Skip to main content
PMC home page
Journal of Dental Research logoLink to Journal of Dental Research
. 2016 Jul 1;95(11):1245–1256. doi: 10.1177/0022034516657003

Association Studies and Direct DNA Sequencing Implicate Genetic Susceptibility Loci in the Etiology of Nonsyndromic Orofacial Clefts in Sub-Saharan African Populations

LJJ Gowans 1,2,3,4,*, WL Adeyemo 5,*, M Eshete 6,*, PA Mossey 7, T Busch 4, B Aregbesola 8, P Donkor 2,9, FKN Arthur 1, SA Bello 10, A Martinez 4, M Li 4, EA Augustine-Akpan 4, W Deressa 6, P Twumasi 1, J Olutayo 5, M Deribew 6, P Agbenorku 2,9, AA Oti 2,9, R Braimah 8, G Plange-Rhule 2, M Gesses 11, S Obiri-Yeboah 2,9, GO Oseni 12, PB Olaitan 12, L Abdur-Rahman 13, F Abate 11, T Hailu 11, P Gravem 14, MO Ogunlewe 12, CJ Buxó 15, ML Marazita 16, AA Adeyemo 17, JC Murray 3, A Butali 4,✉
PMCID: PMC5076758  PMID: 27369588

Abstract

Orofacial clefts (OFCs) are congenital dysmorphologies of the human face and oral cavity, with a global incidence of 1 per 700 live births. These anomalies exhibit a multifactorial pattern of inheritance, with genetic and environmental factors both playing crucial roles. Many loci have been implicated in the etiology of nonsyndromic cleft lip with or without cleft palate (NSCL/P) in populations of Asian and European ancestries, through genome-wide association studies and candidate gene studies. However, few populations of African descent have been studied to date. Here, the authors show evidence of an association of some loci with NSCL/P and nonsyndromic cleft palate only (NSCPO) in cohorts from Africa (Ghana, Ethiopia, and Nigeria). The authors genotyped 48 single-nucleotide polymorphisms that were selected from previous genome-wide association studies and candidate gene studies. These markers were successfully genotyped on 701 NSCL/P and 163 NSCPO cases, 1,070 unaffected relatives, and 1,078 unrelated controls. The authors also directly sequenced 7 genes in 184 nonsyndromic OFC (NSOFC) cases and 96 controls from Ghana. Population-specific associations were observed in the case-control analyses of the subpopulations, with West African subpopulations (Ghana and Nigeria) showing a similar pattern of associations. In meta-analyses of the case-control cohort, PAX7 (rs742071, P = 5.10 × 10−3), 8q24 (rs987525, P = 1.22 × 10−3), and VAX1 (rs7078160, P = 0.04) were nominally associated with NSCL/P, and MSX1 (rs115200552, P = 0.01), TULP4 (rs651333, P = 0.04), CRISPLD2 (rs4783099, P = 0.02), and NOG1 (rs17760296, P = 0.04) were nominally associated with NSCPO. Moreover, 7 loci exhibited evidence of threshold overtransmission in NSOFC cases through the transmission disequilibrium test and through analyses of the family-based association for disease traits. Through DNA sequencing, the authors also identified 2 novel, rare, potentially pathogenic variants (p.Asn323Asp and p.Lys426IlefsTer6) in ARHGAP29. In conclusion, the authors have shown evidence for the association of many loci with NSCL/P and NSCPO. To the best of this knowledge, this study is the first to demonstrate any of these association signals in any African population.

Keywords: genetic heterogeneity, rare variants, genome-wide association studies (GWAS), candidate genes, craniofacial genetics, population genetics

Introduction

Human orofacial clefts (OFCs) are congenital malformations of the face and oral cavity due to dysregulation of embryologic processes. The global incidence of OFCs is 1 per 700 live births. However, race, ethnicity, geographic locations, environmental factors, and socioeconomic status influence the incidence of OFCs (Gorlin et al. 2001). The highest incidence occurs in Asians, followed by populations of European ancestry, whereas African populations have the lowest incidence (Mossey and Modell 2012). Although there are no national prevalence data for Ghana and Ethiopia, an estimate of 0.5 per 1,000 has been observed for Nigeria (Butali, Adeyemo, et al. 2014). These observations presuppose that the relative contributions of individual susceptibility genes may vary across different human populations. OFCs may be syndromic or nonsyndromic, with the syndromic forms presenting with other congenital anomalies. The etiology of the more common nonsyndromic OFCs (NSOFCs) is complex, exhibiting multifactorial pattern of inheritance. NSOFCs are classified into nonsyndromic cleft lip with or without cleft palate (NSCL/P) and nonsyndromic cleft palate only (NSCPO), and these 2 groups have a heterogeneous genetic architecture. NSCL/P comprises nonsyndromic cleft lip only (NSCL) and nonsyndromic cleft lip and palate (NSCLP; Dixon et al. 2011).

To date, 6 genome-wide association studies (GWASs) and a meta-analysis have been published for NSOFCs, with these signals demonstrating an association with NSCL/P but not NSCPO. In a GWAS involving Europeans, an association was observed between a locus in Chr8q.24 and NSCL/P (Birnbaum et al. 2009). The 8q.24 signal was subsequently replicated in another GWAS of NSCL/P in Europeans from the United States (Grant et al. 2009). A third GWAS that involved cohorts of European ancestries also revealed that 2 additional loci, 17q22 (NOG1) and 10q25 (VAX1), were associated with NSCL/P. Other loci yielded a suggestive association with NSCL/P: 15q13.3 (GREM1), 13q31.1 (SPRY2), and 2p21 (THADA; Mangold et al. 2010). Employing trios of Asian and European ancestries, a GWAS implicated 20q12 (MAFB) and 1p22.1 (ABCA4) in the etiology of NSCL/P, with 17p13 (NTN1) showing a suggestive association. Stratified analyses based on ancestries by the same GWAS showed that some signals were ancestry specific: trios of European ancestry gave the strongest association for 8q.24, whereas those of Asian ancestry were strongly associated with MAFB, ABCA4, and IRF6 (Beaty et al. 2010). A meta-analysis revealed additional NSCL/P susceptibility loci: THADA, SPRY2, 15q22.2 (TPM1), and 1p36 (PAX7; Ludwig et al. 2012). Recently, a GWAS involving Asians implicated 16p13.3 (ADCY9; Sun et al. 2015) in the etiology of NSCL/P, whereas a GWAS involving dogs and a Guatemalan population gave a suggestive association for ADAMTS20 (Wolf et al. 2015).

In the pre- and post-GWAS era, candidate gene and replication studies have been instrumental in identifying cleft susceptibility loci. Pathogenic variants in IRF6 were shown to cause van der Woude syndrome and popliteal pterygium syndrome (Kondo et al. 2002). Subsequently, a missense variant in IRF6 (rs2235371) demonstrated overtransmission in NSCL/P cases of European ancestry (Zucchero et al. 2004). Another IRF6 locus, rs642961, has been shown to be associated with NSCL/P but not NSCPO (Rahimov et al. 2008). Corollary to these observations, some studies (Birnbaum et al. 2009; Kerameddin et al. 2015) have confirmed a role of IRF6 as a NSCL/P risk locus in populations of Asian and European ancestries. Other candidate genes implicated in the etiology of NSCL/P included MSX1 (Rafighdoost et al. 2013), BMP4 (Suzuki et al. 2009), FOXE1 (Moreno et al. 2009), AXIN2 (Letra et al. 2012), CRISPLD2 (Chiquet et al. 2007), NOG1, and FGFR2 (Leslie et al. 2015).

Among Africans, genetic studies on OFCs are limited. A study involving a Nigerian cohort implicated MSX1, but not other loci, in the etiology of NSCL/P (Butali et al. 2011). Other studies that recruited Kenyans (Weatherley-White et al. 2011) and Congolese (Figueiredo et al. 2014) could not replicate the association for cleft susceptibility loci among Africans, probably due to the small sample size and population heterogeneity. Moreover, sequencing of GWAS loci in cohorts from Ethiopia and Nigeria reported some rare, potentially causative variants (Butali, Mossey, et al. 2014). Conducting genetic and genomics studies with a cleft cohort from Africa may identify novel and population-specific signals. However, it is also important for us to investigate the role of identified signals and biologically relevant genes from existing European and Asian studies in the African population. The present study aimed to replicate the association between reported GWASs and candidate gene loci in our NSCL/P cohort. We also tested the hypothesis that NSCL/P loci may contribute to NSCPO susceptibility in Africans. Finally, we screened for rare, potentially pathogenic variants in 7 candidate genes at risk loci usually associated with NSCL/P.

Subjects and Methods

We recruited 3,585 participants from Ghana, Ethiopia, and Nigeria (Table 1; Appendix Methods). All sample and data collection at various study sites were approved by the local institutional review boards: College of Health Sciences, KNUST (Ghana; CHRPE/AP/217/13); College of Medicine, University of Lagos (Nigeria; ADM/DCST/HREC/APP/1374); and College of Health Sciences, Addis Ababa University (Ethiopia; 3.10/ 027/2015). Before sample and data collection, written informed consent was obtained from each participating family. DNA processing is shown in the Appendix Methods.

Table 1.

Subphenotypes, Sex, and Sample Types of Study Cohort That Passed Quality Control Checks and Were Included in Statistical Analyses.

