Received: 18 December 2017 | Revised: 8 February 2018 | Accepted: 20 February 2018 DOI: 10.1002/ajmg.a.38672 OR I G I N A L A R T I C L E Williams–Beuren syndrome in diverse populations Paul Kruszka1 | Antonio R. Porras2 | Deise Helena de Souza3 | Angelica Moresco4 | Victoria Huckstadt4 | Ashleigh D. Gill1 | Alec P. Boyle2 | Tommy Hu1 | Yonit A. Addissie1 | Gary T. K. Mok5 | Cedrik Tekendo-Ngongang6 | Karen Fieggen6 | Eloise J. Prijoles7 | Pranoot Tanpaiboon8 | Engela Honey9 | Luk Ho-Ming10 | Ivan F. M. Lo10 | Meow-Keong Thong11 | Premala Muthukumarasamy11 | Kelly L. Jones12 | Khadija Belhassan1,13 | Karim Ouldim13 | Ihssane El Bouchikhi13,14 | Laila Bouguenouch13 | Anju Shukla15 | Katta M. Girisha15 | Nirmala D. Sirisena16 | Vajira H. W. Dissanayake16 | C. Sampath Paththinige16 | Rupesh Mishra16 | Monisha S. Kisling8 | Carlos R. Ferreira8 | María Beatriz de Herreros17 | Ni-Chung Lee18 | Saumya S. Jamuar19 | Angeline Lai19 | Tan Ee Shien19 | Jiin Ying Lim19 | Cham Breana Wen-Min19 | Neerja Gupta20 | Stephanie Lotz-Esquivel21 | Ramses Badilla-Porras22 | Dalia Farouk Hussen23 | Mona O. El Ruby24 | Engy A. Ashaat24 | Siddaramappa J. Patil25 | Leah Dowsett26 | Alison Eaton27 | A. Micheil Innes27 | Vorasuk Shotelersuk28 | E€ben Badoe29 | Ambroise Wonkam6 | María Gabriela Obregon4 | Brian H. Y. Chung5 | Milana Trubnykova30 | Jorge La Serna30 | Bertha Elena Gallardo Jugo30 | Miguel Chavez Pastor30 | Hugo Hernan Abarca Barriga30 | Andre Megarbane31 | Beth A. Kozel32 | Mieke M. van Haelst33 | Roger E. Stevenson7 | Marshall Summar8 | Adeyemo A. Adebowale34 | Colleen A. Morris35 | Danilo Moretti-Ferreira3 | Marius George Linguraru2 | Maximilian Muenke1 1Medical Genetics Branch, National Human Genome Research Institute, The National Institutes of Health, Bethesda, Maryland 2Sheikh Zayed Institute for Pediatric Surgical Innovation, Children’s National Health System, Washington, District of Columbia 3Department of Genetics, Institute of Biosciences, Sao Paulo State University – UNESP, S~ao Paulo, Brazil 4Servicio de Genetica, Hospital de Pediatría Garrahan, Buenos Aires, Argentina 5Department of Paediatrics and Adolescent Medicine, LKS Faculty of Medicine, The University of Hong Kong, Hong Kong Special Administrative Region, Hongkong, China 6Division of Human Genetics, University of Cape Town, Cape Town, South Africa 7Greenwood Genetic Center, Greenwood, South Carolina 8Rare Disease Institute, Children’s National Medical Center, Washington, District of Columbia 1128 | VC 2018Wiley Periodicals, Inc. wileyonlinelibrary.com/journal/ajmga Am J Med Genet. 2018;176A:1128–1136. KRUSZKA ET AL. | 1129 9Department of Genetics, University of Pretoria, Pretoria, South Africa 10Clinical Genetic Service, Department of Health, Hong Kong Special Administrative Region, Hongkong, China 11Department of Paediatrics, Faculty of Medicine, University of Malaya, Kuala Lumpur, Malaysia 12Division of Medical Genetics and Metabolism, Children’s Hospital of The King’s Daughters, Norfolk, Virginia 13Medical Genetics and Oncogenetics Unit, Hassan II University Hospital, Fez, Morocco 14Laboratory of Microbial Biotechnology, Faculty of Sciences and Techniques, University of Sidi Mohammed Ben Abdellah, Fez, Morocco 15Department of Medical Genetics, Kasturba Medical College, Manipal University, Manipal, India 16Human Genetics Unit, Faculty of Medicine, University of Colombo, Colombo, Sri Lanka 17National Secretariat for the Rights of People with Disabilities (SENADIS), Fernando de la Mora, Paraguay 18Department of Pediatrics and Medical Genetics, National Taiwan University Hospital, Taipei, Taiwan 19Genetics Service, Department of Paediatrics, KK Women’s and Children’s Hospital, Singapore, Singapore 20Division of Genetics, Department of Pediatrics, All India Institute of Medical Sciences, New Delhi, India 21Research Department, Hospital San Juan de Dios (CCSS), San Jose, Costa Rica 22Medical Genetics and Metabolism Department, Hospital Nacional de Nin~os (CCSS), San Jose, Costa Rica 23Department of Human Cytogenetics, The National Research Centre, Cairo, Egypt 24Clinical Genetics Department, National Research Centre, Cairo, Egypt 25Mazumdar Shaw Medical Center, Narayana Health City, Bangalore, India 26Kapi’olani Medical Center for Women and Children, Honolulu, Hawaii 27Department of Medical Genetics and Alberta Children’s