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-8832466/v1/7bbbfc80dc3a6b8e104505ac.png\"},{\"id\":105035379,\"identity\":\"e1e17019-c80a-4f9a-8863-aa0e8b9bed8d\",\"added_by\":\"auto\",\"created_at\":\"2026-03-20 07:25:58\",\"extension\":\"png\",\"order_by\":2,\"title\":\"Figure 2\",\"display\":\"\",\"copyAsset\":false,\"role\":\"figure\",\"size\":67685,\"visible\":true,\"origin\":\"\",\"legend\":\"\\u003cp\\u003eSNP-based heritability (h\\u003csup\\u003e2\\u003c/sup\\u003e) for osteochondrosis (OC) and fetlock osteochondral fragments across two to four phenotype strictness levels (A–D). Phenotypes include total OC, hock OC, distal intermediate ridge of the tibia (DIRT) OC, stifle OC, fetlock OC, palmaro-/plantaroproximal osteochondral fragments (POF) of the proximal phalanx, and dorsoproximal osteochondral fragments (DOF) of the proximal phalanx. The strictness level applied to the basis phenotype definition are detailed in corresponding Additional file 3: Supplementary Figures S2-S15. Error bars represent the standard errors of the estimates.\\u003c/p\\u003e\\n\\u003cp\\u003e\\u0026nbsp;\\u003c/p\\u003e\",\"description\":\"\",\"filename\":\"2.png\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-8832466/v1/4fc3b3b866770b86aec74ed6.png\"},{\"id\":105727768,\"identity\":\"7f8a89f4-0481-4953-94d2-5f8b8824ebc4\",\"added_by\":\"auto\",\"created_at\":\"2026-03-30 11:03:22\",\"extension\":\"pdf\",\"order_by\":0,\"title\":\"\",\"display\":\"\",\"copyAsset\":false,\"role\":\"manuscript-pdf\",\"size\":1297984,\"visible\":true,\"origin\":\"\",\"legend\":\"\",\"description\":\"\",\"filename\":\"manuscript.pdf\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-8832466/v1/b7dedff4-74f0-4f8e-a04e-23b8b369d79a.pdf\"},{\"id\":104994530,\"identity\":\"cdd2d6fe-aa04-45ac-83bc-eaed1df2ae5c\",\"added_by\":\"auto\",\"created_at\":\"2026-03-19 16:00:53\",\"extension\":\"docx\",\"order_by\":0,\"title\":\"\",\"display\":\"\",\"copyAsset\":false,\"role\":\"supplement\",\"size\":18169,\"visible\":true,\"origin\":\"\",\"legend\":\"\",\"description\":\"\",\"filename\":\"AdditionalFile1SupplementaryTableS1.docx\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-8832466/v1/57701bd7728462b16ac51f11.docx\"},{\"id\":104994527,\"identity\":\"2c25ab8d-4312-49e0-a424-afb733f17acf\",\"added_by\":\"auto\",\"created_at\":\"2026-03-19 16:00:53\",\"extension\":\"docx\",\"order_by\":1,\"title\":\"\",\"display\":\"\",\"copyAsset\":false,\"role\":\"supplement\",\"size\":19601,\"visible\":true,\"origin\":\"\",\"legend\":\"\",\"description\":\"\",\"filename\":\"AdditionalFile6SupplementaryTableS4.docx\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-8832466/v1/32db77165aff0a3d57947080.docx\"},{\"id\":104994533,\"identity\":\"2bde3c56-5397-402e-9641-a9e5df686437\",\"added_by\":\"auto\",\"created_at\":\"2026-03-19 16:00:53\",\"extension\":\"docx\",\"order_by\":2,\"title\":\"\",\"display\":\"\",\"copyAsset\":false,\"role\":\"supplement\",\"size\":51197,\"visible\":true,\"origin\":\"\",\"legend\":\"\",\"description\":\"\",\"filename\":\"AdditionalFile2SupplementaryFigureS1.docx\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-8832466/v1/57892380eb33a115e90cbbf7.docx\"},{\"id\":105035754,\"identity\":\"e93495d7-85a1-42ee-b16b-145223e819b3\",\"added_by\":\"auto\",\"created_at\":\"2026-03-20 07:26:34\",\"extension\":\"docx\",\"order_by\":3,\"title\":\"\",\"display\":\"\",\"copyAsset\":false,\"role\":\"supplement\",\"size\":31782,\"visible\":true,\"origin\":\"\",\"legend\":\"\",\"description\":\"\",\"filename\":\"AdditionalFile5SupplementaryTableS3.docx\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-8832466/v1/2ebe2608a99782ae8063aaa3.docx\"},{\"id\":104994532,\"identity\":\"d458d907-54ad-4468-aee8-96075ee57f9e\",\"added_by\":\"auto\",\"created_at\":\"2026-03-19 16:00:53\",\"extension\":\"docx\",\"order_by\":4,\"title\":\"\",\"display\":\"\",\"copyAsset\":false,\"role\":\"supplement\",\"size\":83784,\"visible\":true,\"origin\":\"\",\"legend\":\"\",\"description\":\"\",\"filename\":\"AdditionalFile4SupplementaryTableS2.docx\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-8832466/v1/8f680e4c155271d0f14a483e.docx\"},{\"id\":104994534,\"identity\":\"e3e7a450-d7e2-4fb6-a571-958d81068a8d\",\"added_by\":\"auto\",\"created_at\":\"2026-03-19 16:00:54\",\"extension\":\"docx\",\"order_by\":5,\"title\":\"\",\"display\":\"\",\"copyAsset\":false,\"role\":\"supplement\",\"size\":1226872,\"visible\":true,\"origin\":\"\",\"legend\":\"\",\"description\":\"\",\"filename\":\"AdditionalFile3SupplementaryFiguresS2S15.docx\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-8832466/v1/4601d5731972e688b1f00879.docx\"}],\"financialInterests\":\"No competing interests reported.\",\"formattedTitle\":\"Leveraging Joint-Specific Phenotypes for Genome-Wide Association Studies and SNP Heritability Estimation of Equine Osteochondrosis and Fetlock Osteochondral Fragments\",\"fulltext\":[{\"header\":\"Background\",\"content\":\"\\u003cp\\u003eOsteochondrosis (OC) and fetlock osteochondral fragments are common osteoarticular disorders in horses, characterized by failure of the growing skeleton. OC is identified radiographically by abnormal bone contours, with or without adjacent radiopaque fragments, at specific predilection sites. Fetlock osteochondral fragments, particularly dorsoproximal osteochondral fragments (DOF) and palmaro-/plantaroproximal osteochondral fragments (POF) of the proximal phalanx, are identified by radiopaque fragments at their respective sites. The etiopathogenesis of OC, DOF, and POF differs and involves both environmental factors [\\u003cspan citationid=\\\"CR1\\\" class=\\\"CitationRef\\\"\\u003e1\\u003c/span\\u003e], such as the prenatal environment, biomechanical trauma, growth, nutrition, and weaning, and genetic factors [\\u003cspan citationid=\\\"CR2\\\" class=\\\"CitationRef\\\"\\u003e2\\u003c/span\\u003e].\\u003c/p\\u003e \\u003cp\\u003eOC and fetlock osteochondral fragments increase the risk of joint effusion, reduced performance, and eventually lameness [\\u003cspan additionalcitationids=\\\"CR4 CR5\\\" citationid=\\\"CR3\\\" class=\\\"CitationRef\\\"\\u003e3\\u003c/span\\u003e\\u0026ndash;\\u003cspan citationid=\\\"CR6\\\" class=\\\"CitationRef\\\"\\u003e6\\u003c/span\\u003e]. The high prevalence rates of these disorders [\\u003cspan additionalcitationids=\\\"CR8 CR9 CR10 CR11\\\" citationid=\\\"CR7\\\" class=\\\"CitationRef\\\"\\u003e7\\u003c/span\\u003e\\u0026ndash;\\u003cspan citationid=\\\"CR12\\\" class=\\\"CitationRef\\\"\\u003e12\\u003c/span\\u003e], combined with significant economic and welfare costs due to preventive and curative surgeries, has prompted studbooks to implement strict phenotype-based selection criteria [\\u003cspan citationid=\\\"CR2\\\" class=\\\"CitationRef\\\"\\u003e2\\u003c/span\\u003e, \\u003cspan citationid=\\\"CR13\\\" class=\\\"CitationRef\\\"\\u003e13\\u003c/span\\u003e]. However, these efforts have been undermined by the dynamic nature of lesion development and healing, surgical fragment removal without radiological evidence, and environmental factors masking true genetic predisposition [\\u003cspan citationid=\\\"CR2\\\" class=\\\"CitationRef\\\"\\u003e2\\u003c/span\\u003e]. These interfering factors may cause inaccurate breeding decisions, hindering prevalence reduction. Consequently, breeding organizations have invested in the detection of single nucleotide polymorphisms (SNPs) and quantitative trait loci (QTL) associated with OC and fetlock osteochondral fragments, as well as in the development of genomic breeding values for these disorders.\\u003c/p\\u003e \\u003cp\\u003eAcross breeds, heritability estimates for certain types of osteochondral lesions can be as high as 0.52, indicating substantial potential for genetic improvement through selective breeding [\\u003cspan citationid=\\\"CR14\\\" class=\\\"CitationRef\\\"\\u003e14\\u003c/span\\u003e, \\u003cspan citationid=\\\"CR15\\\" class=\\\"CitationRef\\\"\\u003e15\\u003c/span\\u003e]. However, heritability varies significantly among lesion types, with some showing very low estimates, suggesting that environmental factors may have a stronger influence than genetic ones. Consequently, targeted management practices are likely to be more effective in reducing lesion prevalence than selection alone. To support effective breeding decisions and guide selection strategies, accurate joint-specific heritability estimates for specific osteochondral disorder (sub)types are needed [\\u003cspan citationid=\\\"CR2\\\" class=\\\"CitationRef\\\"\\u003e2\\u003c/span\\u003e, \\u003cspan citationid=\\\"CR16\\\" class=\\\"CitationRef\\\"\\u003e16\\u003c/span\\u003e].\\u003c/p\\u003e \\u003cp\\u003eA major challenge in the transition towards using genomic information is the accurate identification of SNPs, QTLs, and genes associated with susceptibility to OC, DOF, and POF. Genetic analyses across ten equine breeds have identified QTLs associated with OC and POF on 27 autosomes and one sex chromosome, leading to the identification of 71 candidate genes [\\u003cspan citationid=\\\"CR2\\\" class=\\\"CitationRef\\\"\\u003e2\\u003c/span\\u003e]. In contrast, no QTLs for DOF had been reported prior to this study. The large number and variability of identified QTLs and candidate genes can be attributed, at least in part, to the lack of uniform phenotyping protocols and phenotype definitions across studies. In addition, fragmented data across breeds and limited sample sizes continue to hinder progress.\\u003c/p\\u003e \\u003cp\\u003eTo address these challenges and provide news insights into the genetics underlying OC and fetlock osteochondral fragments, we conducted joint-specific case\\u0026ndash;control genome‐wide association studies (GWAS) and estimated SNP-based heritabilities using data from 716 horses. We leveraged joint-specific case-control phenotypes derived from a newly developed semi-quantitative scoring system paired with 670K SNP-chip genotypes. Seven base phenotype definitions (total OC, hock OC, distal intermediate ridge of the tibia OC, stifle OC, fetlock OC, DOF and POF) across two to four different strictness levels were used to enhance the specificity of the genetic analyses. Importantly, this represents the first published GWAS of DOF in horses. Our objectives were to: 1) identify significantly associated SNPs; 2) identify novel candidate genes; 3) estimate SNP‐based heritability; and 4) evaluate if and how GWAS results and heritability estimates vary depending on the phenotypes studied.\\u003c/p\\u003e\"},{\"header\":\"Methods\",\"content\":\"\\u003cp\\u003e\\u003cem\\u003eAnimals\\u003c/em\\u003e\\u003c/p\\u003e\\n\\u003cp\\u003eBlood samples, radiographic data and pedigree data were collected from 763 horses admitted at the Faculty of Veterinary Medicine Ghent University in Belgium for radiographic examination (stallion inspection or pre-purchase examination) or arthroscopy (osteochondral fragment removal) from 2012 to 2023. The sample included 469 males and 293 females registered across 12 Warmblood studbooks, with one additional male being an English Thoroughbred. The majority of horses (84.27%) were registered in Belgian Studbooks: Belgian Warmblood Horse, Zangersheide and Belgian Sport Horse (Additional file 1: Supplementary Table S1). At the time of radiographic screening, the horses had a minimum age of 216 days and a maximum age of 6132 days, with a mean age of 1100.6 days (standard deviation (SD)\\u0026thinsp;= 663.1, interquartile range (IQR) 775.5 - 1182.0).\\u003c/p\\u003e\\n\\n\\u003cp\\u003e\\u003cem\\u003eRadiography\\u003c/em\\u003e\\u003c/p\\u003e\\n\\u003cp\\u003eHorses admitted for radiographic examination were sedated with 0.1 ml/100 kg detomidine (Detogesic\\u003csup\\u003e\\u0026reg;\\u003c/sup\\u003e) and 0.1 ml/100 kg butorphanol (Torbugesic\\u003csup\\u003e\\u0026reg;\\u003c/sup\\u003e) administered intravenously. The hooves were trimmed and cleaned, and the sulci of the frog were packed with modelling compound (Play-Doh\\u003csup\\u003e\\u0026reg;\\u003c/sup\\u003e). The radiographic protocol included at least the following projections. Lateromedial, dorso(55\\u0026deg;)proximal-palmarodistal oblique and dorso(65\\u0026deg;)proximal-palmarodistal oblique projections of both front feet. Lateromedial projection of all four fetlock joints. Dorsoplantar, lateromedial, dorso(45\\u0026deg;)lateral-plantaromedial oblique and plantaro(45\\u0026deg;)lateral-dorsomedial oblique projections of both hocks. Caudal(60\\u0026deg;)lateral-dorsomedial oblique projection of both stifles. Complementary projections were performed if a radiographic abnormality was suspected on one of these basic projections or if the client preferred a more extensive radiographic examination. For horses admitted for arthroscopy, the radiographic examination performed by the referring veterinarian was evaluated. If the radiographic screening was incomplete or if there was doubt about radiographic abnormalities on their projections, supplementary radiographic projections were performed.