Cleft Subphenotype of Probands Samples per Population, n
Ghana Ethiopia Nigeria Total
Case-control cohort
NSCL 162 101 77 340
NSCLP 144 143 74 361
NSCPO 102 21 40 163
Unrelated controls 408 357 313 1078
Case-parent trios
NSCL 52 2 20 74
NSCLP 48 3 26 77
NSCPO 34 1 7 42
Case-parent dyads
NSCL 77 84 51 212
NSCLP 76 134 47 257
NSCPO 53 20 32 105
Other trios
NSCL 18 0 0 18
NSCLP 14 0 0 14
NSCPO 11 0 0 11
Other dyads
NSCL 8 0 0 8
NSCLP 3 0 0 3
NSCPO 3 0 0 3
Singletons
NSCL 5 13 6 24
NSCLP 1 8 1 10
NSCPO 2 0 1 3
Tetrads
NSCLP 2 0 0 2
Pentads
NSCLP 1 0 0 1
Open in a new tab

Case probands consisted of 423 males and 441 females, whereas unrelated controls were made up of 441 males and 637 females. The probands in the case-control arm of the study are the same probands in the family-based studies. In some of the designated singletons, parental samples failed data cleaning and were dropped from statistical analyses—hence, the designation of such families as singletons. Singletons were informative in the case-control arm of our study but not the family-based studies. Tetrads and pentads were collected from families where 2 individuals were affected with clefts. “Other trios and dyads” largely refers to case-mother-maternal grandmother trios, case-mother-sibling trios, as well as case-siblings trios and dyads. Case-parent trios, tetrads, and pentads were employed in the transmission disequilibrium test, whereas all sample types, except singletons and unrelated controls, were used for analyses of the family-based association for disease traits. Only case probands and unrelated controls were included in the case-control analyses.

NSCL, nonsyndromic cleft lip; NSCL/P, nonsyndromic cleft lip with or without cleft palate; NSCLP, nonsyndromic cleft lip and palate; NSCPO, nonsyndromic cleft palate only.

Single-Nucleotide Polymorphism Selection

We selected single-nucleotide polymorphisms (SNPs) with a minor allele frequency (MAF) ≥5% in the African population for genotyping; these were previously reported in peer review journals or identified in animal studies and during our resequencing studies. These include SNPs that are associated with NSCL/P in candidate genes studies and GWASs in European and Asian populations (Appendix Table 1).

SNP Genotyping

We genotyped 48 SNPs (Appendix Table 1) on a total of 3,585 samples—872 NSOFC cases (163 NSCPO, 340 NSCL, 361 NSCLP, and 8 “untyped”), 1,635 unaffected relatives, and 1,078 unrelated controls—with the 192.24 Fluidigm SNP genotyping protocol (Appendix Methods). The “untyped” samples (from probands) and other samples, however, failed quality control checks and were not included in the final statistical analyses (Table 1).

Statistical Analyses for Association Studies

During quality control checks, we resolved Mendelian errors in case-parent triads and dropped from the final analyses samples that were not successfully genotyped on at least 95% of the 48 genotyped SNPs. We computed Hardy-Weinberg equilibrium (HWE) through PLINK (http://pngu.mgh.harvard.edu/~purcell/plink/). We then conducted 1) case-control analyses to determine associations in each subpopulation and 2) meta-analyses of the 3 subpopulations based on Table 1. For this test, we used P < 0.05 to denote nominal association and a Bonferroni correction of 141 tests to ascertain a threshold for formal significance of P = 3.54 × 10−4. The 141 tests comprised 47 SNPs that passed HWE × 3 cleft subphenotypes × 1 racial group × 1 test. Of the 48 SNPs, only 1 failed HWE (P < 0.05). Additional analyses to determine overtransmission of the rare alleles were conducted with the transmission disequilibrium test (TDT) and through the family-based association for disease traits (DFAM). The TDT used only the case-parent triad information (Table 1), while the DFAM allowed us to combine triad and dyad data. For these tests, the significant P value was 0.05. Parent-of-origin effects and gene-gene interactions (epistasis) were also calculated. The probands in the case-control arm of the study (Table 1) are the same probands in the family-based studies.

DNA Sequencing

We directly sequenced VAX1, PAX7, ARHGAP29, MSX1, FOXE1, BMP4, and MAFB in 184 NSOFC cases (131 NSCL/P and 53 NSCPO) from Ghana using Sanger Sequencing (Appendix Methods; Butali, Mossey, et al. 2014). We also performed segregation analyses on observed potentially pathogenic missense, frameshift, and splice site variants by sequencing available parental samples. We further sequenced 96 unrelated Ghanaian controls to ascertain whether the novel variants that we encountered in NSOFC cases also occurred in controls.

Results

Association Analyses

In meta-analyses of the case-control cohorts from the 3 subpopulations, we successfully demonstrated nominal association of PAX7 (rs742071, P = 5.10 × 10−3), 8q24 (rs987525, P = 1.22 × 10−3), as well as VAX1 (rs7078160, P = 0.04) with NSCL/P; in addition, MSX1 (rs115200552, P = 0.01), TULP4 (rs651333, P = 0.04), CRISPLD2 (rs4783099, P = 0.02), and NOG1 (rs17760296, P = 0.04) were nominally associated with NSCPO (Table 2), with the direction of effect being the same as reported by earlier studies. Among Ethiopians (Appendix Table 2), PAX7 (rs742071, P = 5.57 × 10−3), IRF6 (rs642961, P = 0.02), DYSF (rs2303596, P = 2.31 × 10−3), 8q24 (rs987525, P = 7.82 × 10−4), and MAFB (rs13041247 and rs11696257, all with P = 0.04) were nominally associated with NSCL/P; ABCA4 (rs481931 and rs4147811, all with P = 0.03) and NTN1 (rs8081823, P = 0.03) were nominally associated with NSCPO. Moreover, subphenotype analyses of the Ethiopian NSCL/P cohort showed that the PAX7, DYSF, MSX1, SPRY2 (rs9574565, P = 7.05 × 10−3) and MAFB signals were particularly stronger for NSCL, whereas the IRF6 (rs642961, P = 9.11 × 10−3) and 8q24 (rs987525, P = 1.07 × 10−3) signals were stronger for NSCLP (Appendix Table 2). Among Ghanaians (Appendix Table 3), ABCA4 (rs560426, P = 0.03) and VAX1 (rs7078160, P = 0.03) were nominally associated with NSCL/P with subphenotype analyses of the NSCL/P cohort showing that the ABCA4 locus was strongly associated with NSCLP. ABCA4 (rs4147811, P = 7.48 × 10−3) and CRISPLD2 (rs4783099, P = 0.04) were nominally associated with NSCL/P and NSCPO, respectively, among Nigerians (Appendix Table 4). Subphenotype analyses of the Nigerian NSCL/P (Appendix Table 4) showed that PAX7 (rs742071, P = 0.02) and ARHGAP29 (rs138751793, P = 0.04) signals were stronger for NSCL, whereas another SNP at the ABCA4 locus (rs481931, P = 2.87 × 10−3) was strongly associated with NSCLP. However, none of these case-control associations passed Bonferroni correction.

Table 2.

Meta-analyses of the Case-Control Cohorts from Ghana, Ethiopia, and Nigeria.