Hospital Research Institute, Cumming School of Medicine, University of Calgary, Calgary, Alberta 28Center of Excellence for Medical Genetics, Department of Pediatrics, Faculty of Medicine, Chulalongkorn University, Bangkok, Thailand 29School of Medicine and Dentistry, College of Health Sciences, University of Ghana, Accra, Ghana 30Instituto Nacional de Salud del Nin~o, Lima, Peru 31Institut Jerôme Lejeune, Paris, France 32National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, Maryland, Department of Pediatrics, Washington University School of Medicine, St. Louis, Missouri 33Department of Genetics, University Medical Centre, Utrecht, Utrecht, The Netherlands 34Center for Research on Genomics and Global Health, National Human Genome Research Institute, The National Institutes of Health, Bethesda, Maryland 35Department of Pediatrics (Genetics Division), University of Nevada School of Medicine, Las Vegas, Nevada Correspondence Williams–Beuren syndrome (WBS) is a common microdeletion syndrome characterized by a 1.5Mb Paul Kruszka and Maximilian Muenke, deletion in 7q11.23. The phenotype of WBS has been well described in populations of European Medical Genetics Branch, National Human Genome Research Institute, The National descent with not as much attention given to other ethnicities. In this study, individuals with WBS Institutes of Health, Bethesda, MD 20892. from diverse populations were assessed clinically and by facial analysis technology. Clinical data Emails: Paul.kruszka@nih.gov; and images from 137 individuals with WBS were found in 19 countries with an average age of 11 mamuenke@mail.nih.gov years and female gender of 45%. The most common clinical phenotype elements were periorbital Funding information fullness and intellectual disability which were present in greater than 90% of our cohort. Addition- Division of Intramural Research at the ally, 75% or greater of all individuals with WBS had malar flattening, long philtrum, wide mouth, National Human Genome Research and small jaw. Using facial analysis technology, we compared 286 Asian, African, Caucasian, and Institute; Government of Abu Dhabi Latin American individuals with WBS with 286 gender and age matched controls and found that the accuracy to discriminate between WBS and controls was 0.90 when the entire cohort was evaluated concurrently. The test accuracy of the facial recognition technology increased signifi- cantly when the cohort was analyzed by specific ethnic population (P-value<0.001 for all comparisons), with accuracies for Caucasian, African, Asian, and Latin American groups of 0.92, 0.96, 0.92, and 0.93, respectively. In summary, we present consistent clinical findings from global populations with WBS and demonstrate how facial analysis technology can support clinicians in making accurate WBS diagnoses. K E YWORD S Africa, Asia, diverse populations, facial analysis technology, Latin America, Middle East, syndrome, Williams, Williams–Beuren 1130 | KRUSZKA ET AL. 1 | INTRODUCTION explained by a concentration of clinical geneticists in developed countries (Limwongse, 2017) and the absence of genetics services in areas such as Williams-Beuren syndrome (WBS) was first characterized as a syn- sub-Saharan Africa (Tekendo-Ngongang et al., 2014). The American drome with dysmorphic facial features, supravalvar aortic stenosis, and Academy of Pediatrics has outlined clinical diagnostic criteria (Committee cognitive impairment in the early 1960’s (Beuren, Apitz, & Harmjanz, on Genetics, 2001), which places emphasis on both facial features and 1962; Williams, Barratt-Boyes, & Lowe, 1961). WBS is one of the com- echocardiography; however, these criteria may be difficult to apply to mon microdeletion syndromes occurring in roughly 1:7500 (Stromme, diverse populations such as sub-Saharan patients given the variation Bjornstad, & Ramstad, 2002) and caused by a 1.5 Mb deletion in in facial features and difficulty obtaining echocardiograms (Tekendo- 7q11.23 which includes 26–28 genes. Individuals with WBS present Ngongang et al., 2014). A few small studies have been conducted in with intellectual disability, hypersocial behavior, distinctive facies, cardi- diverse populations. Tekendo-Ngongang et al. presented three indi- ovascular disease (supravalvar aortic stenosis and peripheral pulmonary viduals with WBS from