Skip to main content

Genome-wide meta-analysis identified novel variant associated with hallux valgus in Caucasians



Hallux valgus, one of the most common structural foot deformities, is highly heritable. However, previous efforts to elucidate the genetic underpinnings of hallux valgus through a genome-wide association study (GWAS) conducted in 4409 Caucasians did not identify genome-wide significant associations with hallux valgus in both gender-specific and sex-combined GWAS meta-analyses. In this analysis, we add newly available data and more densely imputed genotypes to identify novel genetic variants associated with hallux valgus.


A total of 5925 individuals of European Ancestry were categorized into two groups: ‘hallux valgus present’ (n = 2314) or ‘no deformity’ (n = 3611) as determined by trained examiners or using the Manchester grading scale. Genotyping was performed using commercially available arrays followed by imputation to the Haplotype Reference Consortium (HRC) reference panel version 1.1. We conducted both sex-specific and sex-combined association analyses using logistic regression and generalized estimating equations as appropriate in each cohort. Results were then combined in a fixed-effects inverse-variance meta-analyses. Functional Mapping and Annotation web-based platform (FUMA) was used for positional mapping, gene and gene-set analyses.


We identified a novel locus in the intronic region of CLCA2 on chromosome 1, rs55807512 (OR = 0.48, p = 2.96E-09), an expression quantitative trait locus for COL24A1, a member of the collagen gene family.


In this report of the largest GWAS of hallux valgus to date, we identified a novel genome-wide significant locus for hallux valgus. Additional replication and functional follow-up will be needed to determine the functional role of this locus in hallux valgus biology.

Peer Review reports


Hallux valgus, one of the most common structural foot deformities, is characterized by abduction of the great toe (hallux) with respect to the first metatarsal joint [1]. Hallux valgus is associated with pain, functional limitation, increased risk for falls, and diminished quality of life [2,3,4]. The condition is multifactorial in origin and the etiology is not completely understood. Hallux valgus is associated with female sex, older age, lower body mass index (BMI), and certain footwear types [1, 5,6,7]. Structural factors, such as metatarsal length and head shape, first ray hypermobility, and hind-foot pronation, are also considered to be important in hallux valgus development [6]. Hallux valgus is heritable, with estimates ranging from 0.29 to 0.89, suggesting that genetics may influence the development of this deformity [8, 9]. Identifying genetic variants associated with hallux valgus using an agnostic genome-wide approach may provide insights into the development of hallux valgus and lead to new treatment strategies.

The first and only genome-wide association study (GWAS) of hallux valgus was conducted as a meta-analysis in 4409 Caucasians based on a combined analysis of the Framingham Heart Study (FHS), the Genetics of Generalized Osteoarthritis (GOGO) Study, and the Johnston County Osteoarthritis Project (JoCoOA) [10]. This study did not find genome-wide significant associations with hallux valgus in either gender-specific or sex-combined GWAS meta-analyses. In this report, we expand the prior genome-wide association analysis by including association results from the Osteoarthritis Initiative (OAI), in which hallux valgus has also been measured and genome-wide genotyping is available.

The objective of the present paper is to identify novel genetic variants associated with hallux valgus in this expanded sample and with deeper genotype imputation performed (i.e., from 1000 Genomes to the Haplotype Reference Consortium (HRC) reference panel). With the addition of the OAI, the GWA sample size increased to 5925 Caucasian participants, representing a 34% increase in size from the prior GWA sample of 4409 subjects.


Study cohorts and assessment of hallux valgus

The meta-analysis included participants of European ancestry from four cohort studies: the Framingham Heart Study (FHS), the Genetics of Generalized Osteoarthritis (GOGO) Study, the Johnston County Osteoarthritis Project (JoCoOA), and the Osteoarthritis Initiative (OAI).

Framingham Heart Study

FHS is a community-based prospective study that began in 1948 with 5209 Framingham residents primarily white men and women of European-ancestry [11]. In 1972, 5124 offspring of the Original Cohort and their spouses were enrolled into the Offspring Cohort [12]. Our sample is limited to 2264 participants from Original and Offspring cohorts who were successfully genotyped and enrolled into Framingham Foot Study, an ancillary study of the FHS that was designed to examine the contribution of foot disorders to functional limitations [13]. Foot disorders, including hallux valgus, were assessed using a validated Foot Assessment Clinical Tool that captures the main features of common foot disorders by trained clinical examiners14 15. The validity of this tool was evaluated in a sample of elderly residents by comparing podiatry clinic findings to the results from the study examiners. The inter-observer and intra-observer reliability for hallux valgus were excellent [14, 15]. Hallux valgus was considered to be present if the angle of the hallux towards the lesser toes on either foot was observed to be greater than 15 degrees while weight-bearing, in either foot.

Genetics of Generalized Osteoarthritis

GOGO is a multisite collaboration involving seven sites in the United States and United Kingdom (UK). The purpose of study was to identify chromosomal regions associated with increased predisposition to generalized osteoarthritis (OA). The GOGO cohort is a sample of 2728 participants with and without hand OA from 1145 qualified families (at least two siblings with polyarticular OA). The study design has been previously reported [16]. A total of 1231 participants were successfully genotyped and completed clinical examination of the feet, including hallux valgus assessment (same method as JoCoOA, described in next section).

Johnston County Osteoarthritis Project

JoCoOA is an ongoing, community-based, prospective study of the occurrence of OA in Caucasian and African American residents in a rural North Carolina county [17, 18]. A total of 3187 participants were recruited at the 1991–97 baseline with an additional 1015 participants recruited into an enrichment cohort during 2003–2004. During the 2006–10 follow-up visit, 1695 participants completed clinical examination of the foot, including hallux valgus, performed by a trained clinical examiner. Of these, 919 successfully genotyped Caucasian participants were included into this study.

In GOGO and JoCoOA, structural deformities and conditions of the foot were classified as present and absent. Hallux valgus was assessed for each foot using a laminated foot diagram with two lines intersecting at 15°. Participants stood on the diagram with the medial edge of one foot against one line and their first metatarsophalangeal joint at the apex of the two lines. Hallux valgus was recorded as present if the angle of the great toe was greater than 15 degrees in either foot [5, 19, 20]. In JoCoOA, the inter-rater reliability for the hallux valgus measure was excellent for the left foot (kappa 0.84, 95% CI 0.73, 0.96) and good for the right foot (kappa 0.71, 95% CI 0.57, 0.92) [5].