\\u003c/p\\u003e\\n\\n\\u003cp\\u003e\\u003cem\\u003ePhenotyping\\u003c/em\\u003e\\u003c/p\\u003e\\n\\u003cp\\u003eAs part of the equine hospital routine, all the radiographic examinations were evaluated by at least one specialist and one resident of the European College of Veterinary Diagnostic Imaging. Subsequently, all radiographic images and accompanying reports were reevaluated and scored by the first author. To classify radiographic findings indicative of osteochondrosis, a new five-grade scoring system was developed. This system scored the bone contours of specific anatomical locations as follows: 0 - normal, 1 - flattened, 2 - irregular, 3 - notch, and 4 - notch with adjacent osteochondral fragment(s). The common predilection sites of OC were evaluated using this system, including the lateral trochlear ridge of the femur, the distal intermediate ridge of the tibia, the lateral trochlear ridge of the talus, the medial malleolus of the tibia, and the dorsal part of the sagittal ridge of the third metacarpal/metatarsal bone [7, 17, 18]. The most dorsoproximal part of the sagittal ridge of the third metacarpal/metatarsal bone was not taken into account during scoring because there is a lot of radiographic variation (irregularity, indentation, radiolucency, notch) in this region [19] which does not histologically resemble OC lesions [20]. Additionally, uncommon locations of OC such as the medial trochlear ridge of the femur, medial trochlear ridge of the talus and the articular surface of the patella using the same system were scored [7]. Fetlock osteochondral fragments were scored as either 0 (absent) or 1 (present) and were subdivided into three categories: DOF, POF, and other osteochondral fragments in the fetlock joint, such as ununited palmar/plantar eminences, synovial pad fragments, and fractures of the proximal sesamoid bones. \\u003c/p\\u003e\\n\\n\\u003cp\\u003e\\u003cem\\u003eGenotyping\\u003c/em\\u003e\\u003c/p\\u003e\\n\\u003cp\\u003eAll 763 horses were genotyped using the Axiom Equine Genotyping Array (670,796 SNPs) in three different batches: 184 were genotyped at SEGALAB (Laborat\\u0026oacute;rio de Sanidade Animal e Seguran\\u0026ccedil;a Alimentar SA), 359 at CeGen-ISCIII (Spanish National Centre of Genotyping), and 220 at the Array and Analysis facility Uppsala University. For 396 horses, SNP positions were remapped from the EquCab 2.0 reference genome to EquCab 3.0 using ThermoFisher Scientific\\u0026rsquo;s axiom analysis suite software. PLINK 1.9 [21, 22] and R [23] were used to merge the genotypes and to identify the 535,330 common SNP positions consisting of those remapped from EquCab 2.0 and those originally mapped to EquCab 3.0.\\u003c/p\\u003e\\n\\n\\u003cp\\u003e\\u003cem\\u003eQuality control\\u003c/em\\u003e\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cem\\u003e \\u003c/em\\u003eQuality control was performed using PLINK 1.9 [21, 22] following a protocol based on Anderson et al. [24]. The per-individual quality control resulted in the removal of 2 horses with discordant sex information and the removal of 1 horse with outlying heterozygosity rate. Additionally, a parentage test was performed in which the genomic relatedness was compared with the pedigree relatedness resulting in the removal of 44 horses. The principal component analysis could not identify large differences in ancestry between breeds (Additional file 2: Supplementary Figure S1), therefore the remaining 716 horses were considered as one population. The per-marker quality control excluded: (1) 22,070 SNPs located on sex chromosomes, (2) 212 SNPs with a call rate \\u0026lt;95%, (3) 3,817 SNPs with significant deviation from the Hardy\\u0026ndash;Weinberg equilibrium (P \\u0026lt;0.0001) and (4) 74,920 SNPs with low minor allele frequency (\\u0026lt;1%) resulting in 434,311 remaining SNPs. The remaining SNPs were pruned for linkage disequilibrium at a squared correlation coefficient (R\\u003csup\\u003e2\\u003c/sup\\u003e) level of 0.5 thereby retaining 177,029 SNPs.\\u003c/p\\u003e\\n\\n\\u003cp\\u003e\\u003cem\\u003eAnalysis\\u003c/em\\u003e\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cem\\u003e \\u003c/em\\u003eFor the case/control genome-wide association analysis we tested two statistical approaches: a linear regression model and a mixed linear model that incorporated a sparse genetic relationship matrix as a random effect to account for population structure and relatedness. Despite the theoretical advantages of the mixed linear model, its implementation had minimal impact on the results. Consequently, we selected the linear regression model as the final analytical approach due to its computational efficiency and comparable performance. Sex was included as a fixed effect, and to control for residual population structure, the first six principal components of the genotypic data were additionally included as fixed effects. These six principal components, determined by a scree test, accounted for 46.9% of the total genetic variation. \\u003c/p\\u003e\\n\\n\\u003cp\\u003eSuggestive signals were identified using a threshold of -log10(1/177,029) (P \\u0026lt; 5.6\\u0026times;10\\u003csup\\u003e-6\\u003c/sup\\u003e) [25], while genome-wide significant signals were identified using the Bonferroni-corrected threshold of -log10(0.05/177,029) (P \\u0026lt; 2.8\\u0026times;10\\u003csup\\u003e\\u0026minus;7\\u003c/sup\\u003e) [26, 27]. GWA analysis and visualisation were performed in R using genome-wide complex trait analysis (GCTA) [28, 29] and the qqman package [30]. To define genomic regions of interest, we applied the following criteria. GWAS case\\u0026ndash;control groups were required to include at least 40 cases and more than 40 controls. In addition, at least three distinct SNPs within a 3 Mb region had to reach at least suggestive significance across the different phenotype-definition strictness levels. Genes within each identified genomic region, including an additional 0.5 Mb range upstream and downstream, were identified using Ensemble genome browser BioMart tool [31, 32].\\u003c/p\\u003e\\n\\n\\u003cp\\u003eCase and control selection was conducted on a joint-specific basis, using progressively stricter phenotype definitions to increase specificity (A to D). For total OC, the basis definition (A) for cases was the presence of a bone contour notch, with or without adjacent osteochondral fragment(s) (score \\u0026ge; 3), at the common predilection site(s) of OC of the hock and/or stifle and/or fetlock. Controls were defined by the presence of a normal bone contour (score 0) at the same site(s). The additional strictness level (B) required both cases and controls to have no (score 0) osteochondral fragments in the fetlock joint and a normal bone contour (score 0) at uncommon OC locations.\\u003c/p\\u003e\\n\\n\\u003cp\\u003eFor OC of the hock, stifle, and fetlock separately, the basis definition (A) for cases was the presence of a bone contour notch, with or without adjacent osteochondral fragment(s) (score \\u0026ge; 3), at the common predilection site(s) of OC within the respective joint. Controls were defined by the presence of a normal bone contour (score 0) at the same site(s). The additional strictness levels, applied cumulatively and required for both cases and controls, were: (B) a normal bone contour (score 0) at other common OC predilection sites; (C) the absence of osteochondral fragments (score 0) in the fetlock joint and a normal bone contour (score 0) at uncommon OC locations. \\u003c/p\\u003e\\n\\n\\u003cp\\u003eFor OC of the distal intermediate ridge of the tibia (DIRT OC), which is the most common location of hock OC [17], a specific GWAS was performed. The basis definition (A) required cases to have a bone contour notch, with or without adjacent osteochondral fragments (score \\u0026ge; 3), at the distal intermediate ridge. Controls were required to have a normal bone contour (score 0) at the same site. The additional strictness levels, applied cumulatively and to both cases and controls, were: (B) a normal bone contour (score 0) at the lateral trochlear ridge of the talus and the medial malleolus of the tibia (other common predilection sites of hock OC); (C) a normal bone contour (score 0) at other common OC predilection sites; (D) the absence of osteochondral fragments (score 0) in the fetlock joint and a normal bone contour (score 0) at uncommon OC locations.\\u003c/p\\u003e\\n\\n\\u003cp\\u003eFor both POF and DOF, the basis definition (A) classified cases by the presence of the respective condition in at least one fetlock joint, while controls were defined by its absence in all fetlock joints. The additional strictness levels, applied cumulatively and required for both cases and controls, were: (B) the absence of the other condition (POF or DOF) and other osteochondral fragments in the fetlock joint; (C) a normal bone contour (score 0) at the common location for fetlock OC; and (D) a normal bone contour (score 0) at other common and uncommon OC locations.\\u003c/p\\u003e\\n\\n\\u003cp\\u003eSNP-based heritability for the different phenotype definitions was estimated using the restricted maximum likelihood (REML) approach in GCTA [28]. A genomic relationship matrix (GRM) was constructed from autosomal SNPs, and case-control analyses were performed under progressively stricter phenotype definitions to evaluate their impact on heritability estimates. The same covariates from the GWAS analysis (sex and the first six principal components) were included. Heritability was calculated based on the genetic and residual variance estimated from the model.\\u003c/p\\u003e\"},{\"header\":\"Results\",\"content\":\"\\u003cp\\u003eManhattan plots for the GWAS of total OC, hock OC, DIRT OC and POF are presented in Figure 1, with a complete overview of Manhattan and and Quantile-Quantile plots for all phenotypes provided in\\u0026nbsp;Additional file 3: Supplementary Figures S2-S15. Across the different phenotype definitions, we identified 214 SNPs with suggestive significance, 50 of which reached genome-wide significance. Genome-wide significant association signals were identified for total OC (score \\u0026ge; 3 at the fetlock and/or hock and/or stifle OC) on Equus caballus chromosome (ECA) 3; for hock OC (score \\u0026ge; 3 at the distal intermediate ridge of the tibia and/or lateral trochlear ridge of the talus and/or medial malleolus of the tibia) on ECA 20 and 24; for DIRT OC (score \\u0026ge; 3 at the distal intermediate ridge of the tibia) on ECA 17 and 29; for stifle OC (score \\u0026ge; 3 at the lateral trochlear ridge of the femur) on ECA 30; for fetlock OC (score \\u0026ge; 3 at the dorsal part of the sagittal ridge of the third metacarpal/metatarsal bone) on ECA 17, 22 and 27; for POF (score = 1 for POF in at least one of the fetlock joints) on ECA 1, 2, 4, 5, 6, 7, 10, 14, 15, 18, 19, 20, 23, 24 and 25; and for DOF (score = 1 for DOF in at least one of the fetlock joints) on ECA 10, 15 and 18. All SNPs with suggestive and genome-wide significance and their corresponding significance levels for each phenotype are listed in Additional file 4: Supplementary Table S2.\\u003c/p\\u003e\\n\\u003cp\\u003eTable 1 provides an overview of the genomic regions that met our criteria, listing the gene that contains the SNP with the lowest P-value as well as other candidate genes within each region. For each region, we reported the number of SNPs reaching suggestive and genome-wide significance, along with the lowest observed P-value. Across the identified genomic regions, including an additional 0.5 Mb range upstream and downstream, we detected 201 protein coding genes (Additional file 5: Supplementary Table S3). To nominate candidate genes, we annotated the gene ontology and identified the following genes associated with metabolic pathways involved in ossification and angiogenesis: \\u003cem\\u003eLDB2\\u003c/em\\u003e, \\u003cem\\u003eFGF6\\u003c/em\\u003e, \\u003cem\\u003eFGF23\\u003c/em\\u003e, \\u003cem\\u003eSCUBE3\\u003c/em\\u003e, \\u003cem\\u003eMAPK13\\u003c/em\\u003e, \\u003cem\\u003eMAPK14\\u003c/em\\u003e, \\u003cem\\u003eIL17A\\u003c/em\\u003e, \\u003cem\\u003eIL17F\\u003c/em\\u003e, \\u003cem\\u003eTRAM2,\\u003c/em\\u003e \\u003cem\\u003eBMP5\\u0026nbsp;\\u003c/em\\u003eand\\u003cem\\u003e\\u0026nbsp;\\u003c/em\\u003e\\u003cem\\u003ePPP2R5C\\u003c/em\\u003e. \\u0026nbsp;\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cstrong\\u003eTable 1.