Cameroon in sub-Saharan Africa and noted stenosis), short stature, connective tissue anomalies, and endocrine that the facial features were not different from many unaffected abnormalities such as hypercalcemia (Morris, 1993, 2010; Sindhar sub-Saharan African individuals (Tekendo-Ngongang et al., 2014). et al., 2016). Facial characteristics include broad forehead, bitemporal Additionally, Lumaka et al. reported one case of WBS in a resource narrowing, periorbital fullness, a stellate iris appearance, short nose, limited area of central Africa and these authors remind us that most malar flattening, long philtrum, thick upper and lower lip vermillion, cases in sub-Saharan Africa are undiagnosed based on insufficient wide mouth, and large ear lobes (Morris, 1993, 2010). training in the field of dysmorphology and scarcity of genetic resour- The diagnosis of WBS is made based on dysmorphic features and ces (Lumaka et al., 2016). intellectual and behavioral findings. Diagnosis is confirmed with molecular Although we know of at least one comparison of different ethnic- testing. Most studies have focused on Caucasians, which can be ities and WBS, where Zitzer-Comfort et al. compared global sociability TABLE 1 Summary of clinical exam findings of individuals with Williams–Beuren syndrome from diverse backgrounds Present study Perez Jurado et al. (1996) Patil et al. (2012) African American, Asian, Caucasian, Latin American Asian African Latin American Indian n5105 n524 n58 p values n565 n5 27 Average age (years) 11.9 8.1 7.7 5.5 Male gender 55% 50% 75% 74% Molecular diagnosis 100% 96% 75% 94% (56/59) 100% Cardiovascular disease 73% 71% 88% .64 50% (24/48)a 63%a Wide mouth 91% 78% (18/23) 88% <.001 100% Short nose 74% 75% 88% .71 90% (37/41) 100% Periorbital fullness 95% 92% 100% .62 96% (42/44) 100% Malar flattening 99% 75% 100% .001 100% (43/43) 85% Small jaw 82% 75% 75% .69 na 85% Long philtrum 93% 79% 88% .10 83% (35/42) 85% Epicanthic folds 73% 63% 13% .001 71% (27/38) 52% Malocclusion 59% (55/94) 47% (8/17) 38% .39 81% (25/31) 44% Widely spaced teeth 47% (35/74) 93% (15/16) 71% (5/7) .002 41% Broad eyebrow 63% 58% 63% .92 67% (22/33)b 37% Stellate iris 85% (82/97) 12% (2/16) 14% (1/7) <.001 15% Strabismus 57% (59/104) 6% (1/17) 25% <.001 11% Intellectual disability 100% (103/103) 95% (18/19) 100% (7/7) .05 91% (42/46)c Growth abnormalities 91% (93/102) 53% (9/17) 25% <.001 18% (8/44)d aSupraventricular aortic stenosis. bDescribed as medial eyebrow flare in Perez Jurado et al. (1996). cIQ 75. dWeight<3rd centile. KRUSZKA ET AL. | 1131 between Japanese and United States individuals with WBS TABLE 2 Population data used in facial analysis technology, which (Zitzer-Comfort, Doyle, Masataka, Korenberg, & Bellugi, 2007), we are includes 286 individuals with Williams–Beuren syndrome unaware of a dysmorphology and diagnostic comparison. In line with Williams–Beuren Controls other publications on genetic syndromes in diverse populations, we Number % Number % explore the phenotype of WBS in different ancestral populations using Age both clinical exam and facial analysis technology (Kruszka, Addissie, <30 days 0 0 0 0 et al., 2017; Kruszka, Porras, et al., 2017; Kruszka, Porras, Sobering, 1–24 months 49 17 49 17 et al., 2017; Muenke, Adeyemo, & Kruszka, 2016). 25–60 months 47 16 47 16 5–12 years 71 25 71 25 13–18 years 28 10 28 10 2 | METHODS >18 years 91 32 91 32 Total 286 286 2.1 | Review of medical literature Ethnicity African Descent 28 10 28 10 A Medline search was conducted with the following terms: WBS, Asian 26 9 26 9 Africa, Asia, Latin America, Middle East, diverse populations, and facial Caucasian 121 42 121 42 Latino 111 39 111 39 analysis technology. Reference lists of journal studies were used to find Total 286 286 further relevant journal articles. After obtaining journal permissions, Gender photos of individuals with WBS were used to supplement study partici- Male 150 52 150 52 pants described below (Delgado et al., 2013; Honjo et al., 2015; Jiang Female 136 48 136 48 & Liu, 2015; Lumaka et al., 2016; Mazumdar, Sarkar, Badveli, & Total 286 286 Majumder, 2016; Morris, 1993, 2010; Patil, Madhusudhan, Shah, & Suresh, 2012; Sakhuja, Whyte, Kamath, Martin, & Chitayat, 2015; American groups for the purpose of facial analysis. In Table 2, we show Smoot, Zhang, Klaiman, Schultz, & Pober, 