Osteoarthritis Initiative

The OAI is a multi-center, longitudinal, prospective study, designed to identify risk factors for the development and progression of symptomatic knee OA [21]. Participants were recruited at clinical centers in Columbus, Ohio; Baltimore, Maryland; Pittsburgh, Pennsylvania; and Providence, Rhode Island who either were at risk for or had symptomatic radiographic knee OA. A total of 4796 received a baseline evaluation between 2004 and 2006 and were invited to annual follow-up visits for up to 8 years. Hallux valgus was assessed at the 96 month follow-up visit. First, participants were asked if they had ever had a bunionectomy on one or both feet (yes/no). Next, the presence and severity of hallux valgus was determined using the Manchester grading scale, which is recommended as a simple, non-invasive screening tool for clinical and research purposes [22, 23]. A trained and certified examiner compared the participants’ feet to photographs showing four grades and assigned a grade of hallux valgus deformity (grades 1–4: no deformity, mild deformity, moderate deformity, severe deformity) for each participant’s right and left foot separately. Because the severity of hallux valgus was not measured in FHS, GOGO and JoCoOA, the Manchester grades were collapsed into dichotomous categories to indicate presence and absence of hallux valgus based on recommendations from Menz et al. [23, 24]. In the publication by Menz et al., re-test reliability and agreement between dichotomous scores obtained by the examiners and the participants were similar to the levels reported for four severity categories [24]. For our main analyses in OAI, hallux valgus was considered present if participants reported a prior bunionectomy or if one or both feet had a Manchester grade of 3 or 4 (moderate or severe deformity). Hallux valgus was considered absent if participants reported no prior bunionectomy and had a Manchester grade of 1 (no deformity) in both feet. In a sensitivity analysis, OAI participants with Manchester grade of 2 (mild deformity) were added to the ‘no deformity’ group. Therefore, OAI provided two sets of GWAS results: (1) for the main analyses with the original definition of hallux valgus (N = 1511), and (2) for a sensitivity analysis allowing mild deformity to be included in the ‘no deformity’ group (N = 2120).

Genotyping, quality control (QC) and imputation

Details on genotyping and calling for each cohort were described elsewhere [10, 25]. In brief, genotyping was performed using commercially available arrays. To increase the number of tested SNPs and the overlap of variants available for analysis between different arrays, all Caucasian cohorts imputed genotypes to the most current HRC v1.1 reference panel [26] on the Michigan Imputation Server [27]. Additional details on genotyping and pre-imputation quality control in each study are listed in Supplementary Table 1.

Genome-wide association analyses

Following imputation, each study conducted GWAS under an additive genetic model, for the total sample and for women and men separately, to test the effect of imputed allelic dose on presence vs. absence of hallux valgus. For JoCoOA and OAI, the logistic regression model in PLINK v1.90 software was applied [28]. To account for within-family correlations in FHS and GOGO, the generalized estimating equations (GEE) model with the kinship matrix implemented in the R package GEE-pack [29] was used. In sex-specific GWAS, the models were adjusted for age at the time of foot examination, BMI, recruitment site (for OAI and GOGO), and population structure using the principal components. In analyses combining results for men and women, the models were additionally adjusted for sex.

Prior to meta-analysis, we performed post-GWAS harmonization and QC of GWAS results from each cohort to track possible errors in the study-specific analyses. We used the standard protocol accompanied by EasyQC R package [30]. Specifically, we removed single nucleotide polymorphisms (SNPs) with low minor allele frequencies (MAF) (< 0.01), low imputation quality (< 0.6), low minor allele count (<=10), large absolute values of beta coefficients and standard errors (> = 10), low call rate (< 0.95), and deviations from Hardy-Weinberg equilibrium (p < 10− 6).

The association results were combined using an inverse variance weighted fixed-effects meta-analysis in METAL software [31], with correction for genomic control. This method weights effect size estimates using the inverse of the corresponding standard errors. As noted previously, in each of the main analyses conducted in men, women, and both sexes combined, we excluded OAI participants categorized with mild hallux valgus deformity (grade 2), but included these participants in a sensitivity analysis. Heterogeneity was assessed using the I2 metric from the complete study-level meta-analysis. Between-study heterogeneity was tested using the Cochran Q statistic and considered significant at p = 0.1. A genome-wide significance threshold was set at the level of p = 5.0 × 10− 8. The Manhattan plots were generated in R. LocusZoom ( was used to provide regional visualization of results. We performed approximate conditional analysis (e.g., association analysis conditioning on the primary associated SNPs) using Genome-wide Complex Trait Analysis tool (GCTA v1.24) [32] to identify independent signals in suggestive loci. We defined a locus as a chromosomal region at which adjacent pairs of associated SNPs are less than 1 Mb distant. The collinearity threshold was set at r2 = 0.9, so that highly correlated SNPs are not selected in model.

Finally, we attempted to replicate findings from the discovery analysis in the UK Biobank by looking up findings in a GWAS of hallux valgus that has been made publicly available by the Neale lab at the Broad Institute The Neale lab conducted GWAS for 2419 phenotypes in the UK Biobank, which included hallux valgus defined by self-report. For the purpose of simplifying the process of association testing, the linear model with adjustment for sex and 10 principal components was fitted for all outcomes. Fitting a linear model to a binary outcome such a hallux valgus can introduce biases in coefficients and p-values due to violation of asymptotic assumptions of a linear model, especially for SNPs with low MAF in studies with relatively small sample sizes. Therefore, we followed the authors’ recommendations to remove SNPs below an allele frequency threshold defined as 25 divided by the smallest case group or 25/2314 = 0.01. We considered SNPs to replicate if they reached a nominal significance of p = 0.05 in the Neale lab data.

Functional annotation of SNPs and gene mapping

We performed functional annotation of GWAS results using Functional Mapping and Annotation of GWAS platform (FUMA) [33]. FUMA matches variants by chromosome, base-pair position, reference and alternate alleles to multiple publicly available databases to predict functional consequences for these SNPs, retrieve information on previously known SNP trait-association from the GWAS catalog, accommodate gene mapping, and to provide gene-based, pathway and tissue enrichment results. We also used PhenoScanner v2 to evaluate whether any of our associated or near-associated SNPs have been previously associated with musculoskeletal traits.