\\u003c/strong\\u003e Genomic regions and candidate genes associated with osteochondrosis phenotypes in horses based on a case/control GWAS using a linear regression model. Regions are defined as genomic intervals in which at least three SNPs within a 3 Mb window show suggestive significance across the varying phenotype definition strictness levels. Trait refers to the phenotype definition (see supplementary figure referenced in parentheses for details on phenotype definitions and number of horses). \\u0026lsquo;N SNPs\\u0026rsquo; indicates the number of SNPs with suggestive significance (P \\u0026lt; 5.6\\u0026times;10\\u003csup\\u003e-6\\u003c/sup\\u003e), with the number reaching genome-wide significance (P \\u0026lt; 2.8\\u0026times;10\\u003csup\\u003e\\u0026minus;7\\u003c/sup\\u003e) in parentheses. The column \\u0026ldquo;\\u0026ndash;log₁₀(\\u003cem\\u003eP\\u003c/em\\u003e)\\u0026rdquo; shows the smallest \\u003cem\\u003eP\\u003c/em\\u003e-value among SNPs in the region. Candidate genes include those located within the identified region plus 0.5 Mb upstream and downstream.\\u0026nbsp;\\u003c/p\\u003e\\n\\u003ctable border=\\\"0\\\" cellspacing=\\\"0\\\" cellpadding=\\\"0\\\"\\u003e\\n \\u003ctbody\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd valign=\\\"top\\\"\\u003e\\n \\u003cp\\u003eECA\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\"\\u003e\\n \\u003cp\\u003eRegion (Mb)\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\"\\u003e\\n \\u003cp\\u003eTrait\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\"\\u003e\\n \\u003cp\\u003eN SNP\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\"\\u003e\\n \\u003cp\\u003e-log\\u003csub\\u003e10\\u003c/sub\\u003e(P)\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\"\\u003e\\n \\u003cp\\u003eCandidate Genes\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd valign=\\\"top\\\"\\u003e\\n \\u003cp\\u003e3\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\"\\u003e\\n \\u003cp\\u003e108.28-108.50\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\"\\u003e\\n \\u003cp\\u003eT (S2)\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\"\\u003e\\n \\u003cp\\u003e3 (1)\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\"\\u003e\\n \\u003cp\\u003e6.77\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\"\\u003e\\n \\u003cp\\u003e\\u003cem\\u003eLDB2\\u0026nbsp;\\u003c/em\\u003e(LIM domain binding 2)\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd valign=\\\"top\\\"\\u003e\\n \\u003cp\\u003e6\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\"\\u003e\\n \\u003cp\\u003e32.94-33.92\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\"\\u003e\\n \\u003cp\\u003ePOF (S12)\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\"\\u003e\\n \\u003cp\\u003e7 (1)\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\"\\u003e\\n \\u003cp\\u003e6.62\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\"\\u003e\\n \\u003cp\\u003e\\u003cem\\u003eFGF23\\u0026nbsp;\\u003c/em\\u003e(fibroblast growth factor 23);\\u0026nbsp;\\u003cbr\\u003e\\u003cem\\u003eFGF6\\u003c/em\\u003e (fibroblast growth factor 6)\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd valign=\\\"top\\\"\\u003e\\n \\u003cp\\u003e7\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\"\\u003e\\n \\u003cp\\u003e5.80-8.03\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\"\\u003e\\n \\u003cp\\u003ePOF (S12)\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\"\\u003e\\n \\u003cp\\u003e3 (1)\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\"\\u003e\\n \\u003cp\\u003e6.70\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\"\\u003e\\n \\u003cp\\u003e\\u0026nbsp;\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd valign=\\\"top\\\"\\u003e\\n \\u003cp\\u003e20\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\"\\u003e\\n \\u003cp\\u003e35.58-38.87\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\"\\u003e\\n \\u003cp\\u003eH (S4)\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\"\\u003e\\n \\u003cp\\u003e4 (2)\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\"\\u003e\\n \\u003cp\\u003e7.28\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\"\\u003e\\n \\u003cp\\u003e\\u003cem\\u003eSCUCBE3\\u0026nbsp;\\u003c/em\\u003e(signal peptide, CUB domain and EGF like domain containing 3);\\u003cbr\\u003e\\u003cem\\u003eMAPK13\\u003c/em\\u003e (mitogen-activated protein kinase 13);\\u003cem\\u003e\\u0026nbsp;\\u003cbr\\u003e\\u0026nbsp;MAPK14\\u0026nbsp;\\u003c/em\\u003e(mitogen-activated protein kinase 14)\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd valign=\\\"top\\\"\\u003e\\n \\u003cp\\u003e20\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\"\\u003e\\n \\u003cp\\u003e50.18-60.96\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\"\\u003e\\n \\u003cp\\u003ePOF (S12)\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\"\\u003e\\n \\u003cp\\u003e16 (7)\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\"\\u003e\\n \\u003cp\\u003e8.52\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\"\\u003e\\n \\u003cp\\u003e\\u003cem\\u003eIL17A\\u003c/em\\u003e (interleukin 17A);\\u003cem\\u003e\\u0026nbsp;\\u003cbr\\u003e\\u0026nbsp;IL17F\\u003c/em\\u003e (interleukin 17F);\\u003cem\\u003e\\u0026nbsp;\\u003cbr\\u003e\\u0026nbsp;TRAM2\\u0026nbsp;\\u003c/em\\u003e(translocation associated membrane protein 2);\\u0026nbsp;\\u003cbr\\u003e\\u003cem\\u003eBMP5\\u0026nbsp;\\u003c/em\\u003e(bone morphogeneticprotein 5)\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd valign=\\\"top\\\"\\u003e\\n \\u003cp\\u003e24\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\"\\u003e\\n \\u003cp\\u003e43.66-44.61\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\"\\u003e\\n \\u003cp\\u003ePOF (S12)\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\"\\u003e\\n \\u003cp\\u003e5 (3)\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\"\\u003e\\n \\u003cp\\u003e13.26\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\"\\u003e\\n \\u003cp\\u003e\\u003cem\\u003ePPP2R5C\\u003c/em\\u003e (protein phosphatase 2 regulatory subunit B\\u0026apos;gamma)\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd valign=\\\"top\\\"\\u003e\\n \\u003cp\\u003e24\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\"\\u003e\\n \\u003cp\\u003e43.66-45.03\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\"\\u003e\\n \\u003cp\\u003eH (S4)\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\"\\u003e\\n \\u003cp\\u003e3 (2)\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\"\\u003e\\n \\u003cp\\u003e6.89\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\"\\u003e\\n \\u003cp\\u003e\\u0026nbsp;\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd valign=\\\"top\\\"\\u003e\\n \\u003cp\\u003e29\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\"\\u003e\\n \\u003cp\\u003e14.03-16.56\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\"\\u003e\\n \\u003cp\\u003eDIRT (S6)\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\"\\u003e\\n \\u003cp\\u003e3 (1)\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\"\\u003e\\n \\u003cp\\u003e7.24\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\"\\u003e\\n \\u003cp\\u003e\\u0026nbsp;\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003c/tbody\\u003e\\n\\u003c/table\\u003e\\n\\u003cp\\u003eDOF, Dorsoproximal osteochondral fragments of the proximal phalanx; ECA, Equus caballus chromosome; F, Fetlock osteochondrosis; H, Hock osteochondrosis; DIRT, Osteochondrosis at the distal intermediate ridge of the tibia; Mb, Megabase; POF, Palmaro-/ plantaroproximal osteochondral fragments of the proximal phalanx;; S, Stifle osteochondrosis; SNP, Single nucleotide polymorphism; T, Fetlock and/or hock and/or stifle osteochondrosis. \\u0026nbsp;\\u003c/p\\u003e\\n\\u003cp\\u003eSNP-based heritability estimates, adjusted for sex and population structure, are presented in Additional file 6: Supplementary Table S4 and illustrated in Figure 2. For total OC, heritability was moderate, ranging from 26.2% (standard error (SE) = 11.9%) at strictness level A to 27.1% (SE = 16.8%) at level B. Higher heritability estimates were observed for hock OC and DIRT OC across strictness levels, with values ranging from 53.9% (SE = 11.2%) to 61.0% (SE = 16.2%) and 49.4% (SE = 11.6%) to 61.2% (SE = 22.4%), respectively. In contrast, stifle and fetlock OC had low heritability estimates, ranging from 1.5% (SE = 23.7%) to 14.2% (SE = 10.2%) and 0.0% (SE = 24.8%) to 13.7% (SE = 11.7%), respectively. Heritability estimates for POF and DOF showed a wide range depending on the phenotype definition. For POF, the heritability estimates ranged from 10.0% (SE = 22.7%) at strictness level D to 54.8% (SE = 14.0%) at strictness level B, whereas for DOF they ranged from 5.7% (SE = 10.2%) at strictness level A to 39.6% (SE = 27.8%) at strictness level D.\\u003cbr\\u003e\\u0026nbsp;\\u003c/p\\u003e\"},{\"header\":\"Discussion\",\"content\":\"\\u003cp\\u003eIn this study, we present joint-specific case\\u0026ndash;control GWAS and SNP heritability estimation for OC and fetlock osteochondral fragments in a cohort of 716 horses. The population is almost entirely composed of Warmblood horses (99.9%), with only one English Thoroughbred included, and 84.27% are registered in Belgian studbooks. Therefore, the findings are most relevant for Warmblood horse populations, particularly those registered in Belgium.\\u003c/p\\u003e \\u003cp\\u003eTo improve the specificity of the genetic analyses, we applied progressively refined phenotype definitions based on a newly developed semi-quantitative scoring system paired with high-density 670K SNP-chip genotypes. Increasing phenotype definitions strictness improved the specificity of case and control classifications, thereby reducing potential confounding. However, this refinement came at the cost of a reduced sample size, which in turn lowered statistical power of the GWAS. The smaller sample size introduced greater variability and potential unreliability, particularly under the most restrictive phenotype definitions. The reduced statistical power was reflected in the Manhattan and QQ plots (Additional file 3: Supplementary Figures \\u003cspan refid=\\\"MOESM2\\\" class=\\\"InternalRef\\\"\\u003eS2\\u003c/span\\u003e-S15), which showed signs of genomic inflation, false-positive signals, and spurious associations under the strictest phenotype definitions. To mitigate these challenges, we applied the predefined criteria for identifying genomic regions of interest, balancing the advantages of improved phenotype specificity against the limitations of reduced sample size.\\u003c/p\\u003e \\u003cp\\u003eAcross seven base phenotype definitions (total OC, hock OC, distal intermediate ridge of the tibia OC, stifle OC, fetlock OC, DOF and POF), each evaluated at two to four strictness levels, we identified 214 genome-wide suggestive SNPs, 50 of which reached genome-wide significance and 8 genomic regions of interest that met our predefined criteria. From these regions, we nominated 11 candidate genes based on gene ontology annotations.\\u003c/p\\u003e \\u003cp\\u003eTo compare our findings with previous research, we used the review by Van Mol et al. [\\u003cspan citationid=\\\"CR2\\\" class=\\\"CitationRef\\\"\\u003e2\\u003c/span\\u003e], which provides an overview of the genomic coordinates and phenotype definitions of QTLs previously reported for equine OC and osteochondral fetlock fragments. This allowed us to evaluate whether the identified genomic regions in our study overlapped with previously reported QTLs and whether any of the nominated candidate genes had been described in prior literature. Given that our analyses were based on the EquCab 3.0 reference genome, while most prior studies used EquCab 2.0, we remapped our genomic coordinates to EquCab 2.0 using the UCSC LiftOver tool [\\u003cspan citationid=\\\"CR33\\\" class=\\\"CitationRef\\\"\\u003e33\\u003c/span\\u003e] to enable direct comparison. Since the review did not include gene expression studies, we additionally screened the literature for genes with a significantly altered expression in early OC lesions [\\u003cspan additionalcitationids=\\\"CR35 CR36 CR37 CR38 CR39 CR40\\\" citationid=\\\"CR34\\\" class=\\\"CitationRef\\\"\\u003e34\\u003c/span\\u003e\\u0026ndash;\\u003cspan citationid=\\\"CR41\\\" class=\\\"CitationRef\\\"\\u003e41\\u003c/span\\u003e]. We focused on early OC lesions, as gene expression changes observed in later stages may reflect secondary degenerative responses rather than primary pathogenic processes.