2005; Tekendo-Ngongang ages, gender, and ethnicity of the facial analysis technology cohort. et al., 2014; van Kogelenberg et al., 2010; Wu et al., 2002). With feature extraction, feature selection, and classification as out- put variables, our algorithms analyzed study participants’ images. From a 2.2 | Patients set of 44 landmarks placed on the frontal face images, a total of 126 facial features, including both geometric and texture biomarkers, were Individuals with WBS were evaluated from 19 countries. All partici- isolated. Figure 1 shows the landmark locations and the geometric fea- pants (Supporting Information Table 1) had WBS diagnosed by both tures extracted. The geometric biomarkers are distances and angles cal- clinical evaluation and/or molecular diagnosis. In a few cases molecular culated between the different inner facial landmarks. Texture patterns diagnosis was not done secondary to resource limitations. Geographic (Cerrolaza et al., 2016) were calculated at each of the 33 inner facial area of origin or ethnicity (African and African American, Asian, Latin landmarks to quantify texture information (Figure 1). Using the method American, and Middle Eastern) was used to categorize patients. Local proposed previously (Cai, Zhang, & He, 2010), from the collection of clinical geneticists examined patients for established clinical features geometric and texture features, the most significant ones were selected. found in WBS (Committee on Genetics, 2001). For each feature set, a support vector machine classifier (Cortes & Consent was obtained by local institutional review boards and the Vapnik, 1995) was trained using a leave-one-out cross-validation strat- Personalized Genomics protocol at the National Institutes of Health egy (Elisseeff & Pontil, 2003). The optimal number of features was (11-HG-0093). Exam findings from the current study and those from selected as the minimum number for which the classification accuracy the medical literature (Patil et al., 2012; Perez Jurado, Peoples, Kaplan, converged to its maximum; Supporting Information Figures 1–5 graphi- Hamel, & Francke, 1996) are recorded in Table 1. cally demonstrate how the addition of features improves the measures of sensitivity, specificity, and accuracy. The p value of each feature was 2.3 | Facial analysis technology also estimated using the Student’s t-test as an estimator of the individual As described in our previous studies (Kruszka, Addissie, et al., 2017; discriminant power of each feature selected. We evaluated the improve- Kruszka, Porras, et al., 2017; Kruszka, Porras, Sobering, et al., 2017), ments of using classification models trained specifically for each ethnicity digital facial analysis technology (Cerrolaza et al., 2016; Zhao et al., to detect WBS compared to using one single classification model trained 2013; Zhao, Okada, et al., 2014; Zhao, Werghi, et al., 2014) evaluated using all the cases available from all ethnicities. The statistical signifi- 286 frontal images of individuals with WBS, and 286 healthy controls cance of their differences was assessed using Fisher’s exact test. (matched for ethnicity, gender, and age) from our previously described database (Zhao et al., 2013; Zhao, Okada, et al., 2014; Zhao, Werghi, 3 | RESULTS et al., 2014). The 286 individuals with WBS used for facial analysis technology included individuals from Supporting Information Table 1 Clinical information (Table 1) was collected on 137 individuals and and additional archival images of individuals with WBS. A Caucasian images (Figures 2–5; Supporting Information Table 1) on 128 individu- ethnic group was identified in addition to African, Asian, and Latin als (17 individuals were obtained from the medical literature). The 1132 | KRUSZKA ET AL. F IGURE 1 Facial landmarks on three patients with WBS. Inner facial landmarks are represented in red, while external landmarks are represented in blue. Blue lines indicate the calculated distances. Green circles represent the corners of the calculated angles. Texture features are extracted only from the inner facial landmarks. [Color figure can be viewed at wileyonlinelibrary.com] participants were from 19 countries, average age was 11.0 years (range wide mouth, malar flattening, epicanthal folds, widely spaced teeth, stel- newborn to 42 years), and 45% were females (Table 1). Individuals of late iris, strabismus, and growth abnormalities (p< .05; v2 test). African descent are shown in Figure 2, Asian in Figure 3, Latin Ameri- As a more objective measure of phenotype, facial analysis technol- can in Figure 4, and Middle Eastern patients in Figure 5. Table 1 does ogy was applied to 286 individuals (Caucasian, African, Asian, and Latin not show individuals from Middle East due to insufficient clinical American) with results shown in Table 3. The accuracy to discriminate information. between WBS and controls was 0.90 when the entire cohort was eval- From the medical literature in Table 1, we show facial and other uated concurrently. The test accuracy of the facial recognition technol- phenotype elements from two studies that each evaluated over 25 par- ogy increased significantly when the cohort was analyzed by specific ticipants from diverse backgrounds (Patil et al., 2012; Perez Jurado ethnic population (p value< .001 for all comparisons), with accuracies et al., 1996). We compared unpublished patients from the present for Caucasian, African, Asian, and Latin American groups of 0.92, 0.96, study with the above-mentioned studies from the medical literature 0.92, and 0.93, respectively (Table 3). Supporting Information Tables (Table 1). The most common phenotype element in both the present 2–6 show the geometric and texture feature comparisons between study and the medical literature was periorbital fullness and intellectual individuals with WBS and unaffected individuals. Interestingly, the disability which was present in greater than 90% of our cohort (Table angle at the nose root is the most significant geographic discriminator 1). In all studies in Table 1, 75% or greater of all individuals with WBS between WBS and controls across all ethnicities. had malar flattening, long philtrum, wide mouth, and small jaw (wide mouth and small jaw not reported in Perez Jurado et al., 1996). 4 | DISCUSSION As seen in Table 1, the majority of clinical exam findings in the present study were consistent between the different population groups WBS is a common microdeletion syndrome that has recognizable facial with the following exam elements differing statistically amongst groups: characteristics, intellectual disability, a characteristic friendly personality, FIGURE 2 Frontal and lateral facial profiles of individuals of African descent with WBS. Gender, age, and country of origin are presented in Supporting Information Table 1. [Color figure can be viewed at wileyonlinelibrary.com] KRUSZKA ET AL. | 1133 FIGURE 3 Frontal and lateral facial profiles of Asian individuals with WBS. Gender, age, and country of origin are presented in Supporting Information Table 1. [Color figure can be viewed at wileyonlinelibrary.com] and often cardiovascular disease. Given the well characterized pheno- Table 1 lists the clinical phenotype of 137 individuals from Latin Ameri- type of WBS, there is still a paucity of cases of WBS from developing can, Asian, and African ancestry and Figures 2–5 show 128 facial images countries in the medical literature (Lumaka et al., 2016; Tekendo- of individuals from diverse populations. Although there are some statisti- Ngongang et al., 2014). The first goal of this study was to assemble and cally significant differences in phenotype elements across population characterize a cohort of individuals with WBS from diverse populations. groups, there are multiple well-known characteristics that are present in FIGURE 4 Frontal and lateral facial profiles of Latin Americans with WBS. Gender, age, and country of origin are presented in Supporting Information Table 1. [Color figure can be viewed at wileyonlinelibrary.com] 1134 | KRUSZKA ET AL. F IGURE 5 Frontal and lateral facial profiles of individuals from the Middle East with WBS. Gender, age, and country of origin are presented in Supporting Information Table 1. [Color figure can be viewed at wileyonlinelibrary.com] 75% or more of all groups, including periorbital fullness, wide mouth, by clinicians; however, the angle at the nose root increases for shorter malar flattening, small jaw, long philtrum, and intellectual disability (Table noses, which is a well-known feature in patients with Williams syn- 1). In addition to this study, we have also made a publicly available data- drome as seen in Table 1. Interestingly, the only population group for base that shows images of individuals with WBS and syndromes in which the width of the mouth was not depicted as a top feature of diverse populations (http://www.genome.gov/atlas) (Koretzky et al., WBS by our technology was the African group. 