Part A: Meta-analyses of NSCL/P and NSCPO Case-Control Cohorts from All 3 Countries
NSCL/P
 NSCPO
SNP Probable Gene/Loci Minor Allelesa African MAF P OR I P OR I
rs1801131 MTHFR C/Ab 0.15 0.32 1.08 0.00 0.19 0.79 0.00
rs1801133 MTHFR A/Gc 0.09 0.49 1.08 18.19 0.44 0.83 0.00
rs766325 PAX7 G/Ab,d,e 0.18 0.29 0.92 0.00 0.23 0.82 0.00
rs742071 PAX7 T/Gb 0.39 5.10E-03f 1.19 54.68 0.76 0.96 0.00
rs560426 ABCA4 C/Tb,g 0.49 0.10 0.90 6.15 0.16 1.18 0.00
rs481931 ABCA4 T/Gc 0.10 0.40 1.09 11.13 0.49 0.85 0.00
rs4147811 ABCA4 T/Cc 0.11 0.23 1.13 67.35 0.93 1.02 0.00
rs138751793 ARHGAP29 C/Th 0.02 0.24 1.32 0.00 0.47 1.34 27.90
rs6677101 SLC25A24 G/Tb,e,g 0.33 0.80 0.98 12.11 0.87 1.02 53.89
rs861020 IRF6 A/Gb 0.11 0.23 1.11 0.00 0.83 0.96 24.15
rs34743335 IRF6 T/A 0.02 0.59 0.90 0.00 0.84 0.89 38.34
rs642961 IRF6 A/Gb 0.09 0.32 1.11 68.47 0.57 0.88 44.17
rs7590268 THADA G/Tb 0.20 0.74 0.98 0.00 0.38 0.87 0.00
rs4332945 DYSF T/Gb,e,g 0.16 0.94 0.99 0.00 0.97 1.01 0.00
rs2303596 DYSF T/Cc,d,e 0.22 0.20 0.91 75.32 0.57 1.09 73.54
rs227782 DYSF A/Gb,g 0.42 0.33 1.06 0.00 0.35 1.12 61.90
rs115200552 MSX1 C/Gh 0.02 0.38 1.16 28.63 0.01f 1.81 0.00
rs12532 MSX1 G/Ac,e 0.44 0.49 0.96 0.00 0.37 0.90 0.43
rs2674394 Gene desert A/Cb 0.17 0.62 1.04 0.00 0.68 1.07 0.00
rs651333 TULP4 C/Tb,d,g 0.34 0.97 1.00 0.00 0.04f 1.29 0.00
rs6558002 EPHX2 C/Tb,g 0.24 0.39 1.06 0.00 0.87 1.02 0.00
rs987525 8q24 A/Cb,g 0.38 1.22E-03f 0.81 40.55 0.22 0.86 0.00
rs894673 FOXE1 A/Tc 0.33 0.42 0.95 0.00 0.93 1.01 0.00
rs3758249 FOXE1 T/Cc 0.33 0.56 0.96 0.00 0.90 1.02 0.00
rs7078160 VAX1 A/Gb 0.25 0.04f 1.16 0.00 0.88 1.02 0.00
rs4752028 VAX1 C/Tb,g 0.45 0.51 0.96 0.00 0.80 0.97 0.00
rs10785430 ADAMTS20 G/Ab 0.32 0.90 0.99 0.00 0.49 1.09 0.00
rs9574565 SPRY2 T/Cc,g 0.35 0.75 1.02 0.00 0.45 1.10 0.00
rs8001641 SPRY2 G/Ac,d,e,g 0.10 0.35 1.08 0.00 0.37 0.85 0.00
rs17563 BMP4 T/Cb,d,e,g 0.18 0.95 0.99 0.00 0.77 1.04 0.00
rs1258763 GREM1 C/Tc,d,e,g 0.49 0.11 1.11 0.00 0.50 0.92 0.00
rs8049367 ADCY9 C/Tc,d,e 0.30 0.20 1.09 0.00 0.10 0.81 0.00
rs16260 CDH1 A/Cb 0.13 0.59 1.05 0.00 0.39 0.85 0.00
rs11642413 CDH1 G/Ab,e,g 0.28 0.83 1.02 0.00 0.21 0.83 0.00
rs1546124 CRISPLD2 G/Cb,e 0.25 0.60 0.96 0.00 0.89 0.98 0.00
rs4783099 CRISPLD2 T/Cb 0.33 0.59 1.04 0.00 0.02f 0.74 0.00
rs8069536 NTN1 T/Gb 0.32 0.13 1.11 0.97 0.88 0.98 0.00
rs8081823 NTN1 A/Gc 0.24 0.08 0.88 0.00 0.63 0.94 32.54
rs17760296 NOG1 G/Tb 0.02 0.92 0.99 0.00 0.04f 1.74 0.00
rs227731 NOG1 G/Tb,g 0.22 0.86 0.99 0.00 0.26 1.17 0.00
rs7224837 AXIN2 G/Ab 0.11 0.75 1.04 0.00 0.81 0.95 0.00
rs3923086 AXIN2 A/Cb,d,e,g 0.02 0.25 1.15 0.00 NA NA NA
rs17820943 MAFB T/Cc 0.25 0.33 0.93 15.15 0.68 1.06 22.99
rs13041247 MAFB C/Tc 0.25 0.37 0.94 34.01 0.42 1.12 0.00
rs11696257 MAFB T/Cc 0.25 0.30 0.93 32.24 0.61 1.07 0.00
Part B: Meta-analyses of Subphenotypes of NSCL/P Cohorts from the 3 Countries
NSCL
 NSCLP
rs1801131 MTHFR C/Ab 0.15 0.78 1.03 0.00 0.22 1.13 0.00
rs1801133 MTHFR A/Gc 0.09 0.71 1.06 8.24 0.30 0.30 0.00
rs766325 PAX7 G/Ab,d,e 0.18 0.91 0.99 0.00 0.17 0.86 0.00
rs742071 PAX7 T/Gb 0.39 0.02f 1.23 68.74 0.03f 1.19 0.00
rs560426 ABCA4 C/Tb 0.49 0.73 1.03 0.00 0.03f 1.20 10.33
rs481931 ABCA4 T/Gc 0.10 0.81 0.97 0.00 0.08 1.27 63.75
rs4147811 ABCA4 T/Cc 0.11 0.50 1.10 65.82 0.15 1.21 15.35
rs138751793 ARHGAP29 C/Th 0.02 0.19 1.53 66.38 0.41 1.29 0.00
rs6677101 SLC25A24 G/Tb,e,g 0.33 0.92 0.99 0.00 0.98 1.00 58.97
rs861020 IRF6 A/Gb 0.11 0.18 1.17 17.72 0.57 1.07 0.00
rs34743335 IRF6 T/A 0.02 0.87 0.96 0.00 0.50 0.85 23.72
rs642961 IRF6 A/Gb 0.09 0.96 0.99 15.60 0.15 1.21 62.97
rs7590268 THADA G/Tb 0.20 0.45 0.92 0.00 0.50 1.07 0.00
rs4332945 DYSF T/Gb,e,g 0.16 0.54 0.94 10.40 0.71 1.04 0.00
rs2303596 DYSF T/Cc,d,e 0.22 0.29 0.89 63.58 0.44 0.93 75.54
rs227782 DYSF A/Gb,g 0.42 0.85 0.98 0.00 0.13 1.14 0.00
rs115200552 MSX1 C/Gh 0.02 0.18 1.37 61.30 0.68 1.10 0.00
rs12532 MSX1 G/Ac,e 0.44 0.55 0.95 0.00 0.51 0.95 0.00
rs2674394 Gene desert A/Cb 0.17 0.06 1.22 0.00 0.42 0.91 0.00
rs651333 TULP4 C/Tb,d,g 0.34 0.63 0.96 0.00 0.74 0.97 0.00
rs6558002 EPHX2 C/Tb,g 0.24 0.82 1.02 0.00 0.11 0.11 0.00
rs987525 8q24 A/Cb,g 0.38 5.38E-03f 1.28 0.00 0.01f 0.80 54.21
rs894673 FOXE1 A/Tc 0.33 0.54 0.95 42.39 0.45 0.94 0.00
rs3758249 FOXE1 T/Cc 0.33 0.53 0.94 46.73 0.68 0.96 0.00
rs7078160 VAX1 A/Gb 0.25 0.03f 1.23 0.00 0.20 1.13 24.04
rs4752028 VAX1 C/Tb,g 0.45 0.55 1.05 16.64 0.50 0.95 0.00
rs10785430 ADAMTS20 G/Ab 0.32 0.88 1.01 41.30 0.86 0.98 3.00
rs9574565 SPRY2 T/Cc,g 0.35 0.53 1.06 72.62 0.43 1.07 65.44
rs8001641 SPRY2 G/Ac,d,e,g 0.10 0.99 1.00 0.00 0.26 1.13 0.00
rs17563 BMP4 A/Gb,d,e,g 0.18 0.89 0.99 25.84 0.98 1.00 0.00
rs1258763 GREM1 C/Tc,d,e,g 0.49 0.22 0.90 0.00 0.10 1.15 0.00
rs8049367 ADCY9 C/Tc,d,e 0.30 0.36 1.09 10.19 0.35 1.08 0.00
rs16260 CDH1 A/Cb 0.13 0.46 0.91 10.51 0.20 1.16 0.00
rs11642413 CDH1 G/Ab,e,g 0.28 0.98 1.00 0.00 0.55 1.05 0.00
rs1546124 CRISPLD2 G/Cb,e 0.25 0.26 0.90 0.00 0.88 1.01 0.00
rs4783099 CRISPLD2 T/Cb 0.33 0.85 1.02 0.00 0.32 1.09 0.00
rs8069536 NTN1 T/Gb 0.32 0.72 1.03 3.47 0.04f 1.20 0.00
rs8081823 NTN1 A/Gc 0.24 0.55 0.95 0.00 0.05 0.83 0.00
rs17760296 NOG1 G/Tb 0.02 0.83 1.04 5.85 0.85 0.97 0.00
rs227731 NOG1 G/Tb,g 0.22 0.38 0.92 0.00 0.59 1.05 0.00
rs7224837 AXIN2 G/Ab 0.11 0.61 1.08 0.00 0.81 1.04 0.00
rs3923086 AXIN2 A/Cb,d,e,g 0.02 0.62 1.10 40.28 NA NA 0.00
rs17820943 MAFB T/Cc 0.25 0.25 0.89 15.55 0.43 0.93 0.00
rs13041247 MAFB C/Tc 0.25 0.25 0.89 31.03 0.54 0.94 0.00
rs11696257 MAFB T/Cc 0.25 0.24 0.89 27.17 0.40 0.92 0.00
Open in a new tab

All P values reported are for the minor alleles. All initial studies were carried out in Asians and/or Caucasians but not Africans. Source of minor alleles and MAF: http://browser.1000genomes.org.

I, test of heterogeneity of which 0 to 40 represents no heterogeneity; MAF, minor allele frequency; NA, not applicable; NSCL, nonsyndromic cleft lip; NSCL/P, nonsyndromic cleft lip with or without cleft palate; NSCLP, nonsyndromic cleft lip and palate; NSCPO, nonsyndromic cleft palate only; OR, odds ratio; SNP, single-nucleotide polymorphism.

a

The first allele is the minor allele in Europeans unless otherwise indicated. The first allele is also the minor allele in East Asians, South Asians, and Africans.

b

Minor allele was the risk allele in initial study.

c

Minor allele was protective in initial study.

d

The first allele is the major allele, while the second allele is the minor allele in South Asians.

e

The first allele is the major allele, while the second allele is the minor allele in East Asians.

f

Loci that reached nominal significance in meta-analyses (in bold).

g

The first allele is the major allele, while the second allele is the minor allele in Africans.

h

The first allele is the minor allele, and the variation exists only in Africans.