We assigned functional annotations to significant SNPs (p ≤ 5.0 × 10− 7 for analyses in the total sample; p ≤ 5.0 × 10− 6 for sex-specific analyses) and SNPs in linkage disequilibrium (LD) with significant SNPs (r2 > 0.6) using the SNP2GENE FUMA function, which incorporates tools from ANNOVAR, CADD, and RegulomeDB. ANNOVAR annotates functional effects of variants with respect to genes [34]. CADD predicts deleteriousness of the effect of a SNP on protein function. Higher CADD score refers to the more deleterious variants [35]. RegulomeDB scores variants based on information from expression quantitative trait loci (eQTLs) and chromatin marks. The score ranges from 1a to 7, where lower scores indicate increasing evidence that a variant is located in a functional region [36]. All LD information was calculated from the 1000 Genomes Phase 3 release reference panel.

SNPs were mapped to genes based on positional, eQTL, and 3D chromatin interaction mapping. Positional mapping was performed by selecting exonic and splicing SNPs with CADD score > =12.37. This threshold is recommended to restrict the mapping to deleterious coding SNPs [35]. We used eQTLs with false discovery rate (FDR) < 0.05 in 7 tissue types (adipose subcutaneous, whole blood, artery tibial, muscle skeletal, nerve tibial, cells transformed fibroblasts, skin sun exposed lower leg) from the Genotype Tissue Expression database (GTEx v7) [37, 38] and from additional data repositories (eQTLGen, xQTLServer [39], and MuTHER [40]). For chromatin interactions, Hi-C data in two tissues (psoas and mesenchymal stem cell) from GSE87112 were used; interactions were filtered by FDR < 10− 6. The MHC region was excluded from the analysis. We used MAGMA v1.07, which is integrated in FUMA to generate p-values quantifying the degree of association of genes and gene sets with hallux valgus [41]. GWAS summary statistics were aggregated to the level of whole genes to test the joint association of all markers in the gene with hallux valgus. This aggregation reduces the number of tests that are performed and identifies effects consisting of multiple weaker associations. Individual genes were then aggregated into groups of genes sharing certain biological, functional or other characteristics. We applied a default competitive model to test whether genes in a gene set are more strongly associated with hallux valgus than other gene sets. Tissue enrichment analyses were conducted in FUMA using two types of tissues from GTEx: 30 general tissue types from multiple organs and 53 specific tissue types within these organs.


Characteristics of participants and prevalence of hallux valgus in the discovery sample

Sample characteristics of the 5925 Caucasian participants (2314 categorized as ‘hallux valgus present’ and 3611 categorized as ‘no deformity’) who were included in the main analysis are summarized in Table 1. The mean age of participants is 66, ranging from 39 to 100 years. JoCoOA participants were older and had higher BMI compared to the three other cohorts. Within cohorts, cases were more likely to be female and older compared to controls. Hallux valgus was less prevalent and the proportion of men was higher in FHS compared to the other cohorts. In the total sample, cases were slightly older (mean age 67.8 vs 64.5), and proportion of females was higher among cases than among those without deformity. There were no case-control differences with respect to BMI.

Table 1 Baseline characteristics of the participants of European Ancestry from the Framingham Heart Study (FHS), the Genetics of Generalized Osteoarthritis (GOGO) Study, the Johnston County Osteoarthritis Project (JoCoOA), and the Osteoarthritis Initiative (OAI)

GWAS meta-analysis for total sample (Caucasians)

After removal of SNPs that failed to meet the post-GWAS QC criteria, the number of variants included in meta-analysis was 7,410,639 in FHS, 7,695,976 in JoCoOA, 7,646,026 in GOGO, and 7,729,175 in OAI. The results of gender-combined meta-analysis are summarized in the Manhattan plot (Fig. 1).

Fig. 1

Manhattan plot represents the p-values of the entire GWAS on genomic scale. The p-values are ordered by chromosome and position on the corresponding chromosome (x-axis). The value on the y-axis is the (–log10) of the p-value and is equivalent to the number of zeros after the decimal point plus one. The graph looks like a Manhattan skyline because of local correlation of SNPs. The red line shows the threshold for genome-wide significance

A genome-wide significant association was found for two variants located in an intronic region of chromosome 1 within the CLCA2 gene: rs55807512 (MAF = 4%, OR = 0.48, p = 2.96E-09) and rs12124247 (MAF = 3%, OR = 2.19, p = 7.38E-09). Effect direction was consistent across all four data sets (Fig. 2); these SNPs are in a weak LD (r [2] = 0.46). In conditional analysis, the effect of rs12124247 was attenuated and did not remain significant when conditioned on rs55807512 and vice versa indicating that both SNPs tag the same signal. No other SNPs were in a high LD (r2 > =0.8) with the top variant as shown in the regional plot (Fig. 3). Thirty additional SNPs were associated with hallux valgus at p < 5.0 × 10− 6 (Table 2, Supplementary Table 2). In the sensitivity analysis with an additional 609 OAI participants in the control group, the two top-hits remained significant (Table 2, Supplementary Table 3, Supplementary Figure 1) and no additional loci were identified.

Fig. 2

Forest plot of rs55807512 (CLCA2, chr1) association with hallux valgus in the meta-analysis

Fig. 3

Regional association plot of CLCA2 locus

Table 2 Top 10 SNPs from meta-analysis, adjusted for sex, BMI, age, study site, and principal components (total sample, N = 5925)

Sex-specific GWAS meta-analysis

The association signals diminished in sex-specific analyses (Supplementary Tables 4 and 5). In both men and women, the top-hits from the sex-combined analysis did not reach the genome-wide significance level at 5.0 × 10− 8. In men, we found only a single SNP passing the post-GWAS QC in the three cohorts to be significantly associated with hallux valgus: rs141161671 (MAF = 1%, OR = 6.50, p = 3.22E-08), located in the intronic region of chromosome 2 within AC007682.1 gene. The remaining SNPs with p < 5.0 × 10− 6 are listed in Supplementary Table 4. In women, we did not find any SNPs to be significantly associated with hallux valgus. However, rs55807512, the lead variant in the total sample analysis, was associated with hallux valgus with a p-value of 1.73E-06 (MAF = 4%, OR = 0.47). The remaining SNPs with p < 5.0 × 10− 6 are listed in Supplementary Table 5.