2016; Muenke et al., 2016). The study has several limitations. We acknowledge that ascertain- The second goal of this study was to test whether a diagnosis was ment bias exists with only the most severe phenotypes or those with more difficult in different ethnicities as has been suggested (Patil et al., severe congenital heart disease seeking medical attention. Thus, the 2012; Tekendo-Ngongang et al., 2014). To answer this question, we milder cases of WBS are most likely missed. Due to relatively small used the objectivity of facial analysis technology. The facial analysis tech- sample sizes, this study grouped populations by large geographical nology accurately discriminated between individuals with WBS and con- areas. For example, individuals from India, Thailand, and China are trols with accuracy above 92% in all population groups (Table 3). The test grouped into the category “Asia.” In the future, we plan to narrow this accuracy of the facial recognition technology increased significantly when geographic constraint. Another limitation is that much of the clinical the cohort was analyzed by specific ethnic population (p value< .001 for data is subjective and based on provider judgement. We have all comparisons; Fisher’s test), in other words, when the computer was attempted to address this issue with the use of objective measure- trained on an ethnic specific data set, the accuracy improved. ments using digital face analysis technology. Some of the characteristic features of WBS in the global popula- We conclude by acknowledging that WBS can be a difficult diag- tion determined by facial analysis technology are: wide mouth, short nosis to make (average age of diagnosis of WBS is 3.7–5.3 years in nose, and texture of eyelids/epicanthic folds, which were also noted in developed countries) (Ferrero et al., 2007; Huang, Sadler, O’Riordan, & the clinical evaluation of most of the cases. We would like to make spe- Robin, 2002). This study and similar reports (Kruszka, Addissie, et al., cial mention of the angle of the nose root. As noted in the results, the 2017; Kruszka, Porras, et al., 2017; Kruszka, Porras, Sobering, et al., angle at the nose root is the most significant geographic discriminator 2017) and our recently created website, http://www.genome.gov/atlas between WBS and controls across all ethnicities (Supporting Informa- are designed to have widespread clinical significance for the diagnosis tion Tables 2–6). The angle at the nose root is not typically measured of individuals with WBS, especially in countries without access to genetic services or genetic testing where the simplicity of facial analysis TABLE 3 Measures of diagnostic accuracy for facial analysis tech- technology may be a useful asset. nology that discriminate between Williams-Beuren syndrome and unaffected individuals, stratified by populations ACKNOWLEDGMENTS Number We are grateful to the individuals and their families who participated of features AUC Accuracy Sensitivity Specificity in our study. P.K., Y.A.A, A.D.G., T.H., A.A.A., and M.M. are supported Global 17 0.95 0.90 0.92 0.88 by the Division of Intramural Research at the National Human Caucasian 15 0.97 0.92 0.89 0.95 Genome Research Institute, NIH. Partial funding of this project was African and 9 0.96 0.96 0.96 0.96 from a philanthropic gift from the Government of Abu Dhabi to the African Children’s National Health System. V.S. is supported by the Chulalong- American korn Academic Advancement into its 2nd century project. Asian 8 0.95 0.92 0.96 0.88 Latin American 15 0.97 0.93 0.95 0.92 ORCID AUC5 area under the receiver operating characteristic curve Paul Kruszka http://orcid.org/0000-0003-4949-0875 KRUSZKA ET AL. | 1135 Luk Ho-Ming http://orcid.org/0000-0003-4066-4066 Kruszka, P., Porras, A. R., Addissie, Y. A., Moresco, A., Medrano, S., Mok, Anju Shukla http://orcid.org/0000-0002-8347-429X G. T. K., . . . Muenke, M. (2017). Noonan syndrome in diverse populations. American Journal of Medical Genetics Part A, 173(9), Katta M. Girisha