For the TDT and DFAM (Tables 3 and 4) for all 3 subpopulations, 7 loci demonstrated formal significance with NSOFCs at P ≤ 0.05. Formal significance for the TDT and DFAM was evaluated at P ≤ 0.05 because these are secondary analyses compared with case-control analyses and are not true independent tests. All family-based studies suggested that the minor allele of ABCA4 (rs560426) was overtransmitted in NSCLP cases among Africans. PAX7 (rs742071) also consistently showed evidence of overtransmission in NSCL cases in the TDT and DFAM. MSX1 (rs115200552) and AXIN2 (rs3923086) also demonstrated strong overtransmission in NSCLP cases in DFAM analyses, whereas MTHFR (rs1801131) and DYSF exhibited overtransmission in NSCL cases in TDT and DFAM analyses, respectively. Only an SNP of VAX1 demonstrated overtransmission in NSCPO cases.

Table 3.

Transmission Disequilibrium Test for Case-Parent Trios Only.

Part A: Transmission Disequilibrium Test Analyses for NSCL/P and NSCPO
NSCL/P NSCPO
SNP Probable Gene/Loci T:NT P OR (95% CI) T:NT P OR (95% CI)
rs1801131 MTHFR 27:34 0.37 0.79 (0.48 to 1.32) 10:9 0.82 1.11 (0.45 to 2.73)
rs1801133 MTHFR 22:23 0.88 0.96 (0.53 to 1.72) 6:8 0.59 0.75 (0.26 to 2.16)
rs766325 PAX7 43:52 0.36 0.83 (0.55 to 1.24) 11:11 1.00 1.00 (0.43 to 2.31)
rs742071 PAX7 82:75 0.58 1.09 (0.80 to 1.50) 16:11 0.34 1.46 (0.68 to 3.13)
rs560426 ABCA4 78:59 0.10 1.32 (0.94 to 1.85) 18:18 1.00 1.00 (0.52 to 1.92)
rs481931 ABCA4 28:25 0.68 1.12 (0.65 to 1.92) 3:8 0.13 0.38 (0.10 to 1.41)
rs4147811 ABCA4 26:25 0.89 1.04 (0.60 to 1.80) 5:10 0.20 0.50 (0.17 to 1.46)
rs138751793 ARHGAP29 5:7 0.56 0.71 (0.23 to 2.25) 1:2 0.56 0.50 (0.05 to 5.51)
rs6677101 SLC25A24 65:75 0.40 0.87 (0.62 to 1.21) 21:14 0.24 1.50 (0.76 to 2.95)
rs861020 IRF6 35:29 0.45 1.21 (0.74 to 1.97) 3:7 0.21 0.43 (0.11 to 1.66)
rs34743335 IRF6 4:2 0.41 2.00 (0.37 to 10.92) 0:0 NA NA (NA)
rs642961 IRF6 29:29 1.00 1.00 (0.60 to 1.67) 2:7 0.10 0.29 (0.06 to 1.38)
rs7590268 THADA 49:48 0.92 1.02 (0.69 to 1.52) 8:8 1.00 1.00 (0.38 to 2.66)
rs4332945 DYSF 43:40 0.74 1.08 (0.70 to 1.65) 11:8 0.49 1.38 (0.55 to 3.42)
rs2303596 DYSF 45:57 0.23 0.79 (0.53 to 1.18) 12:8 0.37 1.50 (0.61 to 3.67)
rs227782 DYSF 73:65 0.50 1.12 (0.80 to 1.57) 20:13 0.22 1.54 (0.77 to 3.09)
rs115200552 MSX1 10:13 0.53 0.77 (0.34 to 1.75) 7:2 0.10 3.50 (0.72 to 16.85)
rs12532 MSX1 77:71 0.62 1.09 (0.79 to 1.50) 20:22 0.76 0.91 (0.50 to 1.67)
rs2674394 Gene desert 40:44 0.66 0.91 (0.59 to 1.40) 9:9 1.00 1.00 (0.40 to 2.52)
rs651333 TULP4 56:59 0.78 0.95 (0.66 to 1.37) 21:16 0.41 1.31 (0.68 to 2.52)
rs6558002 EPHX2 47:40 0.45 1.18 (0.77 to 1.79) 13:12 0.84 1.08 (0.49 to 2.37)
rs987525 8q24 71:59 0.29 1.20 (0.85 to 1.70) 19:20 0.87 0.95 (0.51 to 1.78)
rs894673 FOXE1 60:67 0.53 0.90 (0.63 to 1.29) 16:15 0.86 1.07 (0.53 to 2.16)
rs3758249 FOXE1 59:66 0.53 0.89 (0.63 to 1.27) 16:15 0.86 1.07 (0.53 to 2.16)
rs7078160 VAX1 60:44 0.12 1.36 (0.92 to 2.01) 18:10 0.13 1.80 (0.83 to 3.90)
rs4752028 VAX1 73:76 0.81 0.96 (0.70 to 1.32) 27:13 0.03a 2.08 (1.07 to 4.03)
rs10785430 ADAMTS20 61:59 0.86 1.03 (0.72 to 1.48) 15:11 0.43 1.36 (0.63 to 2.97)
rs9574565 SPRY2 69:55 0.21 1.26 (0.88 to 1.79) 18:17 0.87 1.06 (0.55 to 2.05)
rs8001641 SPRY2 22:22 1.00 1.00 (0.55 to 1.81) 9:6 0.44 1.50 (0.53 to 4.21)
rs17563 BMP4 44:44 1.00 1.00 (0.66 to 1.52) 10:15 0.32 0.67 (0.30 to 1.48)
rs1258763 GREM1 73:58 0.19 1.26 (0.89 to 1.78) 19:21 0.75 0.90 (0.49 to 1.68)
rs8049367 ADCY9 67:67 1.00 1.00 (0.71 to 1.40) 12:13 0.84 0.92 (0.42 to 2.02)
rs16260 CDH1 31:28 0.70 1.11 (0.66 to 1.85) 6:13 0.11 0.46 (0.18 to 1.21)
rs11642413 CDH1 62:49 0.22 1.27 (0.87 to 1.84) 14:11 0.55 1.27 (0.58 to 2.80)
rs1546124 CRISPLD2 53:44 0.36 1.21 (0.81 to 1.80) 9:14 0.30 0.64 (0.28 to 1.49)
rs4783099 CRISPLD2 75:64 0.35 1.17 (0.84 to 1.64) 15:21 0.32 0.71 (0.37 to 1.39)
rs8069536 NTN1 67:70 0.80 0.96 (0.68 to 1.34) 14:13 0.85 1.08 (0.51 to 2.29)
rs8081823 NTN1 58:56 0.85 1.04 (0.72 to 1.50) 14:15 0.85 0.93 (0.45 to 1.93)
rs17760296 NOG1 7:8 0.80 0.88 (0.32 to 2.41) 2:0 0.16 NA (NA)
rs227731 NOG1 47:49 0.84 0.96 (0.64 to 1.43) 20:11 0.11 1.82 (0.87 to 3.80)
rs7224837 AXIN2 19:27 0.24 0.70 (0.39 to 1.27) 1:6 0.06 0.17 (0.02 to 1.38)
rs3923086 AXIN2 2:3 0.65 0.67 (0.11 to 3.99) 1:0 0.32 NA (NA)
rs17820943 MAFB 49:42 0.46 1.17 (0.77 to 1.76) 15:12 0.56 1.25 (0.59 to 2.67)
rs13041247 MAFB 49:43 0.53 1.14 (0.76 to 1.72) 15:12 0.56 1.25 (0.59 to 2.67)
rs11696257 MAFB 48:43 0.60 1.12 (0.74 to 1.69) 14:12 0.69 1.17 (0.54 to 2.52)
Part B: Transmission Disequilibrium Test Subphenotype Analyses for NSCL/P
NSCL
 NSCLP
rs1801131 MTHFR 9:20 0.04a 0.45 (0.20 to 0.99) 18:14 0.48 1.29 (0.64 to 2.59)
rs1801133 MTHFR 7:8 0.80 0.88 (0.31 to 2.41) 15:15 1.00 1.00 (0.49 to 2.05)
rs766325 PAX7 18:24 0.35 0.75 (0.41 to 1.38) 25:28 0.68 0.89 (0.52 to 1.53)
rs742071 PAX7 50:30 0.03a 1.67 (1.06 to 2.62) 32:45 0.14 0.71 (0.45 to 1.12)
rs560426 ABCA4 32:35 0.71 0.91 (0.57 to 1.48) 46:24 8.55E-03a 1.92 (1.17 to 3.14)
rs481931 ABCA4 10:13 0.53 0.77 (0.34 to 1.75) 18:12 0.27 1.50 (0.72 to 3.14)
rs4147811 ABCA4 8:10 0.64 0.80 (0.32 to 2.03) 18:15 0.60 1.20 (0.60 to 2.38)
rs138751793 ARHGAP29 1:2 0.56 0.50 (0.05 to 5.51) 4:5 0.74 0.80 (0.21 to 2.98)
rs6677101 SLC25A24 26:41 0.07 0.63 (0.39 to 1.04) 39:34 0.56 1.15 (0.72 to 1.82)
rs861020 IRF6 20:14 0.30 1.43 (0.72 to 2.83) 15:15 1.00 1.00 (0.49 to 2.05)
rs34743335 IRF6 2:1 0.56 2.00 (0.18 to 22.06) 2:1 0.56 2.00 (0.18 to 22.06)
rs642961 IRF6 16:15 0.86 1.07 (0.53 to 2.16) 13:14 0.85 0.93 (0.44 to 1.98)
rs7590268 THADA 21:32 0.13 0.66 (0.38 to 1.14) 28:16 0.07 1.75 (0.95 to 3.23)
rs4332945 DYSF 21:17 0.52 1.24 (0.65 to 2.34) 22:23 0.88 0.96 (0.53 to 1.72)
rs2303596 DYSF 18:22 0.53 0.82 (0.44 to 1.53) 27:35 0.31 0.77 (0.47 to 1.27)
rs227782 DYSF 33:28 0.52 1.18 (0.71 to 1.95) 40:37 0.73 1.08 (0.69 to 1.69)
rs115200552 MSX1 6:3 0.32 2.00 (0.50 to 8.00) 4:10 0.11 0.40 (0.13 to 1.28)
rs12532 MSX1 39:32 0.41 1.22 (0.76 to 1.95) 38:39 0.91 0.97 (0.62 to 1.52)
rs2674394 Gene desert 21:17 0.52 1.24 (0.65 to 2.34) 19:27 0.24 0.70 (0.39 to 1.27)
rs651333 TULP4 26:26 1.00 1.00 (0.58 to 1.72) 30:33 0.71 0.91 (0.55 to 1.49)
rs6558002 EPHX2 15:18 0.60 0.83 (0.42 to 1.65) 32:22 0.17 1.46 (0.85 to 2.50)
rs987525 8q24 35:28 0.38 1.25 (0.76 to 2.06) 36:31 0.54 1.16 (0.72 to 1.88)
rs894673 FOXE1 27:31 0.60 0.87 (0.52 to 1.46) 33:36 0.72 0.92 (0.57 to 1.47)
rs3758249 FOXE1 27:31 0.60 0.87 (0.52 to 1.46) 32:35 0.71 0.91 (0.57 to 1.48)
rs7078160 VAX1 37:23 0.07 1.61 (0.96 to 2.71) 23:21 0.76 1.10 (0.61 to 1.98)
rs4752028 VAX1 32:38 0.47 0.84 (0.53 to 1.35) 41:38 0.74 1.08 (0.69 to 1.68)
rs10785430 ADAMTS20 25:28 0.68 0.89 (0.52 to 1.53) 36:31 0.54 1.16 (0.72 to 1.88)
rs9574565 SPRY2 35:29 0.45 1.21 (0.74 to 1.97) 34:26 0.30 1.31 (0.78 to 2.18)
rs8001641 SPRY2 12:12 1.00 1.00 (0.45 to 2.27) 10:10 1.00 1.00 (0.42 to 2.40)
rs17563 BMP4 22:16 0.33 1.38 (0.72 to 2.62) 22:28 0.40 0.79 (0.45 to 1.37)
rs1258763 GREM1 31:27 0.60 1.15 (0.69 to 1.92) 42:31 0.20 1.36 (0.85 to 2.16)
rs8049367 ADCY9 25:28 0.68 0.89 (0.52 to 1.53) 42:39 0.74 1.08 (0.70 to 1.67)
rs16260 CDH1 12:14 0.69 0.86 (0.40 to 1.85) 19:14 0.38 1.36 (0.68 to 2.71)
rs11642413 CDH1 25:22 0.66 1.14 (0.64 to 2.02) 37:27 0.21 1.37 (0.83 to 2.25)
rs1546124 CRISPLD2 25:22 0.66 1.14 (0.61 to 2.02) 28:22 0.40 1.27 (0.73 to 2.23)
rs4783099 CRISPLD2 39:35 0.64 1.11 (0.71 to 1.76) 36:29 0.39 1.24 (0.76 to 2.02)
rs8069536 NTN1 32:35 0.71 0.91 (0.57 to 1.48) 35:35 1.00 1.00 (0.63 to 1.60)
rs8081823 NTN1 30:20 0.16 1.50 (0.85 to 2.64) 28:36 0.32 0.78 (0.47 to 1.27)
rs17760296 NOG1 5:2 0.26 2.50 (0.49 to 12.89) 2:6 0.16 0.33 (0.07 to 1.65)
rs227731 NOG1 22:26 0.56 0.85 (0.48 to 1.49) 25:23 0.77 1.09 (0.62 to 1.92)
rs7224837 AXIN2 10:9 0.82 1.11 (0.45 to 2.73) 9:18 0.08 0.50 (0.22 to 1.11)
rs3923086 AXIN2 1:2 0.56 0.50 (0.05 to 5.51) 1:1 1.00 1.00 (0.06 to 15.99)
rs17820943 MAFB 18:22 0.53 0.82 (0.44 to 1.53) 31:20 0.12 1.55 (0.88 to 2.72)
rs13041247 MAFB 18:22 0.53 0.82 (0.44 to 1.53) 31:21 0.17 1.48 (0.85 to 2.57)
rs11696257 MAFB 18:22 0.53 0.82 (0.44 to 1.53) 30:21 0.21 1.43 (0.82 to 2.50)
Open in a new tab