\\u003c/p\\u003e \\u003cp\\u003eBased on this comparison, all 11 candidate genes identified in our study appear to be novel. This result is remarkable but not entirely surprising considering the differences in study design. Previous studies generally relied on broader phenotype definitions, whereas the present study employed more stringent, joint-specific phenotypes. This may have revealed loci that remained undetected when lesions are grouped. Furthermore, advances in functional annotation may have facilitated the identification of genes that were previously overlooked. The inability to confirm previously reported candidate genes emphasizes the complex genetic architecture underlying these conditions and demonstrates the need for larger studies based on standardized joint-specific phenotypes to establish reproducible associations.\\u003c/p\\u003e \\u003cp\\u003eIn the following sections, we describe the identified genomic regions and compare them with regions previously reported in association with equine OC or fetlock osteochondral fragments. We discuss the biological plausibility of the newly nominated candidate genes in relation to the pathogenesis of these conditions. Finally, we examine how GWAS results and heritability estimates varied depending on the phenotype definitions and assess the implications of these findings for future genetic research.\\u003c/p\\u003e \\u003cdiv id=\\\"Sec11\\\" class=\\\"Section2\\\"\\u003e \\u003ch2\\u003eTotal osteochondrosis\\u003c/h2\\u003e \\u003cp\\u003eIn association with the total OC phenotype (Fig.\\u0026nbsp;\\u003cspan refid=\\\"Fig1\\\" class=\\\"InternalRef\\\"\\u003e1\\u003c/span\\u003eA), we identified a genomic region on ECA 3 spanning 108.28\\u0026ndash;108.50 Mb (corresponding to 106.45-106.67 Mb in EquCab 2.0). Although this exact region has not been reported in prior studies, Drabbe et al. [\\u003cspan citationid=\\\"CR42\\\" class=\\\"CitationRef\\\"\\u003e42\\u003c/span\\u003e] identified a nearby QTL at 115.91-115.92 Mb (EquCab 3.0) in Belgian Warmbloods, using the same phenotype definition. Additionally, Naccache et al. [\\u003cspan citationid=\\\"CR14\\\" class=\\\"CitationRef\\\"\\u003e14\\u003c/span\\u003e] identified another nearby QTL at 105.40-105.94 Mb (EquCab 2.0) for total osteochondrosis dissecans (OCD) in Hanoverian Warmbloods, although, their phenotype definition differed from ours, as they included only horses with both an abnormal bone contour and an adjacent fragment as cases and considered two additional predilection sites. Notably, the genome-wide significant SNP on ECA 3 in our initial analysis (Fig.\\u0026nbsp;\\u003cspan refid=\\\"Fig1\\\" class=\\\"InternalRef\\\"\\u003e1\\u003c/span\\u003eA) disappeared when stricter criteria were applied to both cases and controls (Fig.\\u0026nbsp;\\u003cspan refid=\\\"Fig1\\\" class=\\\"InternalRef\\\"\\u003e1\\u003c/span\\u003eB).\\u003c/p\\u003e \\u003cp\\u003eOn ECA 3 the SNP with the lowest P-value associated with total OC is located within LIM domain binding 2 (\\u003cem\\u003eLDB2\\u003c/em\\u003e), which has an important role in mediating transcription through the formation of higher-order transcription complexes and is expressed in endothelial cells. One of the genes regulated by \\u003cem\\u003eLDB2\\u003c/em\\u003e is delta-like ligand 4 (\\u003cem\\u003eDLL4\\u003c/em\\u003e), a gene important for sprouting angiogenesis and vascular remodeling [\\u003cspan citationid=\\\"CR43\\\" class=\\\"CitationRef\\\"\\u003e43\\u003c/span\\u003e]. Given that the initial stage of OC involves the focal failure of epiphyseal growth cartilage canal vessels, and that the final outcome depends on whether this area of ischemic chondronecrosis can heal [\\u003cspan additionalcitationids=\\\"CR45 CR46 CR47 CR48 CR49 CR50 CR51 CR52 CR53 CR54\\\" citationid=\\\"CR44\\\" class=\\\"CitationRef\\\"\\u003e44\\u003c/span\\u003e\\u0026ndash;\\u003cspan citationid=\\\"CR55\\\" class=\\\"CitationRef\\\"\\u003e55\\u003c/span\\u003e], \\u003cem\\u003eLDB2\\u003c/em\\u003e represents a promising candidate gene for further investigation. No other candidate genes were detected in the associated genomic region on ECA 3.\\u003c/p\\u003e \\u003c/div\\u003e \\u003cdiv id=\\\"Sec12\\\" class=\\\"Section2\\\"\\u003e \\u003ch2\\u003eHock osteochondrosis\\u003c/h2\\u003e \\u003cp\\u003eFor hock OC, our analysis (Fig.\\u0026nbsp;\\u003cspan refid=\\\"Fig1\\\" class=\\\"InternalRef\\\"\\u003e1\\u003c/span\\u003eD) revealed two novel genomic regions of interest: one on ECA 20 (35.58\\u0026ndash;38.87 Mb) and another on ECA 24 (43.66\\u0026ndash;45.03 Mb0). To our knowledge, this is the first study in which SNPs on ECA 20 and ECA 24 have been associated with any form of OC in horses. Notably, one of the two SNP with suggestive significance on ECA 3 in Fig.\\u0026nbsp;\\u003cspan refid=\\\"Fig1\\\" class=\\\"InternalRef\\\"\\u003e1\\u003c/span\\u003eC, located at 107.64 Mb (corresponding to 105.82 Mb in EquCab 2.0) falls within a previously reported QTL for hock OC spanning 100.39-107.92 Mb (EquCab 2.0) [\\u003cspan citationid=\\\"CR56\\\" class=\\\"CitationRef\\\"\\u003e56\\u003c/span\\u003e, \\u003cspan citationid=\\\"CR57\\\" class=\\\"CitationRef\\\"\\u003e57\\u003c/span\\u003e]. However, this signal was no longer evident under the stricter phenotype definition applied in Fig.\\u0026nbsp;\\u003cspan refid=\\\"Fig1\\\" class=\\\"InternalRef\\\"\\u003e1\\u003c/span\\u003eD. Importantly, neither of these studies used the same phenotype definition as in our study [\\u003cspan citationid=\\\"CR56\\\" class=\\\"CitationRef\\\"\\u003e56\\u003c/span\\u003e, \\u003cspan citationid=\\\"CR57\\\" class=\\\"CitationRef\\\"\\u003e57\\u003c/span\\u003e]. Orr et al. [\\u003cspan citationid=\\\"CR56\\\" class=\\\"CitationRef\\\"\\u003e56\\u003c/span\\u003e] included only horses with OCD (abnormal bone contour\\u0026thinsp;+\\u0026thinsp;adjacent fragment) as cases and considered four additional predilection sites, while Teyssedre et al. [\\u003cspan citationid=\\\"CR57\\\" class=\\\"CitationRef\\\"\\u003e57\\u003c/span\\u003e] did not specify the precise anatomical locations affected by osteochondrosis in the hock. Interestingly, Lykkjen et al. [\\u003cspan citationid=\\\"CR58\\\" class=\\\"CitationRef\\\"\\u003e58\\u003c/span\\u003e], who used a phenotype definition similar to that in Fig.\\u0026nbsp;\\u003cspan refid=\\\"Fig1\\\" class=\\\"InternalRef\\\"\\u003e1\\u003c/span\\u003eC, identified a significant SNP on ECA 3 at 113.50 Mb (EquCab 2.0). In contrast to these and other GWAS that have identified QTLs for hock OC [\\u003cspan citationid=\\\"CR14\\\" class=\\\"CitationRef\\\"\\u003e14\\u003c/span\\u003e, \\u003cspan citationid=\\\"CR58\\\" class=\\\"CitationRef\\\"\\u003e58\\u003c/span\\u003e, \\u003cspan citationid=\\\"CR59\\\" class=\\\"CitationRef\\\"\\u003e59\\u003c/span\\u003e], our study is the first to use a phenotype definition that includes only the most common predilection sites of hock OC and excludes animals with concurrent stifle and/or fetlock OC (Fig.\\u0026nbsp;\\u003cspan refid=\\\"Fig1\\\" class=\\\"InternalRef\\\"\\u003e1\\u003c/span\\u003eD), thereby reducing potential confounding.\\u003c/p\\u003e \\u003cp\\u003eThe most significantly associated SNPs with hock OC were located on ECA 20 and ECA 24, within the ribosomal protein S10 (\\u003cem\\u003eRPS10\\u003c/em\\u003e) and CDC42 binding protein kinase beta (\\u003cem\\u003eCDC42BPB\\u003c/em\\u003e) genes, respectively. Neither gene appear to have a direct connection with osteochondrosis, and no candidate genes were identified within the associated region on ECA 24. However, within the associated genomic region of ECA 20, several candidate genes were identified, such as signal peptide, CUB domain and EGF like domain containing 3 (\\u003cem\\u003eSCUBE3)\\u003c/em\\u003e and mitogen-activated protein kinase 13 and 14 (\\u003cem\\u003eMAPK13\\u003c/em\\u003e and \\u003cem\\u003eMAPK14\\u003c/em\\u003e). \\u003cem\\u003eSCUBE3\\u003c/em\\u003e encodes a signal peptide that promotes bone morphogenesis, playing an important role in bone morphology and metabolism [\\u003cspan citationid=\\\"CR60\\\" class=\\\"CitationRef\\\"\\u003e60\\u003c/span\\u003e]. GWAS have identified SNPs of \\u003cem\\u003eSCUBE3\\u003c/em\\u003e among height associated loci in both pigs [\\u003cspan citationid=\\\"CR61\\\" class=\\\"CitationRef\\\"\\u003e61\\u003c/span\\u003e] and humans [\\u003cspan citationid=\\\"CR62\\\" class=\\\"CitationRef\\\"\\u003e62\\u003c/span\\u003e]. \\u003cem\\u003eSCUBE3\\u003c/em\\u003e knockout mice have viable phenotypes but show impaired BMP-mediated chondrogenesis and osteogenesis, which, among other defects, leads to abnormal endochondral bone development. Moreover, mutations in \\u003cem\\u003eSCUBE3\\u003c/em\\u003e have been linked to various human skeletal disorders [\\u003cspan citationid=\\\"CR60\\\" class=\\\"CitationRef\\\"\\u003e60\\u003c/span\\u003e, \\u003cspan citationid=\\\"CR63\\\" class=\\\"CitationRef\\\"\\u003e63\\u003c/span\\u003e].\\u003c/p\\u003e \\u003cp\\u003eMAPK13 and MAPK14, members of the p38 MAPK family, are important for cellular stress responses, tissue development, and homeostasis. The MAPK pathway regulates important processes in skeletogenesis and bone maintenance, including chondrocyte and osteoblast differentiation, extracellular matrix deposition, and mineralization. MAPK13, with selective tissue expression, modulates inflammatory responses and insulin secretion [\\u003cspan citationid=\\\"CR64\\\" class=\\\"CitationRef\\\"\\u003e64\\u003c/span\\u003e]. Disruption of the insulin balance has been associated with increased occurrence of OC [\\u003cspan additionalcitationids=\\\"CR66 CR67 CR68\\\" citationid=\\\"CR65\\\" class=\\\"CitationRef\\\"\\u003e65\\u003c/span\\u003e\\u0026ndash;\\u003cspan citationid=\\\"CR69\\\" class=\\\"CitationRef\\\"\\u003e69\\u003c/span\\u003e]. Conversely, MAPK14, which is ubiquitously expressed, influences a broader range of processes, including inflammatory responses, core cellular processes, and the Wnt signaling pathway [\\u003cspan citationid=\\\"CR64\\\" class=\\\"CitationRef\\\"\\u003e64\\u003c/span\\u003e]. Wnt signaling is altered in OC lesions and may be associated with disease pathogenesis [\\u003cspan citationid=\\\"CR38\\\" class=\\\"CitationRef\\\"\\u003e38\\u003c/span\\u003e, \\u003cspan citationid=\\\"CR70\\\" class=\\\"CitationRef\\\"\\u003e70\\u003c/span\\u003e, \\u003cspan citationid=\\\"CR71\\\" class=\\\"CitationRef\\\"\\u003e71\\u003c/span\\u003e]. MAPK14 predominates in growth-plate chondrocytes, particularly in the pre-hypertrophic zone, where it regulates hypertrophic chondrocyte differentiation. MAPK14 is also highly expressed and essential for osteoclastogenesis, and deletions of \\u003cem\\u003eMAPK14\\u003c/em\\u003e have been associated with skeletal abnormalities in mice [\\u003cspan citationid=\\\"CR64\\\" class=\\\"CitationRef\\\"\\u003e64\\u003c/span\\u003e, \\u003cspan citationid=\\\"CR72\\\" class=\\\"CitationRef\\\"\\u003e72\\u003c/span\\u003e]. The MAPK pathway mediates the effects of multiple growth and transcription factors influencing bone cell function and differentiation [\\u003cspan citationid=\\\"CR64\\\" class=\\\"CitationRef\\\"\\u003e64\\u003c/span\\u003e] of which some are found to be differently expressed in cartilage of OC lesions: TGF-β [\\u003cspan citationid=\\\"CR34\\\" class=\\\"CitationRef\\\"\\u003e34\\u003c/span\\u003e, \\u003cspan citationid=\\\"CR73\\\" class=\\\"CitationRef\\\"\\u003e73\\u003c/span\\u003e], SOX9 [\\u003cspan citationid=\\\"CR40\\\" class=\\\"CitationRef\\\"\\u003e40\\u003c/span\\u003e] and RUNX2 [\\u003cspan citationid=\\\"CR36\\\" class=\\\"CitationRef\\\"\\u003e36\\u003c/span\\u003e]. In our opinion, the overlap between the aforementioned processes regulated by the MAPK pathway and their links to OC, identifies \\u003cem\\u003eMAPK13\\u003c/em\\u003e and \\u003cem\\u003eMAPK14\\u003c/em\\u003e as compelling candidate genes for further research.\\u003c/p\\u003e \\u003c/div\\u003e \\u003cdiv id=\\\"Sec13\\\" class=\\\"Section2\\\"\\u003e \\u003ch2\\u003eDistal intermediate ridge of the tibia osteochondrosis\\u003c/h2\\u003e \\u003cp\\u003eTo enhance signal detection, we refined the phenotype of hock OC even further by focusing exclusively on OC of the distal intermediate ridge of the tibia (Additional file 3: Supplementary Figure \\u003cspan refid=\\\"MOESM6\\\" class=\\\"InternalRef\\\"\\u003eS6\\u003c/span\\u003e). This revealed a region on ECA 29 (Fig.\\u0026nbsp;\\u003cspan refid=\\\"Fig1\\\" class=\\\"InternalRef\\\"\\u003e1\\u003c/span\\u003eE, F) spanning 14.03\\u0026ndash;16.56 Mb (corresponding to 12.96\\u0026ndash;15.49 Mb in EquCab 2.0), not detected in the broader hock OC phenotype definition (Fig.