http://orcid.org/0000-0002-0139-8239 2323–2334. https://doi.org/10.1002/ajmg.a.38362 Carlos R. Ferreira http://orcid.org/0000-0002-2697-1046 Kruszka, P., Porras, A. R., Sobering, A. K., Ikolo, F. A., La Qua, S., Siddaramappa J. Patil http://orcid.org/0000-0001-7154-3449 Shotelersuk, V., . . . Muenke, M. (2017). Down syndrome in diverse Hugo Hernan Abarca Barriga http://orcid.org/0000-0002-0276- populations. American Journal of Medical Genetics Part A, 173(1), 2557 42–53. https://doi.org/10.1002/ajmg.a.38043 Roger E. Stevenson http://orcid.org/0000-0002-1806-6345 Limwongse, C. (2017). Medical genetic services in a developing country: Lesson from Thailand. Current Opinion in Pediatrics, 29(1), 634–639. https://doi.org/10.1097/MOP.0000000000000544 REFERENCES Lumaka, A., Lukoo, R., Mubungu, G., Lumbala, P., Mbayabo, G., Mupuala, Beuren, A. J., Apitz, J., & Harmjanz, D. (1962). Supravalvular aortic steno- A., . . . Devriendt, K. (2016). Williams-Beuren syndrome: Pitfalls for sis in association with mental retardation and a certain facial appear- diagnosis in limited resources setting. Clinical Case Reports, 4(3), 294– ance. Circulation, 26, 1235–1240. 297. https://doi.org/10.1002/ccr3.476 Cai, D., Zhang, C., & He, X. (2010). Unsupervised feature selection for Mazumdar, J., Sarkar, R., Badveli, A., & Majumder, B. (2016). Double multi-cluster data. Paper presented at the Proceedings of the 16th chamber right ventricle in Williams syndrome: A rare cardiac anomaly ACM SIGKDD international conference on Knowledge discovery and reported. Springerplus, 5, 275. https://doi.org/10.1186/s40064-016- data mining. ACM, 333–342. 1897-y Cerrolaza, J. J., Porras, A. R., Mansoor, A., Zhao, Q., Summar, M., & Morris, C. A. (1993). Williams syndrome. In M. P. Adam, H. H. Ardinger, Linguraru, M. G. (2016). Identification of dysmorphic syndromes using R. A. Pagon, S. E. Wallace, L. J. H. Bean, H. C. Mefford, K., . . . N. landmark-specific local texture descriptors. Paper presented at the Bio- Ledbetter (Eds.), GeneReviews®. Seattle, WA: University of Washing- medical Imaging (ISBI), 2016 IEEE 13th International Symposium on ton. Available from http://www-ncbi-nlm-nih-gov.ezproxy.nihlibrary Biomedical Imaging. IEEE, 1080–1083. .nih.gov/books/NBK1249/ Committee on Genetics. (2001). American Academy of Pediatrics: Health Morris, C. A. (2010). Introduction: Williams syndrome. American Journal care supervision for children with Williams syndrome. Pediatrics, 107 of Medical Genetics Part C: Seminars in Medical Genetics, 154C(2), (5), 1192–1204. 203–208. https://doi.org/10.1002/ajmg.c.30266 Cortes, C., & Vapnik, V. (1995). Support-vector networks. Machine Learn- Muenke, M., Adeyemo, A., & Kruszka, P. (2016). An electronic atlas of ing, 20(3), 273–297. human malformation syndromes in diverse populations. Genetics in Delgado, L. M., Gutierrez, M., Augello, B., Fusco, C., Micale, L., Merla, G., Medicine, 18(11), 1085–1087. https://doi.org/10.1038/gim.2016.3 & Pastene, E. A. (2013). A 1.3-mb 7q11.23 atypical deletion identified Patil, S. J., Madhusudhan, B. G., Shah, S., & Suresh, P. V. (2012). Facial in a cohort of patients with Williams-Beuren syndrome. Molecular Syn- phenotype at different ages and cardiovascular malformations in chil- dromology, 4(3), 143–147. https://doi.org/10.1159/000347167 dren with Williams-Beuren syndrome: A study from India. American Elisseeff, A., & Pontil, M. (2003). Leave-one-out error and stability of Journal of Medical Genetics Part A, 158A(7), 1729–734. https:// learning algorithms with applications. NATO Science Series Sub Series, doi.org/10.1002/ajmg.a.35443 III: Computer and Systems Sciences, 190, 111–130. Perez Jurado, L. A., Peoples, R., Kaplan, P., Hamel, B. C., & Francke, U. Ferrero, G. B., Biamino, E., Sorasio, L., Banaudi, E., Peruzzi, L., Forzano, (1996). Molecular definition of the chromosome 7 deletion in S., . . . Silengo, M. C. (2007). Presenting phenotype and clinical evalu- Williams syndrome and parent-of-origin effects on growth. American ation in a cohort of 22 Williams-Beuren syndrome patients. European Journal of Human Genetics, 59(4), 781–792. Journal of Medical Genetics, 50(5), 327–337. https://doi.org/10.1016/ Sakhuja, P., Whyte, H., Kamath, B., Martin, N., & Chitayat, D. (2015). j.ejmg.2007.05.005 Williams syndrome presenting with findings consistent with Alagille