In the UK Biobank data (according to the summary statistics provided by the Neale Lab), neither rs55807512 nor rs12124247 were associated with hallux valgus. Several SNPs with p < 5.0 × 10− 6 in this meta-analysis showed nominal evidence (p < 0.05) for association with hallux valgus in the UK Biobank data (Supplementary Tables 23).

Functional annotation

FUMA identified one genomic risk locus on chromosome 1 tagged by the genome-wide significant lead SNP, rs55807512 (Fig. 4). No information on previously known SNP-trait associations was found for independent significant and tagged SNPs. Functional annotation of hallux valgus associated variants in CLCA2 revealed that rs55807512 is among the top (< 10%) of deleterious mutations in the genome (CADD = 11.89). eQTL mapping showed that our top hits, rs55807512 and rs12124247, which are located in CLCA2, are eQTLs for COL24A1 expression. 3D chromatin interactions revealed significant interactions between these genome-wide significant variants and 14 other genes on chromosome 1 (Fig. 4).

Fig. 4

Circos plots of mapped gene on chromosome 1 locus. Genomic risk loci are highlighted in blue. Genes are mapped by 3-D chromatin interaction (orange) or eQTLs (green), or both (red)

Gene and gene-set analyses did not show any significant associations. Of 18,722 protein coding genes tested, the most significantly associated gene was RUFY1 (p = 4.8 × 10− 6, Supplementary Table 6). Of 10,673 gene sets tested, the most significantly associated gene sets were “furukawa_dusp6_targets_pci35_up”, “positive regulation of cartilage development”, and “positive regulation of chondrocyte differentiation” (p < 1 × 10− 4, Supplementary Table 7).

Tissue analyses on 30 general tissue types from multiple organs and 53 specific tissue types within these organs) did not reveal any statistically significant associations (Supplementary Figures 23).


In the expanded hallux valgus meta-analysis on individuals of European ancestry, we identified a novel locus for hallux valgus in CLCA2. This study presents an updated meta-analysis of the first genome-wide association screen performed in hallux valgus which did not identify genome-wide significant SNPs [10]. This can, in part, be attributed to relatively modest sample sizes. We increased the sample by including data from the OAI and imputed genotypes to the most current HRC reference panel.

The lead variant, rs55807512, located in an intronic region of chromosome 1 within CLCA2 gene, had MAF around 4% and was not included in the first hallux valgus GWAS. Updating the imputation increased the number of low-frequency variants that were filtered out in previous analyses and can be studied reliably using the HRC reference panel. According to Entrez Gene database, CLCA2 encodes a member of the calcium-activated chloride channel regulator (CLCR) family of proteins that regulates transport of chloride across the plasma membrane. Although another member of CLCA family, CLCA4, has been reported to be associated with osteochondrosis in the horse [42], CLCA2 has not been associated with bone formation or any musculoskeletal disorders. However, COL24A1 may be the true gene of interest since our top hits were eQTLs for COL24A1 expression. COL24A1, a member of the collagen gene family, is developmentally expressed in cornea and bone by osteoblasts and regulates osteoblast differentiation and mineralization through interactions with integrins, which leads to the activation of the TGF-β/ Smad signaling pathway [43,44,45]. Collagen type XXIV may be involved in structural differences between fibrillary collagens and affect fibril diameter [44, 46]. Abnormal collagen fibrils are associated with a wide spectrum of diseases of bone and cartilage, including hallux valgus [47, 48]. Uchiyama et al. [48] demonstrated that feet with hallux valgus have different structures of collagen fibers compared to normal feet. This may be in response to continuous stress to the medial collateral ligament, one of the important joint stabilizers, and lead to altered organization of collagen I and collagen III fibrils that could leave the first metatarsophalangeal joint unprotected during gait [48, 49].

An important paralog of COL24A1 is COL5A1. Mutations in the COL5A1 gene, encoding the alpha 1 of type V collagen, have been identified in patients with Ehlers-Danlos syndrome [50, 51] which has been linked to hallux valgus [52], Achilles tendinopathy [50], acquired injuries such as ACL tears [53], and with range of motion [50].

None of the top SNPs identified from the previous hallux valgus meta-analysis became more significant in our updated meta-analysis. Of the four SNPs that met p < 5E-6 in men, only r10224956 and rs4476613, reached nominal significance (p = 0.02 and p = 0.001, respectively) in our study. Of the six SNPs that met p < 5E-6 in women, only rs12214759 and rs2242411 reached suggestive significance (p = 6.70E-06 and p = 6.67E-05, respectively) in our study with the same direction of effect. Furthermore, none of the previously identified SNPs were associated with hallux valgus in the UK Biobank GWAS.

One of the difficulties in studying the genetics of hallux valgus is the lack of a standardized phenotype. The method of measuring hallux valgus in studies collecting such data is not always clearly described. Furthermore, hallux valgus prevalence in studies using self-report data may be under-reported or inaccurate due to a lack of a validated assessment tool for this condition and lack of standardization for terms used in questionnaires (e.g., “bunion” and “hallux valgus”) [1, 24].An important advantage to our study is the detailed assessment of hallux valgus based on objective criteria rather than self-report. Although the presence of hallux valgus was not measured using weight-bearing radiographs of the feet, the reference standard of angle measurement, the clinical measures we used have been previously validated and were conducted by trained examiners which should minimize potential sources of error. These tools have been reported as alternatives to radiographs due to lower cost and lack of radiographic exposure, particularly for large-scale cohort studies that include asymptomatic participants [23]. It is possible that in the absence of diagnostic tests and in-depth knowledge of participants’ medical history, several clinical diagnoses such as a bursa, prominent medial eminence of the first metatarsal, or bony swelling in joints with osteoarthritis can be misclassified as hallux valgus. However, these conditions are relatively rare in a general population and thus misclassification of these conditions likely had little effect on association results obtained from our meta-analysis. Importantly, another strength of our study is that it was not based on clinical cases only, but rather on a general population and therefore not affected by selection bias.