\\u0026nbsp;\\u003cspan refid=\\\"Fig1\\\" class=\\\"InternalRef\\\"\\u003e1\\u003c/span\\u003eC, D). Naccache et al. [\\u003cspan citationid=\\\"CR14\\\" class=\\\"CitationRef\\\"\\u003e14\\u003c/span\\u003e] reported a nearby QTL for hock OC (16.07\\u0026ndash;16.94 Mb) and total OC (16.74\\u0026ndash;16.94 Mb in EquCab 2.0), though without specifying the exact anatomical sites affected by OC. Notably, the SNPs with the lowest P-value on ECA 1 at 34.59 Mb (Fig.\\u0026nbsp;\\u003cspan refid=\\\"Fig1\\\" class=\\\"InternalRef\\\"\\u003e1\\u003c/span\\u003eE) and ECA 29 at 14.03 Mb (Fig.\\u0026nbsp;\\u003cspan refid=\\\"Fig1\\\" class=\\\"InternalRef\\\"\\u003e1\\u003c/span\\u003eF) were also the lowest P-value SNPs on their respective chromosomes in the hock OC GWAS (Fig.\\u0026nbsp;\\u003cspan refid=\\\"Fig1\\\" class=\\\"InternalRef\\\"\\u003e1\\u003c/span\\u003eC, D), but at lower significance levels. While the SNP on ECA 1 did not meet the established criteria for a significant signal, the SNP on ECA 29 did, achieving at least suggestive significance in all four phenotype definitions, with its highest significance reaching -log\\u003csub\\u003e10\\u003c/sub\\u003e(P)\\u0026thinsp;=\\u0026thinsp;7.24 (Fig.\\u0026nbsp;\\u003cspan refid=\\\"Fig1\\\" class=\\\"InternalRef\\\"\\u003e1\\u003c/span\\u003eF). This SNP lies in a noncoding region, with no candidate genes in the associated genomic region.\\u003c/p\\u003e \\u003c/div\\u003e \\u003cdiv id=\\\"Sec14\\\" class=\\\"Section2\\\"\\u003e \\u003ch2\\u003ePalmaroproximal and plantaroproximal osteochondral fragments\\u003c/h2\\u003e \\u003cp\\u003eThe GWAS for POF (Fig.\\u0026nbsp;\\u003cspan refid=\\\"Fig1\\\" class=\\\"InternalRef\\\"\\u003e1\\u003c/span\\u003eG, H, I) identified four genomic regions of interest on: ECA 6 (32.94\\u0026ndash;33.92 Mb), ECA 7 (5.80\\u0026ndash;8.03 Mb), ECA 20 (50.18\\u0026ndash;60.96 Mb), and ECA 24 (43.66\\u0026ndash;44.61 Mb). To our knowledge, this is the first study associating SNPs on ECA 6, 20 and 24 with POF [\\u003cspan citationid=\\\"CR2\\\" class=\\\"CitationRef\\\"\\u003e2\\u003c/span\\u003e]. On ECA 7, Lykkjen et al. [\\u003cspan citationid=\\\"CR74\\\" class=\\\"CitationRef\\\"\\u003e74\\u003c/span\\u003e] reported a QTL for POF at 69.9\\u0026ndash;80.5 Mb (EquCab 2.0), which does not overlap with the region identified in this study (remapped to 5.33\\u0026ndash;7.55 Mb EquCab 2.0). The most significant SNP associated with POF on ECA 7 lies at 8.03 Mb, within a noncoding region, with no candidate genes in the associated genomic region.\\u003c/p\\u003e \\u003cp\\u003eOn ECA 6, the most significant associated SNP also lies in a noncoding region, with fibroblast growth factors 6 and 23 (\\u003cem\\u003eFGF6\\u003c/em\\u003e and \\u003cem\\u003eFGF23\\u003c/em\\u003e) identified as candidate genes within the associated genomic region. FGFs are spatiotemporally expressed during skeletal development, regulating cellular functions and extracellular matrix proteins metabolism [\\u003cspan additionalcitationids=\\\"CR76\\\" citationid=\\\"CR75\\\" class=\\\"CitationRef\\\"\\u003e75\\u003c/span\\u003e\\u0026ndash;\\u003cspan citationid=\\\"CR77\\\" class=\\\"CitationRef\\\"\\u003e77\\u003c/span\\u003e]. FGF23 regulates bone mineralization, parathyroid hormone secretion, and phosphate/vitamin D balance. Its dysregulation is associated with skeletal diseases characterized by defective bone mineralization, such as hypophosphatemic rickets and hyperphosphatemic calcinosis [\\u003cspan additionalcitationids=\\\"CR77 CR78 CR79\\\" citationid=\\\"CR76\\\" class=\\\"CitationRef\\\"\\u003e76\\u003c/span\\u003e\\u0026ndash;\\u003cspan citationid=\\\"CR80\\\" class=\\\"CitationRef\\\"\\u003e80\\u003c/span\\u003e]. FGF6 contributes to muscle regeneration and bone remodeling [\\u003cspan citationid=\\\"CR76\\\" class=\\\"CitationRef\\\"\\u003e76\\u003c/span\\u003e, \\u003cspan citationid=\\\"CR81\\\" class=\\\"CitationRef\\\"\\u003e81\\u003c/span\\u003e].\\u003c/p\\u003e \\u003cp\\u003eOn ECA 20, the most significant associated SNP is located within \\u003cem\\u003eKHDRBS2\\u003c/em\\u003e (KH RNA binding domain containing, signal transduction associated 2), which has no evident link to osteochondral disorders. Within the associated genomic region, interleukin 17A (\\u003cem\\u003eIL17A\\u003c/em\\u003e), interleukin 17F (\\u003cem\\u003eIL17F\\u003c/em\\u003e), translocation associated membrane protein 2 \\u003cem\\u003e(TRAM2)\\u003c/em\\u003e, and bone morphogenetic protein 5 (\\u003cem\\u003eBMP5\\u003c/em\\u003e) were identified as candidate genes. \\u003cem\\u003eL17A\\u003c/em\\u003e and \\u003cem\\u003eIL17F\\u003c/em\\u003e encode cytokines involved in bone repair and cartilage matrix turnover and have been implicated in the pathogenesis of spondylarthritis [\\u003cspan additionalcitationids=\\\"CR83\\\" citationid=\\\"CR82\\\" class=\\\"CitationRef\\\"\\u003e82\\u003c/span\\u003e\\u0026ndash;\\u003cspan citationid=\\\"CR84\\\" class=\\\"CitationRef\\\"\\u003e84\\u003c/span\\u003e]. \\u003cem\\u003eTRAM2\\u003c/em\\u003e encodes a component of the endoplasmic reticulum translocon, a protein channel that regulates collagen I translocation and calcium transport, thereby supporting bone formation [\\u003cspan additionalcitationids=\\\"CR86\\\" citationid=\\\"CR85\\\" class=\\\"CitationRef\\\"\\u003e85\\u003c/span\\u003e\\u0026ndash;\\u003cspan citationid=\\\"CR87\\\" class=\\\"CitationRef\\\"\\u003e87\\u003c/span\\u003e]. \\u003cem\\u003eBMP5\\u003c/em\\u003e encodes a bone morphogenetic protein important for bone formation, growth, remodeling, and repair and has been linked to various genetic skeletal disorders, including the short-ear mouse phenotype [\\u003cspan additionalcitationids=\\\"CR89\\\" citationid=\\\"CR88\\\" class=\\\"CitationRef\\\"\\u003e88\\u003c/span\\u003e\\u0026ndash;\\u003cspan citationid=\\\"CR90\\\" class=\\\"CitationRef\\\"\\u003e90\\u003c/span\\u003e].\\u003c/p\\u003e \\u003cp\\u003eOn ECA 24, the QTL associated with POF (43.66\\u0026ndash;44.61 Mb) overlapped with the QTL associated with hock OC (43.66\\u0026ndash;45.03 Mb). Within this QTL four closely located SNPs at 43.66, 44.08, 44.24, and 44.45 MB reached genome wide significance, with their association with POF increasing as the phenotype definition became more stringent (Additional file 4: Supplementary Table \\u003cspan refid=\\\"MOESM2\\\" class=\\\"InternalRef\\\"\\u003eS2\\u003c/span\\u003e). The most significant associated SNP (\\u0026minus;\\u0026thinsp;log10(P)\\u0026thinsp;=\\u0026thinsp;13.26) at 44.08 MB lies within protein phosphatase 2 regulatory subunit B'gamma (\\u003cem\\u003ePPP2R5C\\u003c/em\\u003e), which has been implicated in overgrowth syndromes in humans [\\u003cspan citationid=\\\"CR91\\\" class=\\\"CitationRef\\\"\\u003e91\\u003c/span\\u003e]. No other candidate genes were found in the associated genomic region. Similar to the other candidate genes (\\u003cem\\u003eFGF6, FGF23, IL17A, IL17F, TRAM2\\u003c/em\\u003e, and \\u003cem\\u003eBMP5)\\u003c/em\\u003e, \\u003cem\\u003ePPP2R5C\\u003c/em\\u003e can be linked to bone development and remodelling processes. Given that POF is proposed to result from weakened bone and pathological avulsion fractures [\\u003cspan citationid=\\\"CR46\\\" class=\\\"CitationRef\\\"\\u003e46\\u003c/span\\u003e, \\u003cspan citationid=\\\"CR92\\\" class=\\\"CitationRef\\\"\\u003e92\\u003c/span\\u003e], all these genes represent plausible candidate genes contributing to its pathogenesis.\\u003c/p\\u003e \\u003c/div\\u003e \\u003cdiv id=\\\"Sec15\\\" class=\\\"Section2\\\"\\u003e \\u003ch2\\u003eStifle osteochondrosis, fetlock osteochondrosis, and dorsoproximal osteochondral fragments\\u003c/h2\\u003e \\u003cp\\u003eIn the GWAS for stifle OC (Additional file 3: Supplementary Figure S8), fetlock OC (Additional file 3: Supplementary Figure S10), and DOF (Additional file 3: Supplementary Figure S14), none of the SNPs met the established criteria to be classified as a significant signal. To our knowledge, this is the first published GWAS conducted for DOF. The absence of significant associations likely reflects the complex, polygenic nature of these traits, characterized by numerous small-effect loci that require larger sample sizes for detection. For DOF, as with POF, fragment removal leaving no radiological trace introduces the risk of false-negative diagnoses, adding to the challenge [\\u003cspan citationid=\\\"CR2\\\" class=\\\"CitationRef\\\"\\u003e2\\u003c/span\\u003e]. Furthermore, as reviewed by Van Mol et al. [\\u003cspan citationid=\\\"CR1\\\" class=\\\"CitationRef\\\"\\u003e1\\u003c/span\\u003e], osteochondral disorders are influenced by various environmental factors, which may overshadow genetic predisposition for some phenotypes and mask genetic associations if not adequately accounted for.\\u003c/p\\u003e \\u003c/div\\u003e \\u003cdiv id=\\\"Sec16\\\" class=\\\"Section2\\\"\\u003e \\u003ch2\\u003eSingle nucleotide polymorphism based heritability estimates\\u003c/h2\\u003e \\u003cp\\u003eThe higher heritability estimates for hock OC and DIRT OC compared to other OC types are consistent with findings from previous studies [\\u003cspan citationid=\\\"CR57\\\" class=\\\"CitationRef\\\"\\u003e57\\u003c/span\\u003e, \\u003cspan citationid=\\\"CR93\\\" class=\\\"CitationRef\\\"\\u003e93\\u003c/span\\u003e, \\u003cspan citationid=\\\"CR94\\\" class=\\\"CitationRef\\\"\\u003e94\\u003c/span\\u003e]. A limitation of our approach is that SEs increase as the phenotype definition become stricter (levels A -\\u0026gt; D), due to reduced numbers of cases and controls. In the literature, SEs ranged between 0.11\\u0026ndash;0.15 for OC at different anatomical locations [\\u003cspan citationid=\\\"CR57\\\" class=\\\"CitationRef\\\"\\u003e57\\u003c/span\\u003e, \\u003cspan citationid=\\\"CR94\\\" class=\\\"CitationRef\\\"\\u003e94\\u003c/span\\u003e], which is comparable to our SEs at strictness level A across all phenotype definitions ranging between 0.10\\u0026ndash;0.12. At stricter levels (B - D), SEs ranges between 0.12\\u0026ndash;0.28, reducing the reliability of these estimates, and highlighting the need for larger genotype datasets to achieve more precise estimates. The wide variability observed in the heritability of DOF and POF may also reflect potential misclassification of cases as controls, as previously discussed. The higher heritability of hock and DIRT OC suggests that selective breeding programs could more effectively reduce their incidence, making these lesions potentially higher-priority targets for breeding programs than total OC, stifle OC, or fetlock OC. However, caution is warranted given the still relatively high SEs of the estimates. In contrast, the low heritability estimates for stifle and fetlock OC indicate that additive genetic variance explains a smaller proportion of phenotypic variance, implying that environmental factors and/or genotype-environment interactions may play a more substantial role in disease risk. This finding aligns with the absence of significant GWAS signals for stifle and fetlock OC.\\u003c/p\\u003e \\u003c/div\\u003e \\u003cdiv id=\\\"Sec17\\\" class=\\\"Section2\\\"\\u003e \\u003ch2\\u003eImplications and future research\\u003c/h2\\u003e \\u003cp\\u003eThe semi-quantitative scoring system presented in this study could serve as a standardized phenotyping protocol for radiographic screenings of horses presented for studbook admission. Consistent international implementation across studbooks would generate high-quality, comparable phenotypes across populations, which are essential for reliable genetic analyses and, ultimately, genetic selection. Pooling data across studbooks to form multi-population reference datasets will improve the cost\\u0026ndash;benefit ratio of investing in genomic evaluations for OC and fetlock osteochondral fragments. Such datasets would also provide a solid foundation for integrating genomic information into breeding programs\\u003c/p\\u003e \\u003cp\\u003eAn advantage of GWAS is its SNP-based framework, which enables the estimation of SNP-based heritabilities. Incorporating SNP genotyping into selection protocols would further expand datasets and, by reducing SEs, improve the precision of heritability estimates. These estimates will help identifying which conditions are meaningful targets for selection. Notably, horses affected by stifle and/or fetlock OC, which according to the current study show relatively low heritability, are currently excluded from breeding by studbooks. If larger datasets in future research confirm these low heritabilities, this would suggest that excluding horses based on such lesions results in a considerable loss of genetic diversity while offering only limited benefit in reducing disease prevalence.