Honjo, R. S., Dutra, R. L., Furusawa, E. A., Zanardo, E. A., Costa, L. S., syndrome. Clinical Case Reports, 3(1), 24–28. https://doi.org/ Kulikowski, L. D., . . . Kim, C. A. (2015). Williams-Beuren syndrome: A 10.1002/ccr3.138 clinical study of 55 Brazilian patients and the diagnostic use of Sindhar, S., Lugo, M., Levin, M. D., Danback, J. R., Brink, B. D., Yu, E., . . . MLPA. BioMed Research International 2015, 903175. https://doi.org/ Kozel, B. A. (2016). Hypercalcemia in patients with Williams-Beuren 10.1155/2015/903175 syndrome. Journal of Pediatrics, 178, 254–260 e254. https://doi.org/ Huang, L., Sadler, L., O’Riordan, M. A., & Robin, N. H. (2002). Delay in 10.1016/j.jpeds.2016.08.027 diagnosis of Williams syndrome. Clinical Pediatrics (Philadelphia), 41(4), Smoot, L., Zhang, H., Klaiman, C., Schultz, R., & Pober, B. (2005). Medical 257–261. https://doi.org/10.1177/000992280204100410 overview and genetics of Williams-Beuren syndrome. Progress in Pedi- Jiang, M., & Liu, L. (2015). Williams-Beuren syndrome: A case confirmed atric Cardiology, 20(2), 195–205. https://doi.org/10.1016/j.ppedcard. by array-CGH method. Iranian Journal of Pediatrics, 25(1), e247. 2005.04.010 https://doi.org/10.5812/ijp.247 Stromme, P., Bjornstad, P. G., & Ramstad, K. (2002). Prevalence estima- Koretzky, M., Bonham, V. L., Berkman, B. E., Kruszka, P., Adeyemo, A., tion of Williams syndrome. Journal of Child Neurology, 17(4), 269– Muenke, M., & Hull, S. C. (2016). Towards a more representative 271. https://doi.org/10.1177/088307380201700406 morphology: Clinical and ethical considerations for including diverse Tekendo-Ngongang, C., Dahoun, S., Nguefack, S., Gimelli, S., Sloan-Bena, populations in diagnostic genetic atlases. Genetics in Medicine, 18(11), F., & Wonkam, A. (2014). Challenges in clinical diagnosis of Williams- 1069–1074. https://doi.org/10.1038/gim.2016.7 Beuren syndrome in sub-Saharan Africans: Case reports from Came- Kruszka, P., Addissie, Y. A., McGinn, D. E., Porras, A. R., Biggs, E., Share, roon. Molecular Syndromology, 5(6), 287–292. https://doi.org/ M., . . . Muenke, M. (2017). 22q11.2 deletion syndrome in diverse 10.1159/000369421 populations. American Journal of Medical Genetics Part A, 173(4), van Kogelenberg, M., Ghedia, S., McGillivray, G., Bruno, D., Leventer, R., 879–888. https://doi.org/10.1002/ajmg.a.38199 Macdermot, K., . . . Robertson, S. P. (2010). Periventricular 1136 | KRUSZKA ET AL. heterotopia in common microdeletion syndromes. Molecular Syndro- Biology Society, 2014, 754–757. https://doi.org/10.1109/EMBC. mology, 1(1), 35–41. https://doi.org/10.1159/000274491 2014.6943700 Williams, J. C., Barratt-Boyes, B. G., & Lowe, J. B. (1961). Supravalvular Zitzer-Comfort, C., Doyle, T., Masataka, N., Korenberg, J., & Bellugi, U. aortic stenosis. Circulation, 24, 1311–1318. (2007). Nature and nurture: Williams syndrome across cultures. Wu, Y. Q., Bejjani, B. A., Tsui, L. C., Mandel, A., Osborne, L. R., & Shaffer, Developmental Science, 10(6), 755–762. https://doi.org/10.1111/ L. G. (2002). Refinement of the genomic structure of STX1A and j.1467-7687.2007.00626.x mutation analysis in nondeletion Williams syndrome patients. Ameri- can Journal of Medical Genetics, 109(2), 121–124. https://doi.org/ 10.1002/ajmg.10321 SUPPORTING INFORMATION Zhao, Q., Okada, K., Rosenbaum, K., Kehoe, L., Zand, D. J., Sze, R., . . . Additional Supporting Information may be found online in the sup- Linguraru, M. G. (2014). Digital facial dysmorphology for genetic porting information tab for this article. screening: Hierarchical constrained local model using ICA. Medical Image Analysis, 18(5), 699–710. https://doi.org/10.1016/j.media. 2014.04.002 Zhao, Q., Okada, K., Rosenbaum, K., Zand, D. J., Sze, R., Summar, M., & How to cite this article: Kruszka P, Porras AR, de Souza DH, Linguraru, M. G. (2013). Hierarchical constrained local model using et al. Williams–Beuren syndrome in diverse populations. Am J ICA and its application to Down syndrome detection. Medical Image Med Genet Part A. 2018;176A:1128–1136. https://doi.org/10. Computing and Computer-Assisted Intervention, 16(Pt 2), 222–229. 1002/ajmg.a.38672 Zhao, Q., Werghi, N., Okada, K., Rosenbaum, K., Summar, M., & Lingur- aru, M. G. (2014). Ensemble learning for the detection of facial dys- morphology. Conference Proceedings IEEE Engineering in Medicine and