Our results should be interpreted in light of several limitations. First, hallux valgus was assessed across cohorts in two different ways (angular criteria vs. Manchester grading scale), which may lead to phenotypic misclassification and potential loss of statistical power. However, we assessed the distributions of the phenotype by cohort and compared distributions of key factors like age, sex, and BMI to ensure that there were no major differences. In all studies, participants categorized as ‘hallux valgus present’ were slightly older and were more likely to be female than those categorized as ‘no deformity’. As we noted previously, hallux valgus was less prevalent in FHS than in GOGO, JoCo, and OAI. This can be explained by the fact that FHS is a geographically-defined cohort study which did not specifically select individuals with or at risk of OA unlike OAI and GOGO. In addition, the lower prevalence of hallux valgus in FHS can be attributed to 1) differences in BMI and sex distributions and 2) environmental risk factors shared by family members leading to the development or prevention of hallux valgus [54]. Despite efforts to minimize bias and ensure that hallux valgus was classified using a comparable method to JoCo, GOGO, and FHS as described by Menz and others, heterogeneity resulting from pooling data across studies may still be present and we can only speculate how results would change if the OAI cohort had been assessed for hallux valgus using angular criteria. We note though that it is unlikely that our primary findings were driven by OAI or any single study since I2 values were low and showed little evidence for study heterogeneity. Misclassification is a potential problem in the OAI where participants have less severe forms of the condition. Participants with mild deformity, however, were excluded from our main analyses, and including these participants in the sensitivity analysis did not affect our novel findings. Overall, any misclassification and heterogeneity would likely bias associations toward the null and would not affect our findings, but may limit power for additional discoveries. Second, we were unable to assess the severity of hallux valgus because we were limited by the measurements available in the participating studies. As noted previously, using ordinal measurements of hallux valgus such as the Manchester grade can improve the statistical power compared to a dichotomous trait such as hallux valgus presence or absence [10, 22]. Third, we were unable to replicate our findings in a different independent population with a comparable level of phenotyping. To the best of our knowledge, there are no other Caucasian cohorts with well-defined hallux valgus phenotypes and genome-wide genotyping. In the UK Biobank data that we used for replication, the lead variant was not associated with hallux valgus. This may be explained in part by the use of different phenotype criteria and different statistical models (logistic vs. linear regression, BMI adjustment). The prevalence of hallux valgus was much lower (~ 2%) in the UK Biobank compared to our meta-analysis (31–48%). Replication of our findings in additional studies with identical phenotype criteria and design will be important in the future. Fourth, we did not evaluate whether our findings are generalizable to individuals of other ancestry groups. We included only participants of European Ancestry in the analyses. Although GWAS data were available for 600 African American (AA) participants (268 from OAI and 332 from JoCoOA), we did not perform meta-analysis on AA samples due to a small sample size and limited statistical power.

In conclusion, we reported the largest hallux valgus meta-analysis on individuals of European ancestry. Hallux valgus is a common foot disorder that is greatly understudied, particularly its possible genetic aspects. Building upon prior work, we aimed to identify novel genetic variants associated with hallux valgus, and found a novel variant in the gene CLCA2. In addition, our top-hits in CLCA2 are eQTLs for a neighboring COL24A1 gene and potentially pinpoint the true gene of interest from an associated locus. While observed results were attenuated and signal diminished in sex-specific analyses, this study provides new insights into hallux valgus biology and the findings for additional replication and functional follow-up.

Availability of data and materials

JoCo and GOGO are not publicly available data sources, and thus permission from the principal investigators is required for obtaining data (, OAI data is publicly available at Participant-level phenotype and genotype data from the Framingham Heart Study are accessible from the U.S. National Center for Biotechnology Information (NCBI) database of Genotypes and Phenotypes (dbGaP) at to approved scientific investigators pursuing research questions that are consistent with the informed consent agreements provided by individual research participants.


  1. 1.

    Nix S, Smith M, Vicenzino B. Prevalence of hallux valgus in the general population: a systematic review and meta-analysis. J Foot Ankle Res 2010;3:21.

  2. 2.

    Menz HB, Roddy E, Thomas E, Croft PR. Impact of hallux valgus severity on general and foot-specific health-related quality of life. Arthritis Care Res (Hoboken) 2011;63(3):396–404. doi: [published Online First: 2010/11/17].

  3. 3.

    Menz HB, Auhl M, Spink MJ. Foot problems as a risk factor for falls in community-dwelling older people: a systematic review and meta-analysis. Maturitas. 2018;118:7–14.

    Article  Google Scholar 

  4. 4.

    Cho NH, Kim S, Kwon DJ, Kim HA. The prevalence of hallux valgus and its association with foot pain and function in a rural Korean community. J Bone Joint Surg Br. 2009;91b(4):494–8.

    Article  Google Scholar 

  5. 5.

    Golightly YM, Hannan MT, Dufour AB, Renner JB, Jordan JM. Factors associated with hallux Valgus in a community-based cross-sectional study of adults with and without osteoarthritis. Arthrit Care Res. 2015;67(6):791–8.

    Article  Google Scholar 

  6. 6.

    Nix SE, Vicenzino BT, Collins NJ, Smith MD. Characteristics of foot structure and footwear associated with hallux valgus: a systematic review, Osteoarthritis Cartilage. 2012;20(10):1059–74. [published Online First: 2012/07/10].

    CAS  Article  Google Scholar 

  7. 7.

    Nguyen USDT, Hillstrom HJ, Li W, Dufour AB, Kiel DP, Procter-Gray E, Gagnon MM, Hannan MT. Factors associated with hallux valgus in a population-based study of older women and men: the MOBILIZE Boston study. Osteoarthr Cartilage. 2010;18(1):41–6.

    Article  Google Scholar 

  8. 8.

    Hannan MT, Menz HB, Jordan JM, Cupples LA, Cheng CH, Hsu YH. High heritability of hallux Valgus and lesser toe deformities in adult men and women. Arthrit Care Res. 2013;65(9):1515–21.

    Article  Google Scholar 

  9. 9.

    Lee CH, Lee S, Kang H, Jung DE, Song YM, Lee K, Lee K, Hwang J, Sung J. Genetic influences on hallux Valgus in Koreans: the healthy twin study. Twin Res Hum Genet. 2014;17(2):121–6.

    Article  Google Scholar 

  10. 10.

    Hsu YH, Liu Y, Hannan MT, Maixner W, Smith SB, Diatchenko L, Golightly YM, Menz HB, Kraus VB, Doherty M, Wilson AG, Jordan JM. Genome-wide association meta-analyses to identify common genetic variants associated with hallux valgus in Caucasian and African Americans. J Med Genet 2015;52(11):762–9. doi: [published Online First: 2015/09/05].