\\u003c/p\\u003e \\u003cp\\u003eA practical implementation would proceed in several steps. First, harmonize and apply the scoring system across participating studbooks. Second, collect standardized phenotypes and representative genotypes to establish a multi-population reference panel. Third, estimate SNP-based heritabilities and genomic breeding values. Fourth, after appropriate validation and cost\\u0026ndash;benefit analyses, integrate genomic information into breeding protocols. In parallel, future research should prioritize targeted follow-up of the nominated loci. Functional validation of the candidate genes, both from this study and future studies, will be essential to confirm causal relationships. Ultimately, harmonized phenotyping, larger multi-population datasets, and integrated genomic approaches will result in evidence-based breeding strategies that reduce disease prevalence while preserving genetic diversity.\\u003c/p\\u003e \\u003c/div\\u003e\"},{\"header\":\"Conclusion\",\"content\":\"\\u003cp\\u003eAcross 23 case\\u0026ndash;control GWAS of equine osteochondrosis and fetlock osteochondral fragments in a population of 716 horses, we identified 214 suggestive SNPs, 50 of which reached genome-wide significance. Using predefined criteria, 8 genomic regions of interest were identified, leading to the discovery of 11 novel candidate genes. For total OC, a region on ECA 3 was identified containing the candidate gene \\u003cem\\u003eLDB2\\u003c/em\\u003e, which is a regulator of angiogenesis. For hock OC, novel loci on ECA 20 and ECA 24 were identified harbouring candidate genes \\u003cem\\u003eSCUBE3\\u003c/em\\u003e, \\u003cem\\u003eMAPK13\\u003c/em\\u003e, and \\u003cem\\u003eMAPK14\\u003c/em\\u003e, which are important for skeletogenesis. Further refinement to OC at the distal intermediate ridge revealed a genomic region of interest on ECA 29. For POF, a total of four genomic regions on ECA 6, 7, 20 and 24 were identified, pointing to \\u003cem\\u003eFGF6\\u003c/em\\u003e, \\u003cem\\u003eFGF23\\u003c/em\\u003e, \\u003cem\\u003eIL17A\\u003c/em\\u003e, \\u003cem\\u003eIL17F\\u003c/em\\u003e, \\u003cem\\u003eTRAM2\\u003c/em\\u003e, \\u003cem\\u003eBMP5\\u003c/em\\u003e, and \\u003cem\\u003ePPP2R5C\\u003c/em\\u003e as plausible candidate genes due to their link with bone development and remodelling. In contrast, for stifle OC, fetlock OC, and DOF, none of the SNPs met the established significance criteria, reflecting the likely complex polygenic nature and possible overshadowing environmental factors for these traits. Importantly, this study represents the first published GWAS conducted for DOF. Notably, previously reported candidate genes could not be reaffirmed in this study, potentially due to differences in study populations and/or phenotype definitions. SNP-based heritability estimates varied considerably among phenotypes, with moderate estimates for total OC (26.2\\u0026ndash;27.1%), high estimates for hock and DIRT OC (49.4\\u0026ndash;61.2%), low estimates for stifle and fetlock OC (0.0-14.2%), and highly variable estimates for POF and DOF (5.7\\u0026ndash;54.8%). Our findings demonstrate the importance of stringent lesion-specific phenotypes, as even slight differences in phenotype definition strictness within the same population can substantially affect the detection of genomic associations, identification of candidate genes and estimation of heritability.\\u003c/p\\u003e\"},{\"header\":\"Declarations\",\"content\":\"\\u003cp\\u003e\\u003cstrong\\u003eEthics approval and consent to participate\\u003c/strong\\u003e\\u003c/p\\u003e\\n\\u003cp\\u003eAll protocols were approved by the Ethical Committee of the Faculty of Veterinary Medicine, Ghent University, Belgium (EC 2021/060). Informed consent was obtained from the horse owners prior to the inclusion of their horses in the study.\\u0026nbsp;\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cstrong\\u003eConsent for publication\\u003c/strong\\u003e\\u003c/p\\u003e\\n\\u003cp\\u003eNot applicable.\\u0026nbsp;\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cstrong\\u003eAvailability of data and materials\\u003c/strong\\u003e\\u003c/p\\u003e\\n\\u003cp\\u003eThe data that support the findings of this study are available from the Department of Large Animal Surgery, Anaesthesia and Orthopaedics, Faculty of Veterinary Medicine, Ghent University, and the Center for Animal Breeding and Genetics, Department of Biosystems, KU Leuven, but restrictions apply to the availability of these data, which were used under license for the current study and are therefore not publicly available. Data are, however, available from the authors upon reasonable request and with permission from the Department of Large Animal Surgery, Anaesthesia and Orthopaedics, Faculty of Veterinary Medicine, Ghent University, and the Center for Animal Breeding and Genetics, Department of Biosystems, KU Leuven.\\u003cstrong\\u003e\\u0026nbsp;\\u003c/strong\\u003e\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cstrong\\u003eCompeting interests\\u003c/strong\\u003e\\u003c/p\\u003e\\n\\u003cp\\u003eThe authors declare that they have no competing interests. \\u0026nbsp;\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cstrong\\u003eFunding\\u003c/strong\\u003e\\u003c/p\\u003e\\n\\u003cp\\u003eBram Van Mol received a PhD fellowship in fundamental research from The Research Foundation \\u0026ndash; Flanders (FWO-FR 11B3921N). The genotyping was funded by Paardenpunt Vlaanderen as part of the \\u0026lsquo;PaardenGenomica 2020\\u0026rsquo; project.\\u0026nbsp;\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cstrong\\u003eAuthors\\u0026apos; contributions\\u003c/strong\\u003e\\u003c/p\\u003e\\n\\u003cp\\u003eBVM collected, analysed, and interpreted the data and wrote the manuscript. SJ, RM, and LC contributed to the study design and critically reviewed the statistical analyses. SJ, RM, LC and NB critically reviewed the manuscript. All authors read and approved the final manuscript. \\u0026nbsp;\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cstrong\\u003eAcknowledgements\\u003c/strong\\u003e\\u003c/p\\u003e\\n\\u003cp\\u003eThe authors would like to thank the Belgian Warmblood (BWP) studbook, Paardenpunt Vlaanderen and the Belgian Agency for Agriculture and Fisheries for their cooperation.\\u003cstrong\\u003e\\u003cbr\\u003e\\u0026nbsp;\\u003c/strong\\u003e\\u003c/p\\u003e\"},{\"header\":\"References\",\"content\":\"\\u003col\\u003e\\n \\u003cli\\u003eVan Mol B, Oosterlinck M, Janssens S, Buys N, Pille F. Environmental factors of equine osteochondrosis and fetlock osteochondral fragments: A scoping review \\u0026ndash; Part 1. Vet J. 2024;308:106249. doi: https://doi.org/10.1016/j.tvjl.2024.106249.\\u003c/li\\u003e\\n \\u003cli\\u003eVan Mol B, Janssens S, Oosterlinck M, Pille F, Buys N. Genetic factors of equine osteochondrosis and fetlock osteochondral fragments: A scoping review - Part 2. Vet J. 2024;308:106258. doi: https://doi.org/10.1016/j.tvjl.2024.106258.\\u003c/li\\u003e\\n \\u003cli\\u003eDeclercq J, Martens A, Maes D, Boussauw B, Forsyth R, Boening KJ. Dorsoproximal proximal phalanx osteochondral fragmentation in 117 Warmblood horses. Vet Comp Orthop Traumatol. 2009;22(1):1-6. doi: 10.3415/vcot-08-02-0016.\\u003c/li\\u003e\\n \\u003cli\\u003eBrink P, Dolvik NI, Tverdal A. Lameness and effusion of the tarsocrural joints after arthroscopy of osteochondritis dissecans in horses. Vet Rec. 2009;165(24):709-12.\\u003c/li\\u003e\\n \\u003cli\\u003evan Weeren PR. Chapter 89 - Osteochondritis Dissecans. In: Auer JA, Stick JA, K\\u0026uuml;mmerle JM, Prange T, editors. Equine Surgery. 5th ed. Philadelphia, PA: Elsevier Saunders; 2019. p. 1509-28.\\u003c/li\\u003e\\n \\u003cli\\u003eVerwilghen DR, Janssens S, Busoni V, Pille F, Johnston C, Serteyn D. Do developmental orthopaedic disorders influence future jumping performances in Warmblood stallions? Equine Vet J. 2013;45(5):578-81. doi: 10.1111/evj.12027.\\u003c/li\\u003e\\n \\u003cli\\u003evan Grevenhof EM, Ducro BJ, van Weeren PR, van Tartwijk J, van den Belt AJ, Bijma P. Prevalence of various radiographic manifestations of osteochondrosis and their correlations between and within joints in Dutch Warmblood horses. Equine Vet J. 2009;41(1):11-6. doi: 10.2746/042516408x334794.\\u003c/li\\u003e\\n \\u003cli\\u003eArnan P, Hertsch B. R\\u0026ouml;ntgenologische Untersuchung zur Erfassung der Osteochondrosis dissecans in Fessel, Sprung-und Kniegelenk in Vergleich von Fohlen zum Zweij\\u0026auml;hrigen. Proceedings G\\u0026ouml;ttinger Pferdetage. 2004;4:115-24.\\u003c/li\\u003e\\n \\u003cli\\u003eStock KF, Hamann H, Distl O. Prevalence of osseous fragments in distal and proximal interphalangeal, metacarpo- and metatarsophalangeal and tarsocrural joints of Hanoverian Warmblood horses. J Vet Med A Physiol Pathol Clin Med. 2005;52(8):388-94. doi: 10.1111/j.1439-0442.2005.00753.x.\\u003c/li\\u003e\\n \\u003cli\\u003eVos NJ. Incidence of osteochondrosis (dissecans) in Dutch warmblood horses presented for pre-purchase examination. Irish Veterinary Journal. 2008;61(1):33. doi: 10.1186/2046-0481-61-1-33.\\u003c/li\\u003e\\n \\u003cli\\u003eHilla D, Distl O. Heritabilities and genetic correlations between fetlock, hock and stifle osteochondrosis and fetlock osteochondral fragments in Hanoverian Warmblood horses. Journal of Animal Breeding and Genetics. 2014;131(1):71-81. doi: 10.1111/jbg.12062.\\u003c/li\\u003e\\n \\u003cli\\u003eJonsson L, Dalin G, Egenvall A, Nasholm A, Roepstorff L, Philipsson J. Equine hospital data as a source for study of prevalence and heritability of osteochondrosis and palmar/plantar osseous fragments of Swedish Warmblood horses. Equine Vet J. 2011;43(6):695-700. doi: 10.1111/j.2042-3306.2010.00354.x.\\u003c/li\\u003e\\n \\u003cli\\u003eOrtved KF. Surgical Management of Osteochondrosis in Foals. Vet Clin N Am, Equine Pract. 2017;33(2):379-96. doi: https://doi.org/10.1016/j.cveq.2017.03.010.\\u003c/li\\u003e\\n \\u003cli\\u003eNaccache F, Metzger J, Distl O. Genetic risk factors for osteochondrosis in various horse breeds. Equine Vet J. 2018;50(5):556-63. doi: 10.1111/evj.12824.\\u003c/li\\u003e\\n \\u003cli\\u003eMcCoy AM, Norton EM, Kemper AM, Beeson SK, Mickelson JR, McCue ME. SNP-based heritability and genetic architecture of tarsal osteochondrosis in North American Standardbred horses. Anim Genet. 2019;50(1):78-81. doi: 10.1111/age.12738.\\u003c/li\\u003e\\n \\u003cli\\u003eRussell J, Matika O, Russell T, Reardon RJM. Heritability and prevalence of selected osteochondrosis lesions in yearling Thoroughbred horses. Equine Vet J. 2017;49(3):282-7. doi: https://doi.org/10.1111/evj.12613.\\u003c/li\\u003e\\n \\u003cli\\u003eMcIlwraith CW, Foerner JJ, Davis DM. Osteochondritis dissecans of the tarsocrural joint: results of treatment with arthroscopic surgery. Equine Vet J. 1991;23(3):155-62. doi: 10.1111/j.2042-3306.1991.tb02746.x.\\u003c/li\\u003e\\n \\u003cli\\u003eMcIlwraith CW. Inferences from referred clinical cases of osteochondritis dissecans. Equine Vet J. 1993;25(S16):27-30. doi: https://doi.org/10.1111/j.2042-3306.1993.tb04850.x.\\u003c/li\\u003e\\n \\u003cli\\u003eHauspie S, Martens A, Declercq J, Busoni V, Vanderperren K, van Bree H, et al. Radiographic features of the dorsal condylar sagittal ridge of the third metacarpal and metatarsal bones in young Warmblood stallions. Veterinary and Comparative Orthopaedics and Traumatology. 2010;23(6):411-6. doi: 10.3415/vcot-09-12-0130.\\u003c/li\\u003e\\n \\u003cli\\u003eHauspie S, Forsyth R, Vanderperren K, Declercq J, Martens A, Saunders JH. The histological appearance of the proximal aspect of the dorsal condylar sagittal ridge of the third metacarpal and metatarsal bone in young warmblood horses: normal appearance and correlation with detected radiographic variations. Anat Histol Embryol. 2013;42(3):232-8. doi: 10.1111/ahe.12006.\\u003c/li\\u003e\\n \\u003cli\\u003eChang CC, Chow CC, Tellier LC, Vattikuti S, Purcell SM, Lee JJ. Second-generation PLINK: rising to the challenge of larger and richer datasets. Gigascience. 2015;4:7. doi: 10.1186/s13742-015-0047-8.\\u003c/li\\u003e\\n \\u003cli\\u003ePurcell S, Neale B, Todd-Brown K, Thomas L, Ferreira MA, Bender D, et al. PLINK: a tool set for whole-genome association and population-based linkage analyses. American journal of human genetics. 2007;81(3):559-75. doi: 10.1086/519795.\\u003c/li\\u003e\\n \\u003cli\\u003eR Core Team R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria. 2023.\\u003c/li\\u003e\\n \\u003cli\\u003eAnderson CA, Pettersson FH, Clarke GM, Cardon LR, Morris AP, Zondervan KT. Data quality control in genetic case-control association studies. Nature Protocols. 2010;5(9):1564-73. doi: 10.1038/nprot.2010.116.