95% CI, 95% confidence interval; NA, not applicable; NSCL, nonsyndromic cleft lip; NSCL/P, nonsyndromic cleft lip with or without cleft palate; NSCLP, nonsyndromic cleft lip and palate; NSCPO, nonsyndromic cleft palate only; NT, not transmitted; OR, odds ratio; SNP, single nucleotide polymorphism; T, transmitted.

a

Loci that demonstrated overtransmission at threshold significance of P ≤ 0.05 (in bold).

Table 4.

Family-Based Association for Disease Traits for Cases and Relatives.

P Values
SNP Probable Gene/Loci NSCL/P NSCL NSCLP NSCPO
rs1801131 MTHFR 0.70 0.68 0.24 0.67
rs1801133 MTHFR 0.82 0.51 0.59 0.29
rs766325 PAX7 0.61 0.71 0.74 0.24
rs742071 PAX7 0.32 0.02a 0.29 0.96
rs560426 ABCA4 2.59E-02a 0.72 4.75E-03a 0.80
rs481931 ABCA4 0.15 0.55 0.16 0.61
rs4147811 ABCA4 0.29 0.44 0.48 0.51
rs138751793 ARHGAP29 0.38 0.66 0.43 0.40
rs6677101 SLC25A24 1.00 0.80 0.64 0.24
rs861020 IRF6 0.43 0.23 0.98 0.35
rs34743335 IRF6 0.32 0.52 0.47 0.61
rs642961 IRF6 0.83 0.99 0.98 0.15
rs11119388 SYT14 0.83 0.85 0.92 0.91
rs7590268 THADA 0.85 0.30 0.18 0.77
rs4332945 DYSF 0.04a 0.02a 0.60 0.62
rs2303596 DYSF 0.81 0.84 0.53 0.60
rs227782 DYSF 0.36 0.48 0.55 0.47
rs115200552 MSX1 0.89 0.13 3.50E-02a 0.08
rs12532 MSX1 0.67 0.96 0.30 0.43
rs2674394 Gene desert 0.59 0.11 0.58 0.51
rs651333 TULP4 0.92 0.90 0.63 0.20
rs6558002 EPHX2 0.38 0.77 0.27 0.52
rs987525 8q24 0.80 0.50 0.52 0.99
rs894673 FOXE1 0.69 0.88 0.46 0.55
rs3758249 FOXE1 0.69 0.86 0.46 0.55
rs7078160 VAX1 0.21 0.18 0.77 0.28
rs4752028 VAX1 0.88 0.44 0.30 0.06
rs10785430 ADAMTS20 0.84 0.86 0.62 0.66
rs9574565 SPRY2 0.07 0.16 0.28 0.22
rs8001641 SPRY2 0.32 0.19 0.88 0.64
rs375489721 MIR17HG NA NA NA NA
rs185831554 MIR17HG 0.32 0.32 NA NA
rs17563 BMP4 0.66 0.15 0.80 0.70
rs1258763 GREM1 0.14 1.00 0.06 0.98
rs8049367 ADCY9 0.23 0.24 0.56 0.18
rs16260 CDH1 0.59 0.59 0.36 0.46
rs11642413 CDH1 0.33 0.81 0.08 0.88
rs1546124 CRISPLD2 0.30 0.53 0.45 0.15
rs4783099 CRISPLD2 0.17 0.14 0.89 0.37
rs8069536 NTN1 0.58 0.47 0.87 0.23
rs8081823 NTN1 0.97 0.30 0.19 0.89
rs17760296 NOG1 0.63 0.25 0.97 0.63
rs227731 NOG1 0.24 0.41 0.43 0.09
rs7224837 AXIN2 0.20 0.75 0.12 0.35
rs3923086 AXIN2 0.89 0.70 2.88E-03a 0.85
rs17820943 MAFB 0.31 0.88 0.14 0.65
rs13041247 MAFB 0.37 0.83 0.21 0.63
rs11696257 MAFB 0.46 0.89 0.26 0.77
Open in a new tab

NA, not applicable; NSCL, nonsyndromic cleft lip; NSCL/P, nonsyndromic cleft lip with or without cleft palate; NSCLP, nonsyndromic cleft lip and palate; NSCPO, nonsyndromic cleft palate only; SNP, single-nucleotide polymorphism.

a

Loci that demonstrated overtransmission at threshold significance (in bold).