  11. 11.

    Dawber TR, Meadors GF, Moore FE. Epidemiological approaches to heart disease: the Framingham study. Am J Public Health N. 1951;41(3):279–86.

    CAS  Article  Google Scholar 

  12. 12.

    Feinleib M, Kannel WB, Garrison RJ, Mcnamara PM, Castelli WP. Framingham Offspring Study - Design and Preliminary Data. Prev Med. 1975;4(4):518–25.

    CAS  Article  Google Scholar 

  13. 13.

    Dufour AB, Broe KE, Nguyen USDT, Gagnon DR, Hillstrom HJ, Walker AH, Kivell E, Hannan MT. Foot pain: is current or past Shoewear a factor? Arthrit Care Res. 2009;61(10):1352–8.

    Article  Google Scholar 

  14. 14.

    Hannan MT, Zimmer J, Sullivan E, Diel DP. Physical limitations and foot disorders in elders. J Am Geriatr Soc. 2001;49(4):S22.

    Google Scholar 

  15. 15.

    Hannan MT, Murabito JM, Felson DT, Rivinus M, Kaplan J, Kiel DP. The epidemiology of foot disorders and foot pain in men and women: the Framingham study. Arthritis Rheum. 2003;48(9):S672–S72.

    Google Scholar 

  16. 16.

    Kraus VB, Jordan JM, Doherty M, Wilson AG, Moskowitz R, Hochberg M, Loeser R, Hooper M, Renner JB, Crane MM, Hastie P, Sundseth S, Atif U. The genetics of generalized osteoarthritis (GOGO) study: study design and evaluation of osteoarthritis phenotypes. Osteoarthr Cartilage. 2007;15(2):120–7.

    CAS  Article  Google Scholar 

  17. 17.

    Jordan JM, Linder GF, Renner JB, Fryer JG. The impact of arthritis in rural populations. Arthritis Care Res. 1995;8(4):242–50.

    CAS  Article  PubMed  Google Scholar 

  18. 18.

    Jordan JM, Helmick CG, Renner JB, Luta G, Dragomir AD, Woodard J, Fang F, Schwartz TA, Abbate LM, Callahan LF, Kalsbeek WD, Hochberg MC. Prevalence of knee symptoms and radiographic and symptomatic knee osteoarthritis in African Americans and Caucasians: the Johnston County osteoarthritis project. J Rheumatol. 2007;34(1):172–80.

    PubMed  Google Scholar 

  19. 19.

    Hagedorn TJ, Dufour AB, Riskowski JL, Hillstrom HJ, Menz HB, Casey VA, Hannan MT. Foot disorders, foot posture, and foot function: the framingham foot study. PLoS One. 2013;8(9):e74364.

    CAS  Article  PubMed  Google Scholar 

  20. 20.

    Galica AM, Hagedorn TJ, Dufour AB, Riskowski JL, Hillstrom HJ, Casey VA, Hannan MT. Hallux valgus and plantar pressure loading: the Framingham foot study. J Foot Ankle Res. 2013;6:42.

  21. 21.

    Lester G. The osteoarthritis initiative: a NIH public-private partnership. HSS J. 2012;8(1):62–3.

    Article  Google Scholar 

  22. 22.

    Garrow AP, Papageorgiou A, Silman AJ, Thomas E, Jayson MIV, Macfarlane GJ. The grading of hallux valgus - the Manchester scale. J Am Podiat Med Assn. 2001;91(2):74–8.

    CAS  Article  Google Scholar 

  23. 23.

    Menz HB, Munteanu SE. Radiographic validation of the Manchester scale for the classification of hallux valgus deformity. Rheumatology. 2005;44(8):1061–6.

    CAS  Article  Google Scholar 

  24. 24.

    Menz HB, Fotoohabadi MR, Wee E, Spink MJ. Validity of self-assessment of hallux valgus using the Manchester scale. Bmc Musculoskel Dis. 2010;11:215.

  25. 25.

    Yau MS, Yerges-Armstrong LM, Liu YF, Lewis CE, Duggan DJ, Renner JB, Torner J, Felson DT, McCulloch CE, Kwoh CK, Nevitt MC, Hochberg MC, Mitchell BD, Jordan JM, Jackson RD. Genome-wide association study of radiographic knee osteoarthritis in north American Caucasians. Arthritis Rheumatol. 2017;69(2):343–51.

    Article  PubMed  Google Scholar 

  26. 26.

    McCarthy S, Das S, Kretzschmar W, Delaneau O, Wood AR, Teumer A, Kang HM, Fuchsberger C, Danecek P, Sharp K, Luo Y, Sidorel C, Kwong A, Timpson N, Koskinen S, Vrieze S, Scott LJ, Zhang H, Mahajan A, Veldink J, Peters U, Pato C, van Duijn CM, Gillies CE, Gandin I, Mezzavilla M, Gilly A, Cocca M, Traglia M, Angius A, Barrett JC, Boomsma D, Branham K, Breen G, Brummett CM, Busonero F, Campbell H, Chan A, Che S, Chew E, Collins FS, Corbin LJ, Smith GD, Dedoussis G, Dorr M, Farmaki AE, Ferrucci L, Forer L, Fraser RM, Gabriel S, Levy S, Groop L, Harrison T, Hattersley A, Holmen OL, Hveem K, Kretzler M, Lee JC, McGue M, Meitinger T, Melzer D, Min JL, Mohlke KL, Vincent JB, Nauck M, Nickerson D, Palotie A, Pato M, Pirastu N, McInnis M, Richards JB, Sala C, Salomaa V, Schlessinger D, Schoenherr S, Slagboom PE, Small K, Spector T, Stambolian D, Tuke M, Tuomilehto J, Van den Berg LH, Van Rheenen W, Volker U, Wijmenga C, Toniolo D, Zeggini E, Gasparini P, Sampson MG, Wilson JF, Frayling T, PIW d B, Swertz MA, McCarroll S, Kooperberg C, Dekker A, Altshuler D, Willer C, Iacono W, Ripatti S, Soranzo N, Walter K, Swaroop A, Cucca F, Anderson CA, Myers RM, Boehnke M, MI MC, Durbin R, Abecasis G, Marchini J, Consortium HR. A reference panel of 64,976 haplotypes for genotype imputation. Nat Genet. 2016;48(10):1279–83.