\\u003c/li\\u003e\\n \\u003cli\\u003eLander E, Kruglyak L. Genetic dissection of complex traits: guidelines for interpreting and reporting linkage results. Nature Genetics. 1995;11(3):241-7. doi: 10.1038/ng1195-241.\\u003c/li\\u003e\\n \\u003cli\\u003eDudbridge F, Gusnanto A. Estimation of significance thresholds for genomewide association scans. Genetic Epidemiology. 2008;32(3):227-34. doi: https://doi.org/10.1002/gepi.20297.\\u003c/li\\u003e\\n \\u003cli\\u003eLi MX, Yeung JM, Cherny SS, Sham PC. Evaluating the effective numbers of independent tests and significant p-value thresholds in commercial genotyping arrays and public imputation reference datasets. Hum Genet. 2012;131(5):747-56. doi: 10.1007/s00439-011-1118-2.\\u003c/li\\u003e\\n \\u003cli\\u003eYang J, Lee SH, Goddard ME, Visscher PM. GCTA: a tool for genome-wide complex trait analysis. The American Journal of Human Genetics. 2011;88(1):76-82.\\u003c/li\\u003e\\n \\u003cli\\u003eJiang L, Zheng Z, Qi T, Kemper KE, Wray NR, Visscher PM, et al. A resource-efficient tool for mixed model association analysis of large-scale data. Nat Genet. 2019;51(12):1749-55. doi: 10.1038/s41588-019-0530-8.\\u003c/li\\u003e\\n \\u003cli\\u003eTurner S. qqman: an R package for visualizing GWAS results using Q-Q and manhattan plots. Journal of Open Source Software. 2018;3:731. doi: 10.21105/joss.00731.\\u003c/li\\u003e\\n \\u003cli\\u003eHarrison PW, Amode MR, Austine-Orimoloye O, Azov Andrey G, Barba M, Barnes I, et al. Ensembl 2024. Nucleic Acids Research. 2023;52(D1):D891-D9. doi: 10.1093/nar/gkad1049.\\u003c/li\\u003e\\n \\u003cli\\u003eKinsella RJ, K\\u0026auml;h\\u0026auml;ri A, Haider S, Zamora J, Proctor G, Spudich G, et al. Ensembl BioMarts: a hub for data retrieval across taxonomic space. Database (Oxford). 2011;2011:bar030. doi: 10.1093/database/bar030.\\u003c/li\\u003e\\n \\u003cli\\u003eHinrichs AS, Karolchik D, Baertsch R, Barber GP, Bejerano G, Clawson H, et al. The UCSC Genome Browser Database: update 2006. Nucleic Acids Research. 2006;34(suppl_1):D590-D8. doi: 10.1093/nar/gkj144.\\u003c/li\\u003e\\n \\u003cli\\u003eHenson FM, Schofield PN, Jeffcott LB. Expression of transforming growth factor-beta 1 in normal and dyschondroplastic articular growth cartilage of the young horse. Equine Vet J. 1997;29(6):434-9. doi: 10.1111/j.2042-3306.1997.tb03155.x.\\u003c/li\\u003e\\n \\u003cli\\u003eSemevolos SA, Strassheim ML, Haupt JL, Nixon AJ. Expression patterns of hedgehog signaling peptides in naturally acquired equine osteochondrosis. J Orthop Res. 2005;23(5):1152-9. doi: 10.1016/j.orthres.2005.01.024.\\u003c/li\\u003e\\n \\u003cli\\u003eMirams M, Tatarczuch L, Ahmed YA, Pagel CN, Jeffcott LB, Davies HMS, et al. Altered Gene Expression in Early Osteochondrosis Lesions. Journal of Orthopaedic Research. 2009;27(4):452-7. doi: 10.1002/jor.20761.\\u003c/li\\u003e\\n \\u003cli\\u003eRiddick TL, Duesterdieck-Zellmer K, Semevolos SA. Gene and protein expression of cartilage canal and osteochondral junction chondrocytes and full-thickness cartilage in early equine osteochondrosis. Vet J. 2012;194(3):319-25. doi: https://doi.org/10.1016/j.tvjl.2012.04.023.\\u003c/li\\u003e\\n \\u003cli\\u003eKinsley MA, Semevolos SA, Duesterdieck-Zellmer KF. Wnt/-catenin signaling of cartilage canal and osteochondral junction chondrocytes and full thickness cartilage in early equine osteochondrosis. Journal of Orthopaedic Research. 2015;33(10):1433-8. doi: 10.1002/jor.22846.\\u003c/li\\u003e\\n \\u003cli\\u003eMirams M, Ayodele BA, Tatarczuch L, Henson FM, Pagel CN, Mackie EJ. Identification of novel osteochondrosis Associated genes. Journal of Orthopaedic Research. 2016;34(3):404-11. doi: 10.1002/jor.23033.\\u003c/li\\u003e\\n \\u003cli\\u003eGrissom SK, Semevolos SA, Duesterdieck-Zellmer K. Role of cartilage and bone matrix regulation in early equine osteochondrosis. Bone Reports. 2023;18. doi: 10.1016/j.bonr.2023.101653.\\u003c/li\\u003e\\n \\u003cli\\u003eAustbo L, Roed KH, Dolvik NI, Skretting G. Identification of differentially expressed genes associated with osteochondrosis in standardbred horses using rna arbitrarily primed PCR. Animal Biotechnology. 2010;21(2):135-9. doi: 10.1080/10495391003608316.\\u003c/li\\u003e\\n \\u003cli\\u003eDrabbe A, Janssens S, Blott S, Ducro BJ, Fontanel M, Francois L, et al. Genome-Wide Association Analyses of Osteochondrosis in Belgian Warmbloods Reveal Candidate Genes Associated With Chondrocyte Development. Journal of Equine Veterinary Science. 2022;111. doi: 10.1016/j.jevs.2022.103870.\\u003c/li\\u003e\\n \\u003cli\\u003eChoi HJ, Rho SS, Choi DH, Kwon YG. LDB2 regulates the expression of DLL4 through the formation of oligomeric complexes in endothelial cells. BMB Rep. 2018;51(1):21-6. doi: 10.5483/bmbrep.2018.51.1.140.\\u003c/li\\u003e\\n \\u003cli\\u003eCarlson CS, Cullins LD, Meuten DJ. Osteochondrosis of the articular-epiphyseal cartilage complex in young horses: evidence for a defect in cartilage canal blood supply. Vet Pathol. 1995;32(6):641-7. doi: 10.1177/030098589503200605.\\u003c/li\\u003e\\n \\u003cli\\u003eJeffcott L. Developmental diseases affecting growing horses. 2nd European Workshop on Equine Nutrition. ENESAD, Dijon, FRANCE2004. p. 243-55.\\u003c/li\\u003e\\n \\u003cli\\u003eOlstad K, Ekman S, Carlson CS. An Update on the Pathogenesis of Osteochondrosis. Vet Pathol. 2015;52(5):785-802. doi: 10.1177/0300985815588778.\\u003c/li\\u003e\\n \\u003cli\\u003eYtrehus B, Carlson CS, Ekman S. Etiology and Pathogenesis of Osteochondrosis. Vet Pathol. 2007;44(4):429-48. doi: 10.1354/vp.44-4-429.\\u003c/li\\u003e\\n \\u003cli\\u003eOlstad K, Hendrickson EHS, Carlson CS, Ekman S, Dolvik NI. Transection of vessels in epiphyseal cartilage canals leads to osteochondrosis and osteochondrosis dissecans in the femoro-patellar joint of foals; a potential model of juvenile osteochondritis dissecans. Osteoarthritis and Cartilage. 2013;21(5):730-8. doi: 10.1016/j.joca.2013.02.005.\\u003c/li\\u003e\\n \\u003cli\\u003eDik KJ, Enzerink E, van Weeren PR. Radiographic development of osteochondral abnormalities, in the hock and stifle of Dutch Warmblood foals, from age 1 to 11 months. Equine Vet J Suppl. 1999(31):9-15. doi: 10.1111/j.2042-3306.1999.tb05308.x.\\u003c/li\\u003e\\n \\u003cli\\u003eSantschi EM, Prichard MA, Whitman JL, Batten CA, Strathman TA, Canada NC, et al. Stifle radiography in Thoroughbreds from 6 to 18 months of age. Equine Vet Educ. 2020;32:78-84. doi: 10.1111/eve.13194.\\u003c/li\\u003e\\n \\u003cli\\u003eGr\\u0026Oslash;ndahl AM. The incidence of osteochondrosis in the tibiotarsal joint of norwegian standardbred trotters. Journal of Equine Veterinary Science. 1991;11(5):272-4. doi: https://doi.org/10.1016/S0737-0806(06)81313-4.\\u003c/li\\u003e\\n \\u003cli\\u003eCarlsten J, Sandgren B, Dalin G. Development of osteochondrosis in the tarsocrural joint and osteochondral fragments in the fetlock joints of Standardbred trotters. I. A radiological survey. Equine Vet J. 1993;25(S16):42-7. doi: https://doi.org/10.1111/j.2042-3306.1993.tb04853.x.\\u003c/li\\u003e\\n \\u003cli\\u003eBaccarin RYA, Pereira MA, Roncati NV, Bergamaschi RRC, Hagen SCF. Development of osteochondrosis in Lusitano foals: A radiographic study. Canadian Veterinary Journal-Revue Veterinaire Canadienne. 2012;53(10):1079-84.\\u003c/li\\u003e\\n \\u003cli\\u003eJacquet S, Robert C, Valette JP, Denoix JM. Evolution of radiological findings detected in the limbs of 321 young horses between the ages of 6 and 18months. Vet J. 2013;197(1):58-64. doi: https://doi.org/10.1016/j.tvjl.2013.03.042.\\u003c/li\\u003e\\n \\u003cli\\u003eStromberg B. A review of the salient features of osteochondrosis in the horse. Equine Vet J. 1979;11(4):211-4. doi: 10.1111/j.2042-3306.1979.tb01346.x.\\u003c/li\\u003e\\n \\u003cli\\u003eOrr N, Hill EW, Gu J, Govindarajan P, Conroy J, van Grevenhof EM, et al. Genome-wide association study of osteochondrosis in the tarsocrural joint of Dutch Warmblood horses identifies susceptibility loci on chromosomes 3 and 10. Anim Genet. 2013;44(4):408-12. doi: 10.1111/age.12016.\\u003c/li\\u003e\\n \\u003cli\\u003eTeyssedre S, Dupuis MC, Guerin G, Schibler L, Denoix JM, Elsen JM, et al. Genome-wide association studies for osteochondrosis in French Trotter horses. Journal of Animal Science. 2012;90(1):45-53. doi: 10.2527/jas.2011-4031.\\u003c/li\\u003e\\n \\u003cli\\u003eLykkjen S, Dolvik NI, McCue ME, Rendahl AK, Mickelson JR, Roed KH. Genome-wide association analysis of osteochondrosis of the tibiotarsal joint in Norwegian Standardbred trotters. Anim Genet. 2010;41:111-20. doi: 10.1111/j.1365-2052.2010.02117.x.\\u003c/li\\u003e\\n \\u003cli\\u003eMcCoy AM, Beeson SK, Splan RK, Lykkjen S, Ralston SL, Mickelson JR, et al. Identification and validation of risk loci for osteochondrosis in standardbreds. Bmc Genomics. 2016;17. doi: 10.1186/s12864-016-2385-z.\\u003c/li\\u003e\\n \\u003cli\\u003eLin Y-C, Sahoo BK, Gau S-S, Yang R-B. The biology of SCUBE. Journal of Biomedical Science. 2023;30(1):33. doi: 10.1186/s12929-023-00925-3.\\u003c/li\\u003e\\n \\u003cli\\u003eWang L, Zhang L, Yan H, Liu X, Li N, Liang J, et al. Genome-wide association studies identify the loci for 5 exterior traits in a Large White \\u0026times; Minzhu pig population. PLoS One. 2014;9(8):e103766. doi: 10.1371/journal.pone.0103766.\\u003c/li\\u003e\\n \\u003cli\\u003eGudbjartsson DF, Walters GB, Thorleifsson G, Stefansson H, Halldorsson BV, Zusmanovich P, et al. Many sequence variants affecting diversity of adult human height. Nat Genet. 2008;40(5):609-15. doi: 10.1038/ng.122.\\u003c/li\\u003e\\n \\u003cli\\u003eLin Y-C, Niceta M, Muto V, Vona B, Pagnamenta AT, Maroofian R, et al. SCUBE3 loss-of-function causes a recognizable recessive developmental disorder due to defective bone morphogenetic protein signaling. The American Journal of Human Genetics. 2021;108(1):115-33. doi: https://doi.org/10.1016/j.ajhg.2020.11.015.\\u003c/li\\u003e\\n \\u003cli\\u003eThouverey C, Caverzasio J. Focus on the p38 MAPK signaling pathway in bone development and maintenance. Bonekey Rep. 2015;4:711. doi: 10.1038/bonekey.2015.80.\\u003c/li\\u003e\\n \\u003cli\\u003eRalston SL. Hyperglycemia/hyperinsulinemia after feeding a meal of grain to young horses with osteochondritis dissecans (OCD) lesions. Pferdeheilkunde Equine Medicine. 1996;12:320-2.\\u003c/li\\u003e\\n \\u003cli\\u003ePagan J, Geor R, Caddel S, Pryor P, Hoekstra K. The relationship between glycemic response and the incidence of OCD in Thoroughbred weanlings: a field study. Proc Am Ass equine Practnrs. 2001;47.\\u003c/li\\u003e\\n \\u003cli\\u003eHarris P, Staniar W, Ellis AD. Effect of exercise and diet on the incidence of DOD. The Growing Horse: Nutrition and Prevention of Growth Disorders Wageningen Academic Publishers, Wageningen, The Netherlands 2nd European Conference on Equine Nutrition Dijon, France2004. p. 273-90.\\u003c/li\\u003e\\n \\u003cli\\u003eStaniar WB. Linking dietary energy and skeletal development in the horse. Revista Brasileira de Zootecnia. 2010;39:138-44.\\u003c/li\\u003e\\n \\u003cli\\u003eHenson FM, Davenport C, Butler L, Moran I, Shingleton WD, Jeffcott LB, et al. Effects of insulin and insulin-like growth factors I and II on the growth of equine fetal and neonatal chondrocytes. Equine Vet J. 1997;29(6):441-7. doi: https://doi.org/10.1111/j.2042-3306.1997.tb03156.x.\\u003c/li\\u003e\\n \\u003cli\\u003eSerteyn D, Mendoza Garc\\u0026iacute;a L, Sandersen C, Caudron I, piquemal D, Chavatte-Palmer P, et al. High plasma concentrations of sclerostin, an inhibitor of the Wnt signaling pathway in young horses affected by osteochondrosis. Open journal of orthepedics. 2014. doi: 10.4236/ojo.2014.412051.\\u003c/li\\u003e\\n \\u003cli\\u003eMarchant EA, Semevolos SA. Differential Protein Expression of the Marginal Transitional Zone in Foals with Osteochondrosis. Journal of Equine Veterinary Science. 2022;116:104055. doi: https://doi.org/10.1016/j.jevs.2022.104055.\\u003c/li\\u003e\\n \\u003cli\\u003eHan J, Wu J, Silke J. An overview of mammalian p38 mitogen-activated protein kinases, central regulators of cell stress and receptor signaling. F1000Res. 2020;9. doi: 10.12688/f1000research.22092.1.\\u003c/li\\u003e\\n \\u003cli\\u003eSemevolos SA, Nixon AJ, Brower-Toland BD. Changes in molecular expression of aggrecan and collagen types I, II, and X, insulin-like growth factor-I, and transforming growth factor-\\u0026beta;1 in articular cartilage obtained from horses with naturally acquired osteochondrosis. Am J Vet Res. 2001;62(7):1088-94. doi: 10.2460/ajvr.2001.62.1088.\\u003c/li\\u003e\\n \\u003cli\\u003eLykkjen S, Dolvik NI, McCue ME, Rendahl AK, Mickelson JR, Roed KH. Equine developmental orthopaedic diseases - a genome-wide association study of first phalanx plantar osteochondral fragments in Standardbred trotters. Anim Genet. 2013;44(6):766-9. doi: 10.1111/age.12064.