Parent-of-Origin Effects

Parent-of-origin effects were not observed for almost all SNPs, except rs16260 of CDH1. For rs16260, a trend toward association (P = 0.0764) was observed for all clefts. The rs16260 SNP exhibited a maternal imprinting or maternal overtransmission effect.

Gene-Gene Interactions

In gene-gene (G × G) or epistatic interactions, 3 SNPs exhibited evidence of epistasis with other SNPs. Each of these epistatic interactions yielded P = 0.02. A SNP for ABCA4, rs560426, interacted with Chr6, rs2674394 (gene desert). Moreover, rs2303596 of DYSF interacted with rs3923086 of AXIN2. Finally, rs8069536 of NTN1 interacted with rs17820943, rs13041247, and rs11696257, all of MAFB. However, none of these G × G interactions passed Bonferroni correction.

Direct DNA Sequencing of 7 Selected Genes

We observed several rare and/or novel variants in the 7 genes that we sequenced (Table 5, Appendix Table 5). “Rare variants,” as used here, refer to either a novel variant or a variant whose MAF is ≤1%. Some of these variants were predicted to be potentially pathogenic by various bioinformatics tools, whereas others were depicted as benign. A de novo occurrence could not be demonstrated for any of these variants, because either the variant was present in at least 1 parent, or not both parents were available for segregation analysis. Last, some of the novel variants that we observed occurred in controls (e.g., all VAX1 variants), whereas others were not observed in controls (e.g., all ARHGAP29 variants).

Table 5.

Novel, Rare, and Potentially Etiologic Variants Observed in Sequenced Genes.

Part A: Variants Observed in Cases and Some Parents but Not in Controls
HGVS HGVp Total No. of Cases with Variant Subphenotype of Cases with Variant Segregation Analyses
ARHGAP29
c.341-30T>A NA 1 NSCL NA
c.511-107T>C NA 2 NSCLP and NSCPO NA
c.967A>G p.Asn323Asp 1 NSCL Absent in father
c.1277delAinsTA p.Lys426IlefsTer6 1 NSCLP Absent in mother
c.1281+4A>G NA 1 NSCLP Observed in clinically unaffected mother
PAX7
c.1227G>A p.Leu409Leu 1 NSCL NA
Part B: Bioinformatics-Predicted Effects of Potentially Pathogenic Variants
HGVS Polyphen-2 SIFT Human Splice Finder RegulomeDB
ARHGAP29
c.341-30T>A NA NA Alteration of ESS site NA
c.511-107T>C NA NA Alteration of ESS site and creation of new ESE site NA
c.967A>G Benign Deleterious NA NA
c.1277delAinsTA NA NA NA NA
c.1281+4A>G NA NA Alteration of wildtype donor site NA
PAX7
c.1227G>A Benign Tolerated Alteration of an ESE site NA
Open in a new tab

All analyses were based on genome assembly number GRCh37/hg19, 2009 (http://genome.ucsc.edu).

ESE, exonic splicing enhancer; ESS, exonic splicing silencer; NA, not applicable; NSCLP, nonsyndromic cleft lip and palate; NSCL, nonsyndromic cleft lip only; NSCPO, nonsyndromic cleft palate only.

Discussion

We have successfully demonstrated associations (both nominal in case-control analyses and threshold in the TDT and DFAM analyses) between some loci and NSCL/P in cohorts from Africa. We also tested the hypothesis that these loci contribute to NSCPO in Africans, and we observed some interesting associations. The 8q24 locus exhibited the strongest nominal significance with NSCL/P in case-control meta-analyses, with the trends suggesting that this locus may be relevant in all 3 subpopulations. The test of heterogeneity also largely suggested the absence of heterogeneity at this locus among the 3 African populations. We observed that among Africans, the associated minor C allele of rs987525 (http://browser.1000genomes.org) conferred reduced susceptibility, while the major A allele is the risk allele. Irrespective of these differences in minor alleles, our result is in harmony with earlier studies (Birnbaum et al. 2009; Grant et al. 2009; Mangold et al. 2010; Beaty et al. 2010; Ludwig et al. 2012) demonstrating that the A allele of rs987525 is a risk allele for NSCL/P in Europeans. These observations suggest that the actual risk variant is (or variants are) in linkage disequilibrium with the A allele of rs987525. Fine mapping of the African haplotype (which is smaller in the 8q24 region) will help identify the risk variant (or variants). Our observations corroborate those made elsewhere (Beaty et al. 2010; Murray et al. 2012) suggesting that the varied ethnic association of the rs987525 allele largely depends on its MAF in various populations. Current evidence suggests that although the 8q24 window is a gene desert, it harbors very remote cis-acting craniofacial enhancer elements that regulate the expression of oncogenic MYC in the developing face; perturbation of this regulatory network leads to craniofacial dysmorphologies, including sporadic CL/P, in mice (Uslu et al. 2014).

The C677T (rs1801133) SNP of MTHFR but not A1298C (rs1801131) has largely been associated with reduced risk for NSCL/P in Asians (Zhao et al. 2014; Martinelli et al. 2015; Pan et al. 2015) and, to some extent, in European-derived populations (Estandia-Ortega et al. 2014; de Aguiar et al. 2015), though not all studies (Sozen et al. 2009) replicated the association. Interestingly, we have demonstrated in TDT analyses that MTHFR is significantly associated with NSCL among Africans and that it is the C minor allele of the A1298C (rs1801131) SNP that confers a reduced risk, suggesting that A is the risk allele. AXIN2 has been implicated in the etiology of NSOFCs in multiple populations, except Africans, with rs3923086 demonstrating an association with NSCLP among Asians (Letra et al. 2012). Other studies (Mostowska et al. 2012; de Araujo et al. 2015) have replicated the association between AXIN2 and NSCL/P. Here, we have demonstrated that rs3923086 (AXIN2) is also associated with NSCLP among Africans in DFAM analyses. Other candidate genes (e.g., DYSF) also showed evidence of association with NSOFCs among Africans, buttressing the relevance of this approach in etiologic “gene hunting.”

Other SNPs, other than the already-reported ones, may be responsible for the associations between certain loci and NSOFCs in some ethnicities. Through direct DNA sequencing of the MSX1 gene, we observed overtransmission of the minor allele of rs115200552 in NSOFC cases. Subsequent genotyping of this SNP in 3,585 individuals showed that this SNP was associated with NSCPO (P = 0.01) in case-control meta-analyses, although family-based studies suggest that this marker may be a risk allele for NSCLP. Earlier studies involving Africans from Nigeria implicated MSX1 in the etiology of NSCL/P (Butali et al. 2011).

We could not detect a formal association between some GWASs and candidate gene loci and NSCL/P, presupposing that 1) these loci may not play a role in the etiology of NSCL/P in Africans or 2) the genotyped SNPs may not be the tag SNPs for Africans. Lack of statistical power due to sample size and low MAF of the genotyped SNPs in Africans could also be possible reasons. For example, rs2235371—an SNP of IRF6 that is in high-linkage disequilibrium and the same locus as rs642961 and that has been associated with NSCL/P among mostly Asians (Sun et al. 2015) and in some Europeans (Zucchero et al. 2004)—does not exist in the African population (http://browser.1000genomes.org/index.html). It is also possible that even when no associations are detected between reported loci and NSOFCs, potentially pathogenic variants may be observed in NSOFC cases. Therefore, GWASs and whole genome sequencing of NSOFC cases from Africa are required to detect more risk loci.

Subphenotype and subpopulation analyses (even among the same racial group) may be crucial in detecting an association between certain loci and NSOFCs. In both TDT and DFAM analyses, we observed that rs560426 of ABCA4 was associated with NSCLP but not the other OFC subphenotypes. Case-control analyses further suggested that the ABCA4 locus may be crucial in NSOFC etiology in all 3 African populations. PAX7 (rs742071) exhibited nominal association with NSCL/P in case-control meta-analyses, with subpopulation analyses suggesting that this signal originated mainly from the Ethiopian and Nigerian cohorts that exhibited some level of heterogeneity. However, TDT and DFAM subphenotype analyses demonstrated that rs742071 exhibited overtransmission in NSCL cases in all 3 populations. In case-control meta-analyses, VAX1 (rs7078160) was nominally associated with NSCL/P, with subpopulation analyses suggesting the 2 West African countries (largely Ghana) drive this signal.

Rare variants, but not necessarily common variants, may account for the link between certain loci and NSOFCs. We observed many missense mutations and 1 frameshift mutation in sequenced genes. No de novo occurrence was observed for any of these variants due to the unavailability of some parental samples. Moreover, some of the novel variants were also observed in clinically unaffected parents and controls. We sequenced the novel variants in 96 controls from Ghana, and the likelihood of identifying these novel variants in more controls (i.e., >96) is possible. Nonetheless, these variants are absent in >1,000 individuals in the 1000 Genomes database (with >300 Africans), >61,000 individuals in the ExAC database, as well as 6,500 individuals in the EVS database. There is also the need to functionally validate the pathogenicity or otherwise of these variants in vivo. Rare variants in ARHGAP29 (Leslie et al. 2012), PAX7 and VAX1 (Butali et al. 2013; Leslie et al. 2015), BMP4 (Suzuki et al. 2009), FOXE1 (Moreno et al. 2009), MAFB (Butali, Mossey, et al. 2014), and MSX1 (Liang et al. 2012) have been observed in NSOFC cases.