    CAS  Article  PubMed Central  PubMed  Google Scholar 

  27. 27.

    Das S, Forer L, Schonherr S, Sidore C, Locke AE, Kwong A, Vrieze SI, Chew EY, Levy S, McGue M, Schlessinger D, Stambolian D, Loh PR, Iacono WG, Swaroop A, Scott LJ, Cucca F, Kronenberg F, Boehnke M, Abecasis GR, Fuchsberger C. Next-generation genotype imputation service and methods. Nat Genet. 2016;48(10):1284–7.

    CAS  Article  PubMed Central  PubMed  Google Scholar 

  28. 28.

    Chang CC, Chow CC, Tellier LCAM, Vattikuti S, Purcell SM, Lee JJ. Second-generation PLINK: rising to the challenge of larger and richer datasets. Gigascience 2015;4:7.

  29. 29.

    Halekoh U, Hojsgaard S, Yan J. The R package geepack for generalized estimating equations. J Stat Softw. 2006;15(2):1–11.

    Article  Google Scholar 

  30. 30.

    Winkler TW, Day FR, Croteau-Chonka DC, Wood AR, Locke AE, Magi R, Ferreira T, Fall T, Graff M, Justice AE, Luan JA, Gustafsson S, Randall JC, Vedantam S, Workalemahu T, Kilpelainen TO, Scherag A, Esko T, Kutalik Z, Heid IM, Loos RJF, Trai GIA. Quality control and conduct of genome-wide association meta-analyses. Nat Protoc. 2014;9(5):1192–212.

    Article  PubMed Central  PubMed  Google Scholar 

  31. 31.

    Willer CJ, Li Y, Abecasis GR. METAL: fast and efficient meta-analysis of genomewide association scans. Bioinformatics. 2010;26(17):2190–1.

    CAS  Article  PubMed Central  PubMed  Google Scholar 

  32. 32.

    Yang J, Lee SH, Goddard ME, Visscher PM. GCTA: a tool for genome-wide complex trait analysis. Am J Hum Genet. 2011;88(1):76–82. [published Online First: 2010/12/21].

    CAS  Article  PubMed Central  PubMed  Google Scholar 

  33. 33.

    Watanabe K, Taskesen E, van Bochoven A, Posthuma D. Functional mapping and annotation of genetic associations with FUMA. Nat Commun. 2017;8(1):1826. [published Online First: 2017/12/01].

    CAS  Article  PubMed Central  PubMed  Google Scholar 

  34. 34.

    Wang K, Li M, Hakonarson H. ANNOVAR: functional annotation of genetic variants from high-throughput sequencing data. Nucleic Acids Res. 2010;38(16):e164. [published Online First: 2010/07/06].

    CAS  Article  PubMed Central  PubMed  Google Scholar 

  35. 35.

    Kircher M, Witten DM, Jain P, O'Roak BJ, Cooper GM, Shendure J. A general framework for estimating the relative pathogenicity of human genetic variants. Nat Genet 2014;46(3):310–315. doi: [published Online First: 2014/02/04].

    Article  Google Scholar 

  36. 36.

    Boyle AP, Hong EL, Hariharan M, Cheng Y, Schaub MA, Kasowski M, Karczewski KJ, Park J, Hitz BC, Weng S, Cherry JM, Snyder M. Annotation of functional variation in personal genomes using RegulomeDB. Genome Res. 2012;22(9):1790–7. [published Online First: 2012/09/08].

    CAS  Article  PubMed Central  PubMed  Google Scholar 

  37. 37.

    Consortium GT. The Genotype-Tissue Expression (GTEx) project. Nat Genet. 2013;45(6):580–5. [published Online First: 2013/05/30].

    CAS  Article  Google Scholar 

  38. 38.

    Consortium GT. Human genomics. The Genotype-Tissue Expression (GTEx) pilot analysis: multitissue gene regulation in humans. Science. 2015;348(6235):648–60. [published Online First: 2015/05/09].

    CAS  Article  Google Scholar 

  39. 39.

    Ng B, White CC, Klein HU, Sieberts SK, McCabe C, Patrick E, Xu J, Yu L, Gaiteri C, Bennett DA, Mostafavi S, De Jager PL. An xQTL map integrates the genetic architecture of the human brain's transcriptome and epigenome. Nat Neurosci. 2017;20(10):1418–26. [published Online First: 2017/09/05].

    CAS  Article  PubMed Central  PubMed  Google Scholar 

  40. 40.

    Grundberg E, Small KS, Hedman AK, Nica AC, Buil A, Keildson S, Bell JT, Yang TP, Meduri E, Barrett A, Nisbett J, Sekowska M, Wilk A, Shin SY, Glass D, Travers M, Min JL, Ring S, Ho K, Thorleifsson G, Kong A, Thorsteindottir U, Ainali C, Dimas AS, Hassanali N, Ingle C, Knowles D, Krestyaninova M, Lowe CE, Di Meglio P, Montgomery SB, Parts L, Potter S, Surdulescu G, Tsaprouni L, Tsoka S, Bataille V, Durbin R, Nestle FO, O'Rahilly S, Soranzo N, Lindgren CM, Zondervan KT, Ahmadi KR, Schadt EE, Stefansson K, Smith GD, McCarthy MI, Deloukas P, Dermitzakis ET, Spector TD, Multiple Tissue Human Expression Resource C. Mapping cis- and trans-regulatory effects across multiple tissues in twins. Nat Genet 2012;44(10):1084–9. doi: [published Online First: 2012/09/04].

    CAS  Article  Google Scholar 

  41. 41.

    de Leeuw CA, Mooij JM, Heskes T, Posthuma D. MAGMA: generalized gene-set analysis of GWAS data. PLoS Comput Biol. 2015;11(4):e1004219. [published Online First: 2015/04/18].

    CAS  Article  PubMed Central  PubMed  Google Scholar 

  42. 42.

    Bates JT, Jacobs JC, Jr., Shea KG, Oxford JT. Emerging genetic basis of osteochondritis dissecans. Clin Sports Med 2014;33(2):199–220. doi: [published Online First: 2014/04/05].

  43. 43.