\\u003c/li\\u003e\\n \\u003cli\\u003eFayad C, Legeai-Mallet L. FGF Signaling: A Key Pathway During Skeletal Development. In: Rossi A, Zaucke F, editors. The Extracellular Matrix in Genetic Skeletal Disorders. Cham: Springer Nature Switzerland; 2024. p. 247-85.\\u003c/li\\u003e\\n \\u003cli\\u003eXie Y, Su N, Yang J, Tan Q, Huang S, Jin M, et al. FGF/FGFR signaling in health and disease. Signal Transduction and Targeted Therapy. 2020;5(1):181. doi: 10.1038/s41392-020-00222-7.\\u003c/li\\u003e\\n \\u003cli\\u003eOrnitz DM, Marie PJ. Fibroblast growth factor signaling in skeletal development and disease. Genes Dev. 2015;29(14):1463-86. doi: 10.1101/gad.266551.115.\\u003c/li\\u003e\\n \\u003cli\\u003eB\\u0026auml;r L, Stournaras C, Lang F, F\\u0026ouml;ller M. Regulation of fibroblast growth factor 23 (FGF23) in health and disease. FEBS Letters. 2019;593(15):1879-900. doi: https://doi.org/10.1002/1873-3468.13494.\\u003c/li\\u003e\\n \\u003cli\\u003eSu N, Jin M, Chen L. Role of FGF/FGFR signaling in skeletal development and homeostasis: learning from mouse models. Bone Research. 2014;2(1):14003. doi: 10.1038/boneres.2014.3.\\u003c/li\\u003e\\n \\u003cli\\u003eChonchol M, Greene T, Zhang Y, Hoofnagle AN, Cheung AK. Low Vitamin D and High Fibroblast Growth Factor 23 Serum Levels Associate with Infectious and Cardiac Deaths in the HEMO Study. J Am Soc Nephrol. 2016;27(1):227-37. doi: 10.1681/asn.2014101009.\\u003c/li\\u003e\\n \\u003cli\\u003eBosetti M, Leigheb M, Brooks RA, Boccafoschi F, Cannas MF. Regulation of osteoblast and osteoclast functions by FGF-6. Journal of Cellular Physiology. 2010;225(2):466-71. doi: https://doi.org/10.1002/jcp.22225.\\u003c/li\\u003e\\n \\u003cli\\u003eHuangfu L, Li R, Huang Y, Wang S. The IL-17 family in diseases: from bench to bedside. Signal Transduction and Targeted Therapy. 2023;8(1):402. doi: 10.1038/s41392-023-01620-3.\\u003c/li\\u003e\\n \\u003cli\\u003eTanigawa S, Aida Y, Kawato T, Honda K, Nakayama G, Motohashi M, et al. Interleukin-17F affects cartilage matrix turnover by increasing the expression of collagenases and stromelysin-1 and by decreasing the expression of their inhibitors and extracellular matrix components in chondrocytes. Cytokine. 2011;56(2):376-86. doi: https://doi.org/10.1016/j.cyto.2011.08.015.\\u003c/li\\u003e\\n \\u003cli\\u003eJo S, Wang SE, Lee YL, Kang S, Lee B, Han J, et al. IL-17A induces osteoblast differentiation by activating JAK2/STAT3 in ankylosing spondylitis. Arthritis Research \\u0026amp; Therapy. 2018;20(1):115. doi: 10.1186/s13075-018-1582-3.\\u003c/li\\u003e\\n \\u003cli\\u003ePregizer S, Barski A, Gersbach CA, Garc\\u0026iacute;a AJ, Frenkel B. Identification of novel Runx2 targets in osteoblasts: Cell type-specific BMP-dependent regulation of Tram2. Journal of Cellular Biochemistry. 2007;102(6):1458-71. doi: https://doi.org/10.1002/jcb.21366.\\u003c/li\\u003e\\n \\u003cli\\u003eZhang J, Ji Y, Jiang S, Shi M, Cai W, Miron RJ, et al. Calcium\\u0026ndash;Collagen Coupling is Vital for Biomineralization Schedule. Advanced Science. 2021;8(15):2100363. doi: https://doi.org/10.1002/advs.202100363.\\u003c/li\\u003e\\n \\u003cli\\u003eJohnson AE, van Waes MA. The Translocon: A Dynamic Gateway at the ER Membrane. Annual Review of Cell and Developmental Biology. 1999;15(Volume 15, 1999):799-842. doi: https://doi.org/10.1146/annurev.cellbio.15.1.799.\\u003c/li\\u003e\\n \\u003cli\\u003eKing JA, Marker PC, Seung KJ, Kingsley DM. BMP5 and the Molecular, Skeletal, and Soft-Tissue Alterations in short ear Mice. Developmental Biology. 1994;166(1):112-22. doi: https://doi.org/10.1006/dbio.1994.1300.\\u003c/li\\u003e\\n \\u003cli\\u003eStorm EE, Kingsley DM. Joint patterning defects caused by single and double mutations in members of the bone morphogenetic protein (BMP) family. Development. 1996;122(12):3969-79. doi: 10.1242/dev.122.12.3969.\\u003c/li\\u003e\\n \\u003cli\\u003eSalazar VS, Gamer LW, Rosen V. BMP signalling in skeletal development, disease and repair. Nature Reviews Endocrinology. 2016;12(4):203-21. doi: 10.1038/nrendo.2016.12.\\u003c/li\\u003e\\n \\u003cli\\u003eLoveday C, Tatton-Brown K, Clarke M, Westwood I, Renwick A, Ramsay E, et al. Mutations in the PP2A regulatory subunit B family genes PPP2R5B, PPP2R5C and PPP2R5D cause human overgrowth. Human Molecular Genetics. 2015;24(17):4775-9. doi: 10.1093/hmg/ddv182.\\u003c/li\\u003e\\n \\u003cli\\u003eMcCoy AM, Toth F, Dolvik NI, Ekman S, Ellermann J, Olstad K, et al. Articular osteochondrosis: a comparison of naturally-occurring human and animal disease. Osteoarthritis and Cartilage. 2013;21(11):1638-47. doi: https://doi.org/10.1016/j.joca.2013.08.011.\\u003c/li\\u003e\\n \\u003cli\\u003evan Grevenhof EM, Schurink A, Ducro BJ, van Weeren PR, van Tartwijk J, Bijma P, et al. Genetic variables of various manifestations of osteochondrosis and their correlations between and within joints in Dutch warmblood horses. Journal of Animal Science. 2009;87(6):1906-12. doi: 10.2527/jas.2008-1199.\\u003c/li\\u003e\\n \\u003cli\\u003eZimmermann E, Distl O. SNP-Based Heritability of Osteochondrosis Dissecans in Hanoverian Warmblood Horses. Animals. 2023;13(9). doi: 10.3390/ani13091462.\\u003c/li\\u003e\\n\\u003c/ol\\u003e\"}],\"fulltextSource\":\"\",\"fullText\":\"\",\"funders\":[],\"hasAdminPriorityOnWorkflow\":false,\"hasManuscriptDocX\":true,\"hasOptedInToPreprint\":true,\"hasPassedJournalQc\":\"\",\"hasAnyPriority\":false,\"hideJournal\":false,\"highlight\":\"\",\"institution\":\"\",\"isAcceptedByJournal\":false,\"isAuthorSuppliedPdf\":false,\"isDeskRejected\":\"\",\"isHiddenFromSearch\":false,\"isInQc\":false,\"isInWorkflow\":false,\"isPdf\":false,\"isPdfUpToDate\":true,\"isWithdrawnOrRetracted\":false,\"journal\":{\"display\":true,\"email\":\"info@researchsquare.com\",\"identity\":\"bmc-genomics\",\"isNatureJournal\":false,\"hasQc\":true,\"allowDirectSubmit\":false,\"externalIdentity\":\"gics\",\"sideBox\":\"Learn more about [BMC Genomics](http://bmcgenomics.biomedcentral.com/)\",\"snPcode\":\"\",\"submissionUrl\":\"https://www.editorialmanager.com/gics\",\"title\":\"BMC Genomics\",\"twitterHandle\":\"#BMCGenomics\",\"acdcEnabled\":true,\"dfaEnabled\":false,\"editorialSystem\":\"em\",\"reportingPortfolio\":\"BMC Series\",\"inReviewEnabled\":true,\"inReviewRevisionsEnabled\":true},\"keywords\":\"Equine Osteochondrosis, Fetlock osteochondral fragments, horse, Genome-wide association study, SNP-based heritability, Joint-specific phenotypes\",\"lastPublishedDoi\":\"10.21203/rs.3.rs-8832466/v1\",\"lastPublishedDoiUrl\":\"https://doi.org/10.21203/rs.3.rs-8832466/v1\",\"license\":{\"name\":\"CC BY 4.0\",\"url\":\"https://creativecommons.org/licenses/by/4.0/\"},\"manuscriptAbstract\":\"\\u003ch2\\u003eBackground\\u003c/h2\\u003e \\u003cp\\u003eEquine osteochondrosis (OC) and fetlock osteochondral fragments are highly prevalent developmental joint disorders caused by failure of the growing skeleton, often leading to intra-articular fragments that require arthroscopic surgery and impose substantial economic and animal welfare burdens. To reduce prevalence, phenotype-based selection has had limited success, prompting interest in genomic approaches. Progress in genetic research has been hindered by difficulties in assembling large study populations, inconsistent phenotype definitions, and the disease\\u0026rsquo;s dynamic nature. This study aimed to perform joint-specific case\\u0026ndash;control genome‐wide association studies (GWAS) and estimate single nucleotide polymorphism (SNP) based heritabilities for OC and fetlock osteochondral fragments in horses. Seven base phenotypes with two to four levels of strictness were defined using a novel semi-quantitative scoring system and paired with high-density genotypes.\\u003c/p\\u003e\\u003ch2\\u003eResults\\u003c/h2\\u003e \\u003cp\\u003eAcross 23 GWAS on 716 horses, 214 suggestive SNPs were identified, including 50 that reached genome-wide significance. Based on predefined criteria, eight genomic regions of interest were identified, revealing 11 novel candidate genes: \\u003cem\\u003eLDB2\\u003c/em\\u003e for total OC; \\u003cem\\u003eSCUBE3\\u003c/em\\u003e, \\u003cem\\u003eMAPK13\\u003c/em\\u003e and \\u003cem\\u003eMAPK14\\u003c/em\\u003e for hock OC; and \\u003cem\\u003eFGF6\\u003c/em\\u003e, \\u003cem\\u003eFGF23\\u003c/em\\u003e, \\u003cem\\u003eIL17A\\u003c/em\\u003e, \\u003cem\\u003eIL17F\\u003c/em\\u003e, \\u003cem\\u003eTRAM2\\u003c/em\\u003e, \\u003cem\\u003eBMP5\\u003c/em\\u003e, and PPP2R5C for palmaro-/plantaroproximal osteochondral fragments of the proximal phalanx (POF). No candidate genes were detected for distal intermediate ridge of the tibia (DIRT) OC. For stifle OC, fetlock OC, and dorsoproximal fragments of the proximal phalanx (DOF), suggestive SNPs did not define genomic regions of interest. SNP-based heritability estimates ranged from 26.2\\u0026ndash;27.1% for total OC; 53.9\\u0026ndash;61.0% for hock OC; 49.4\\u0026ndash;61.2% for DIRT OC; 1.5\\u0026ndash;14.2% for stifle OC; 0.0-13.7% for fetlock OC; 10.0-54.8% for POF; and 5.7\\u0026ndash;39.6% for DOF.\\u003c/p\\u003e\\u003ch2\\u003eConclusion\\u003c/h2\\u003e \\u003cp\\u003eThis study represents the first GWAS on DOF in horses and identifies novel candidate genes involved in angiogenesis and bone development for total OC, hock OC and POF. The results demonstrate the importance of stringent, joint-specific phenotypes, as even minor differences in phenotype definition substantially affected heritability estimates and the detection of genomic associations and candidate genes. The high heritability of hock and DIRT OC indicates considerable potential for rapid genetic improvement, compared to the lower heritability observed for total OC, fetlock OC, and stifle OC.\\u003c/p\\u003e\",\"manuscriptTitle\":\"Leveraging Joint-Specific Phenotypes for Genome-Wide Association Studies and SNP Heritability Estimation of Equine Osteochondrosis and Fetlock Osteochondral Fragments\",\"msid\":\"\",\"msnumber\":\"\",\"nonDraftVersions\":[{\"code\":1,\"date\":\"2026-03-19 16:00:35\",\"doi\":\"10.21203/rs.3.rs-8832466/v1\",\"editorialEvents\":[{\"type\":\"communityComments\",\"content\":0},{\"type\":\"editorInvitedReview\",\"content\":\"\",\"date\":\"2026-04-19T06:20:35+00:00\",\"index\":\"hide\",\"fulltext\":\"\"},{\"type\":\"reviewerAgreed\",\"content\":\"312123552528406936299781753630740174176\",\"date\":\"2026-03-23T08:16:46+00:00\",\"index\":\"hide\",\"fulltext\":\"\"},{\"type\":\"reviewerAgreed\",\"content\":\"73920247055045231156927664254204466745\",\"date\":\"2026-03-18T10:39:01+00:00\",\"index\":\"hide\",\"fulltext\":\"\"},{\"type\":\"reviewersInvited\",\"content\":\"\",\"date\":\"2026-03-16T10:04:33+00:00\",\"index\":\"\",\"fulltext\":\"\"},{\"type\":\"editorAssigned\",\"content\":\"\",\"date\":\"2026-02-23T06:49:40+00:00\",\"index\":\"\",\"fulltext\":\"\"},{\"type\":\"editorInvited\",\"content\":\"\",\"date\":\"2026-02-20T05:17:58+00:00\",\"index\":\"\",\"fulltext\":\"\"},{\"type\":\"checksComplete\",\"content\":\"\",\"date\":\"2026-02-19T15:47:45+00:00\",\"index\":\"\",\"fulltext\":\"\"},{\"type\":\"submitted\",\"content\":\"BMC Genomics\",\"date\":\"2026-02-19T13:22:45+00:00\",\"index\":\"\",\"fulltext\":\"\"}],\"status\":\"published\",\"journal\":{\"display\":true,\"email\":\"info@researchsquare.com\",\"identity\":\"bmc-genomics\",\"isNatureJournal\":false,\"hasQc\":true,\"allowDirectSubmit\":false,\"externalIdentity\":\"gics\",\"sideBox\":\"Learn more about [BMC Genomics](http://bmcgenomics.biomedcentral.com/)\",\"snPcode\":\"\",\"submissionUrl\":\"https://www.editorialmanager.com/gics\",\"title\":\"BMC Genomics\",\"twitterHandle\":\"#BMCGenomics\",\"acdcEnabled\":true,\"dfaEnabled\":false,\"editorialSystem\":\"em\",\"reportingPortfolio\":\"BMC Series\",\"inReviewEnabled\":true,\"inReviewRevisionsEnabled\":true}}],\"origin\":\"\",\"ownerIdentity\":\"4c907ca6-309a-4604-a700-a10922860634\",\"owner\":[],\"postedDate\":\"March 19th, 2026\",\"published\":true,\"recentEditorialEvents\":[],\"rejectedJournal\":[],\"revision\":\"\",\"amendment\":\"\",\"status\":\"under-review\",\"subjectAreas\":[],\"tags\":[],\"updatedAt\":\"2026-03-19T16:00:35+00:00\",\"versionOfRecord\":[],\"versionCreatedAt\":\"2026-03-19 16:00:35\",\"video\":\"\",\"vorDoi\":\"\",\"vorDoiUrl\":\"\",\"workflowStages\":[]},\"version\":\"v1\",\"identity\":\"rs-8832466\",\"journalConfig\":\"researchsquare\"},\"__N_SSP\":true},\"page\":\"/article/[identity]/[[...version]]\",\"query\":{\"redirect\":\"/article/rs-8832466\",\"identity\":\"rs-8832466\",\"version\":[\"v1\"]},\"buildId\":\"XKTyCvWXoU3ODBz1xrDgd\",\"isFallback\":false,\"isExperimentalCompile\":false,\"dynamicIds\":[84888],\"gssp\":true,\"scriptLoader\":[]}","source_license":"CC-BY-4.0","license_restricted":false}