The incidence of OFC in Africans is much lower than in Europeans and Asians (Mossey and Modell 2012; Butali, Adeyemo, et al. 2014), even though these populations may share the same or similar genetic susceptibility loci for OFCs, as observed in the present study. Although underascertainment due to a lack of birth defect registries in most African countries could be a contributing factor (Butali, Adeyemo, et al. 2014), the low incidence of OFCs among Africans may be real, as African-derived populations in the Caribbean have a low OFC incidence similar to that of their ancestral population (Mossey and Modell 2012). We therefore hypothesize the possible existence of genetic protective variants in the African genome, whose “rescue mission” reduces clefting. The identification and elucidation of such protective variants can be translated to European and Asian populations to bring about reduced OFC incidence and eventually prevention.

Conclusion

The present study has shown evidence of an association of certain loci with NSOFCs at both nominal and threshold significance. For instance, we have for the first time shown that the 8q.24 locus is a risk locus in Africans. Our study has thus corroborated an earlier suggestion that the 8q24 locus may be a risk locus for NSCL/P across major ethnicities, although the effect size is smaller in Asians due to a lower MAF. Subphenotype as well as subpopulation analyses and genotyping of other SNPs, other than those already reported for some loci, may be crucial in identifying NSOFC loci in various ethnicities and populations. We have also demonstrated the existence of rare variants, both novel and known, in NSOFC cases from Africa. In conclusion, we have for the first time demonstrated associations between the SNPs that we studied and NSOFC among Africans. Our study is crucial for understanding the genetic architecture of NSOFCs in Africans and further suggests the need to carry out GWASs and whole genome sequencing for every ethnicity as far as complex traits are concerned.

Author Contributions

L.J.J. Gowans, contributed to conception, design, data acquisition, analysis, and interpretation, drafted and critically revised the manuscript; W.L. Adeyemo and M. Eshete, contributed to conception, design, and data acquisition, critically revised the manuscript; P.A. Mossey, contributed to conception, data acquisition, and analysis, critically revised the manuscript; T. Busch, contributed to design, data acquisition, and interpretation, critically revised the manuscript; B. Aregbesola, contributed to data acquisition and critically revised the manuscript; P. Donkor, contributed to design, data acquisition, and interpretation, critically revised the manuscript; F.K.N. Arthur, contributed to design, data acquisition, and analysis, critically revised the manuscript; S.A. Bello, contributed to data acquisition, critically revised the manuscript; A. Martinez, M. Li, and E. Augustine-Akpan, contributed to data acquisition and analysis, critically revised the manuscript; W. Deressa, contributed to data acquisition, critically revised the manuscript; P. Twumasi, contributed to design, critically revised the manuscript; J. Olutayo, M. Deribew, P. Agbenorku, A.A. Oti, R. Braimah, G. Plange-Rhule, M. Gesses, S. Obiri-Yeboah, G.O. Oseni, P.B. Olaitan, L. Abdur-Rahman, F. Abate, T. Hailu, P. Gravem, and M.O. Ogunlewe, contributed to data acquisition, critically revised the manuscript; C.J. Buxó, M.L. Marazita, and A.A. Adeyemo, contributed to data analysis and interpretation, critically revised the manuscript; J.C. Murray and A. Butali, contributed to conception, design, data acquisition, analysis, and interpretation, critically revised the manuscript. All authors gave final approval and agree to be accountable for all aspects of the work.

Supplementary Material

Supplementary material

Supplementary Material

Supplementary material

Supplementary Material

Supplementary material

Supplementary Material

Supplementary material

Supplementary Material

Supplementary material

Supplementary Material

Supplementary material

Acknowledgments

The authors are grateful to all families that participated in this research. We also appreciate all the doctors and nurses at the various teaching hospitals where sample collections were carried out: Komfo Anokye Teaching Hospital (Ghana), University of Lagos Teaching Hospital (Nigeria), Obafemi Awolowo University Teaching Hospital (Nigeria), Ladoke Akintola University of Technology Teaching Hospital (Nigeria), and Yekatit 12 Hospital Medical college (Ethiopia). Contributions of members of the Murray and Butali laboratories, University of Iowa, are acknowledged.

Footnotes

This work was supported by the National Institute of Dental and Craniofacial Research (R00 DE022378) and the Robert Wood Johnson Foundation (72429; A.B.); the National Institutes of Health (R37 DE-08559 and DE-016148; J.C.M.); the Ghana Cleft Foundation (L.J.J.G.); the National Institute of Dental and Craniofacial Research (R01-DE009886 and R01-DE012472; M.L.M.); and the National Institute of Minority Health and Disparities (U54-MD007587, R25-MD007607; C.J.B.) and the National Institute of Dental and Craniofacial Research (K99-DE024571; C.J.B.).

The authors declare no potential conflicts of interest with respect to the authorship and/or publication of this article.

A supplemental appendix to this article is published electronically only at http://jdr.sagepub.com/supplemental.