    Wang W, Olson D, Liang G, Franceschi RT, Li C, Wang B, Wang SS, Yang S. Collagen XXIV (Col24alpha1) promotes osteoblastic differentiation and mineralization through TGF-beta/Smads signaling pathway. Int J Biol Sci. 2012;8(10):1310–22. [published Online First: 2012/11/10].

    CAS  Article  PubMed Central  PubMed  Google Scholar 

  44. 44.

    Koch M, Laub F, Zhou PH, Hahn RA, Tanaka S, Burgeson RE, Gerecke DR, Ramirez F, Gordon MK. Collagen XXIV, a vertebrate fibrillar collagen with structural features of invertebrate collagens - selective expression in developing cornea and bone. J Biol Chem. 2003;278(44):43236–44.

    CAS  Article  Google Scholar 

  45. 45.

    Matsuo N, Tanaka S, Yoshioka H, Koch M, Gordon MK, Ramirez F. Collagen XXIV (Col24a1) gene expression is a specific marker of osteoblast differentiation and bone formation. Connect Tissue Res. 2008;49(2):68–75.

    CAS  Article  Google Scholar 

  46. 46.

    Bella J, Hulmes DJ. Fibrillar Collagens. Subcell Biochem. 2017;82:457–90. [published Online First: 2017/01/20].

    CAS  Article  Google Scholar 

  47. 47.

    Kuivaniemi H, Tromp G, Prockop DJ. Mutations in fibrillar collagens (types I, II, III, and XI), fibril-associated collagen (type IX), and network-forming collagen (type X) cause a spectrum of diseases of bone, cartilage, and blood vessels. Hum Mutat. 1997;9(4):300–15.<300::AID-HUMU2>3.0.CO;2-9 [published Online First: 1997/01/01].

    CAS  Article  Google Scholar 

  48. 48.

    Uchiyama E, Kitaoka HB, Luo ZP, Grande JP, Kura H, An KN. Pathomechanics of hallux valgus: biomechanical and immunohistochemical study. Foot Ankle Int. 2005;26(9):732–8. [published Online First: 2005/09/22].

    Article  Google Scholar 

  49. 49.

    Perera AM, Mason L, Stephens MM. The Pathogenesis of Hallux Valgus. J Bone Joint Surg Am. 2011;93a(17):1650–61.

    Article  Google Scholar 

  50. 50.

    Collins M, Mokone GG, September AV, van der Merwe L, Schwellnus MP. The COL5A1 genotype is associated with range of motion measurements. Scand J Med Sci Spor. 2009;19(6):803–10.

    CAS  Article  Google Scholar 

  51. 51.

    Symoens S, Malfait F, Renard M, Andre J, Hausser I, Loeys B, Coucke P, De Paepe A. COL5A1 signal peptide mutations interfere with protein secretion and cause classic Ehlers-Danlos syndrome. Hum Mutat. 2009;30(2):E395–403.

    Article  Google Scholar 

  52. 52.

    Mann RA, Coughlin MJ. Surgery of the foot and ankle. 6th ed. St. Louis: Mosby; 1993.

    Google Scholar 

  53. 53.

    Vaughn NH, Stepanyan H, Gallo RA. Dhawan A. Genetic Factors in Tendon Injury: A Systematic Review of the Literature. Orthop J Sports Med. 2017;5(8):2325967117724416.

    Article  Google Scholar 

  54. 54.

    Munteanu SE, Menz HB, Wark JD, Christie JJ, Scurrah KJ, Bui M, Erbas B, Hopper JL, Wluka AE. Hallux Valgus, By Nature or Nurture? A Twin Study. Arthritis Care Res (Hoboken). 2017;69(9):1421–8. [published Online First: 2016/11/20].

    Article  Google Scholar 

Download references


Not applicable.


JoCo was supported in part by S043, S1734, & S3486 from the Centers for Disease Control and Prevention/Association of Schools of Public Health; 5-P60-AR30701 & 5-P60 AR49465–03 from the National Institute of Arthritis Musculoskeletal and Skin Diseases (NIAMS) of the National Institutes of Health (NIH) and Algynomics, Inc. Ms. Arbeeva and Drs. Golightly, Nelson, and Jordan were supported by NIAMS Multidisciplinary Clinical Research Grant 5-P60 AR062760 and CDC 1U01DP006266. The conclusions are attributed to the authors and cannot be attributed to the CDC. The GOGO study was supported by GlaxoSmithKline. FHS was supported by R01-AR060492. The Osteoarthritis Initiative is a public–private partnership funded by the NIH (NIAMS contracts N01-AR-2-2258, N01-AR-2-2259, N01-AR-2-2260, N01-AR-2-2261, and N01-AR-2-2262). Genotyping in OAI was supported by the NIH (NIAMS RC2 AR058950). Additional support was provided by NIH grant P30-DK-072488, T32-AG-00262, and T32-AG-023480.

Author information




LA drafted the paper; MY, BDM, RDJ, YMG, MTH and MCH provided essential comments and contributed to the manuscript editing; RDJ, MTH, JMJ and MCH acquired data; LA, MY, KR analyzed the data; all authors contributed to study design, data interpretation, and approved the final manuscript.

Corresponding author

Correspondence to Liubov Arbeeva.

Ethics declarations

Ethics approval

The study was approved by the IRB at each clinical center. All participants provided informed consent. The OAI study and public use of clinical and imaging data used in this study were approved by the committee on Human Research at the University of California, San Francisco (IRB# 10–00532). JoCo study was approved by the UNC Office of Human Research Ethics (IRB# 18–0438). GoGo study was approved by the UNC Office of Human Research Ethics (IRB# 99–0807). The Framingham Genetics of Foot Disorders Study was approved by the Institutional Review Board of Hebrew SeniorLife, Protocol #10–009.

Consent for publication

Not applicable.

Competing interests

None declared.

Additional information

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit The Creative Commons Public Domain Dedication waiver ( applies to the data made available in this article, unless otherwise stated in a credit line to the data.

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Arbeeva, L., Yau, M., Mitchell, B.D. et al. Genome-wide meta-analysis identified novel variant associated with hallux valgus in Caucasians. J Foot Ankle Res 13, 11 (2020).

Download citation


  • Genome-wide
  • Muscle disease
  • Genetic epidemiology
  • Rheumatology