References

  1. Beaty TH, Murray JC, Marazita ML, Munger RG, Ruczinski I, Hetmanski JB, Liang KY, Wu T, Murray T, Fallin MD, et al. 2010. A genome-wide association study of cleft lip with and without cleft palate identifies risk variants near MAFB and ABCA4. Nat Genet. 42(6):525–529. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Birnbaum S, Ludwig KU, Reutter H, Herms S, Steffens M, Rubini M, Baluardo C, Ferrian M, Almeida de, Assis N, Alblas MA, et al. 2009. Key susceptibility locus for nonsyndromic cleft lip with or without cleft palate on chromosome 8q24. Nat Genet. 41(4):473–477. [DOI] [PubMed] [Google Scholar]
  3. Butali A, Adeyemo WL, Mossey PA, Olasoji HO, Onah II, Adebola A, Efunkoya Akintububo A, James O, Adeosun OO, et al. 2014. Prevalence of orofacial clefts in Nigeria. Cleft Plate Craniofac J. 51(3):320–325. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Butali A, Mossey P, Adeyemo W, Eshete M, Gaines L, Braimah R, Aregbesola B, Rigdon J, Emeka C, Olutayo J, et al. 2014. Rare functional variants in genome-wide association identified candidate genes for non-syndromic clefts in the African population. Am J Med Genet A. 164(10):2567–2571. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Butali A, Mossey PA, Adeyemo WL, Jezewski PA, Onwuamah CK, Ogunlewe MO, Ugboko VI, Adejuyigbe O, Adigun AI, Abdur-Rahman LO, et al. 2011. Genetic studies in the Nigerian population implicate an MSX1 mutation in complex oral facial clefting disorders. Cleft Palate Craniofac J. 48(6):646–653. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Butali A, Suzuki S, Cooper ME, Mansilla AM, Cuenco K, Leslie EJ, Suzuki Y, Niimi T, Yamamoto M, Ayanga G, et al. 2013. Replication of genome wide association identified candidate genes confirm the role of common and rare variants in PAX7 and VAX1 in the etiology of nonsyndromic CL[P]. Am J Med Genet A. 161(5):965–972. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Chiquet BT, Lidral AC, Stal S, Mulliken JB, Moreno LM, Arcos-Burgos M, Valencia-Ramirez C, Blanto SH, Hetch JT. 2007. CRISPLD2: a novel NSCLP candidate gene. Hum Mol Genet. 16(18):2241–2248. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. de Aguiar PK, Coletta RD, de Oliveira AM, Machado RA, Furtado PG, de Oliveira LA, de Aquino SN, Martelli-Junior H, de Almeida, Reis SR, Moreira HS, et al. 2015. Rs1801133C>T polymorphism is a risk factor for nonsyndromic cleft lip with or without cleft palate in the Brazilian population. Birth Defects Res A Clin Mol Teratol. 103(4):292–298. [DOI] [PubMed] [Google Scholar]
  9. de Araujo TK, Secolin R, Félix TM, de Souza LT, Fontes MÍ, Monlleó IL, de Souza J, Fett-Conte AC, Ribeiro EM, Xavier AC, et al. 2015. A multicentric association study between 39 genes and nonsyndromic cleft lip and palate in a Brazilian population. J Craniofacial Surg. 44(1):16–20. [DOI] [PubMed] [Google Scholar]
  10. Dixon MJ, Marazita ML, Beaty TH, Murray JC. 2011. Cleft lip and palate: understanding genetic and environmental influences. Nat Rev Genet. 12(3):167–178. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Estandia-Ortega B, Velázquez-Aragón JA, Alcántara-Ortigoza MA, Reyna-Fabian ME, Villagómez-Martínez S, González-Del Angel A. 2014. 5,10-Methylenetetrahydrofolate reductase single nucleotide polymorphisms and gene-enevironment interaction analysis in non-syndromic cleft lip/palate. Eur J Oral Sci. 122(2):109–113. [DOI] [PubMed] [Google Scholar]
  12. Figueiredo JC, Ly S, Raimondi H, Magee K, Baurley JW, Sanchez-Lara PA, Ihenacho U, Yao C, Edlund CK, van den Berg D, et al. 2014. Genetic risk factors for orofacial clefts in Central Africans and Southeast Asians. Am J Med Genet A. 164(10):2572–2580. [DOI] [PubMed] [Google Scholar]
  13. Gorlin RJ, Cohen MM, Hennekam RCM. 2001. Syndromes of the neck and head. New York (NY): Oxford University Press. [Google Scholar]
  14. Grant SF, Wang K, Zhang H, Glaberson W, Annaiah K, Kim CE, Bradfield JP, Glessner JT, Thomas KA, Garris M, et al. 2009. A genome-wide association study identifies a locus for non-syndromic cleft lip with or without cleft palate on 8q.24. J Pediatr. 155(6):909–913. [DOI] [PubMed] [Google Scholar]
  15. Kerameddin S, Namipashaki A, Ebrahimi S, Ansari-Pour N. 2015. IRF6 is a marker of severity in nonsyndromic cleft lip/palate. J Dent Res. 94(9):226S–232S. [DOI] [PubMed] [Google Scholar]
  16. Kondo S, Schutte BC, Richardson RJ, Bjork BC, Knight AS, Watanabe Y, Howard E, de Lima RL, Daack-Hirsch S, Sander A, et al. 2002. Mutations in IRF6 cause van der Woude and popliteal pterygium syndromes. Nat Genet. 32(2):285–289. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Leslie EJ, Mansilla MA, Biggs LC, Schuette K, Bullard S, Cooper M, Dunnwald M, Lidral AC, Marazita ML, Beaty TH, et al. 2012. Expression and mutation analyses implicate ARHGAP29 as the etiologic gene for the cleft lip with or without cleft palate locus identified by genome wide association on chromosome 1p22. Birth Defects Res A Clin Mol Teratol. 94(11):934–942. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Leslie EJ, Taub MA, Liu H, Steinberg KM, Koboldt DC, Zhang Q, Carlson JC, Hetmanski JB, Wang H, Larson DE, et al. 2015. Identification of functional variants for cleft lip with or without cleft palate in or near PAX7, FGFR2, and NOG by targeted sequencing of GWAS loci. Am J Hum Genet. 96(3):397–411. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Letra A, Bjork B, Cooper ME, Szabo-Rogers H, Deleyiannis FW, Field LL, Czeizel AE, Ma L, Garlet GP, Poletta FA, et al. 2012. Association of AXIN2 with non-syndromic oral clefts in multiple populations. J Dent Res. 91(5):473–478. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Liang J, Zhu L, Meng L, Chen D, Bian Z. 2012. Novel nonsense mutation in MSX1 causes tooth agenesis with cleft lip in a Chinese family. Eur J Oral Sci. 120(4):278–282. [DOI] [PubMed] [Google Scholar]
  21. Ludwig KU, Mangold E, Herms S, Nowak S, Reutter H, Paul A, Becker J, Herberz R, Alchawa T, Nasser E, et al. 2012. Genome-wide meta-analyses of nonsyndromic cleft lip with or without cleft palate identify six new risk loci. Nat Genet. 44(9):968–971. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Mangold E, Ludwig KU, Birnbaum S, Baluardo C, Ferrian M, Herms S, Reutter H, de Assis NA, Chawa TA, Mattheisen M, et al. 2010. Genomewide association study identifies two susceptibility loci for nonsyndromic cleft lip with or without cleft palate. Nat Genet. 42(1):24–26. [DOI] [PubMed] [Google Scholar]
  23. Martinelli M, Girardi A, Cura F, Nouri N, Pinto V, Carinci F, Morselli PG, Salehi M, Scapoli L. 2015. Non-syndromic cleft lip with or without cleft palate in Asian populations: association analysis on three gene polymorphisms of the folate pathway. Arch Oral Biol. 61:79-82. [DOI] [PubMed] [Google Scholar]
  24. Moreno LM, Mansilla MA, Bullard SA, Cooper ME, Busch TD, Machida J, Johnson MK, Brauer D, Krahn K, Daack-Hirsch S, et al. 2009. FOXE1 association with both isolated cleft lip with or without cleft palate, and isolated cleft palate. Hum Mol Genet. 18(24):4879–4896. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Mossey PA, Modell B. 2012. Epidemiology of oral clefts 2012: an international perspective. Front Oral Biol. 16:1–18. [DOI] [PubMed] [Google Scholar]
  26. Mostowska A, Hozyasz KK, Wójcicki P, Lasota A, Dunin-Wilczyńska I, Jagodziński PP. 2012. Association of DVL2 and AXIN2 gene polymorphisms with cleft lip with or without cleft palate in a Polish population. Birth Defects Res A Clin Mol Teratol. 94(11):943–950. [DOI] [PubMed] [Google Scholar]
  27. Murray T, Taub MA, Ruczinski I, Scott AF, Hetmanski JB, Schwender H, Patel P, Zhang TX, Munger RG, Wilcox AJ, et al. 2012. Examining markers in 8q24 to explain differences in evidence for association with cleft lip with/without cleftpalate between Asians and Europeans. Genet Epidemiol. 36(4):392–399. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Pan X, Wang P, Yin X, Liu X, Li D, Li X, Wang Y, Li H, Yu Z. 2015. Association between maternal MTHFR polymorphisms and nonsyndromic cleft lip with or without cleft palate in offspring: a meta-analysis based on 15 case-control studies. Int J Fertil Steril. 8(4):463–480. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Rafighdoost H, Hashemi M, Narouei A, Eskanadri-Nasab E, Dashti-Khadivaki G, Taheri M. 2013. Association between CDH1 and MSX1 gene polymorphisms and the risk of nonsyndromic cleft lip and/or cleft palate in a southeast Iranian population. Cleft Palate Craniofac J. 50(5):e98–e104. [DOI] [PubMed] [Google Scholar]
  30. Rahimov F, Marazita ML, Visel A, Cooper ME, Hitchler MJ, Rubini M, Domann FE, Govil M, Christensen K, Bille C, et al. 2008. Disruption of an AP-2α binding site in an IRF6 enhancer is associated with cleft lip. Nat Genet. 40(11):1341–1347. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Sozen MA, Tolarova MM, Spritz RA. 2009. The common MTHFR C677T and A1298C variants are not associated with the risk of non-syndromic cleft lip/palate in northern Venezuela. J Genet Genomics. 36(5):283–288. [DOI] [PubMed] [Google Scholar]
  32. Sun Y, Huang Y, Yin A, Pan Y, Wang Y, Wang C, Du Y, Wang M, Lan F, Hu Z, et al. 2015. Genome-wide association study identifies a new susceptibility locus for cleft lip with or without a cleft palate. Nat Commun. 6:6414. [DOI] [PubMed] [Google Scholar]
  33. Suzuki S, Marazita ML, Cooper ME, Miwa N, Hing A, Jugessur A, Natsume N, Shimozato K, Ohbayashi N, Suzuki Y, et al. 2009. Mutations in BMP4 are associated with subepithelial, microform, and overt cleft lip. Am J Hum Genet. 84(3):406–411. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Uslu VV, Petretich M, Ruf S, Langenfeld K, Fonseca NA, Marioni JC, Spitz F. 2014. Long-range enhancers regulating Myc expression are required for normal facial morphogenesis. Nat Genet. 46(7):753–758. [DOI] [PubMed] [Google Scholar]
  35. Weatherley-White RC, Ben S, Jin Y, Riccardi S, Arnold TD, Spritz RA. 2011. Analysis of genome-wide association signals for nonsyndromic cleft lip/palate in a Kenya African Cohort. Am J Med Genet A. 155A(10): 2422–2425. [DOI] [PubMed] [Google Scholar]
  36. Wolf ZT, Brand HA, Shaffer JR, Leslie EJ, Arzi B, Willet CE, Cox TC, McHenry T, Narayan N, Feingold E, et al. 2015. Genome-wide association studies in dogs and humans identify ADAMTS20 as a risk variant for cleft lip and palate. PLoS Genet. 11(3):e1005059. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Zhao M, Ren Y, Shen L, Zhang Y, Zhou B. 2014. Association between MTHFR C677T and A1298C polymorphisms and NSCL/P risk in Asians: a meta-analysis. PLoS One. 9(3):e88242. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Zucchero TM, Cooper ME, Maher BS, Daack-Hirsch S, Nepomuceno B, Ribeiro L, Caprau D, Christensen K, Suzuki Y, Machida J, et al. 2004. Interferon regulatory factor 6 (IRF6) gene variants and the risk of isolated cleft lip or palate. N Engl J Med. 351(8):769–780. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary material
DS_10.1177_0022034516657003_AppendixMethods.pdf (168.3KB, pdf)
Supplementary material
DS_10.1177_0022034516657003_AppendixTable1.pdf (182KB, pdf)
Supplementary material
DS_10.1177_0022034516657003_AppendixTable2.pdf (181.1KB, pdf)
Supplementary material
DS_10.1177_0022034516657003_AppendixTable3.pdf (180.2KB, pdf)
Supplementary material
DS_10.1177_0022034516657003_AppendixTable4.pdf (180.5KB, pdf)
Supplementary material
DS_10.1177_0022034516657003_AppendixTable5.pdf (239.1KB, pdf)

Articles from Journal of Dental Research are provided here courtesy of International and American Associations for Dental Research

RESOURCES