Background

Emirati people (citizens of United Arab Emirates, UAE) have diverse ethnicities that include lineages from the Arabian Peninsula, Persia, Baluchistan, and East Africa. The culture is primarily tribal and encourages intra-tribal (consanguineous) marriages [1]. Thus, “founder” mutations are prevalent, which markedly increase the frequency of autosomal recessive disorders [2]. Consistently, the prevalence of inborn errors of metabolism (IEM) in the region is high (about 1 in 1,329) [1]. Genetic disorders are also exceptionally common [3].

Diagnosing IEM is often challenging due to their genetic heterogeneity and atypical or overlapping phenotypes. Using clinical assessments and conventional investigations (having an estimated cost of about $25,000 per patient), the rate of molecular diagnosis in these patients is usually less than 50 % [4]. In these families, the lack of a diagnosis poses significant consequences, such as patients missing treatments and failure by medical staff to provide proper counseling and guidance.

The relatively new term “exome” refers to all exons in the human genome. The term “exons” refers to the ~180,000 genomic sequences, which transcribe and remain in the mature RNA. Exome constitutes only 3 % of the human genome and represents the ~22,000 protein-coding genes, mutations of which are responsible for about 85 % of our significant clinical diseases [5–7]. Therefore, sequencing of the exome is thought to be an efficient method of analyzing a patient's DNA to discover the genetic cause of disease. WES is especially efficient in detecting rare mutations in autosomal recessive diseases in consanguineous families.

The recently introduced WES technology offers an adjunctive analytical tool to locus-specific and gene-panel testing for patients with wide-spectrum phenotypes. This study describes diagnostic rate, advantages, and limitations of WES in the UAE.

Methods

This study was approved by Al-Ain Medical Human Research Ethics Committee (protocol number 10/09). Between January 2012 and December 2014, at Tawam Hospital (Al Ain, Abu Dhabi), 85 WES studies were requested. The risks and benefits of the test were explained to families and patients and written informed consent/ permission was obtained. WES was requested because of atypical phenotypes, overlapping features, or negative biochemical studies.

Blood (5 mL in EDTA-Vacutainer) was obtained from patients, parents and affected family members (when applicable). Extracted DNA was shipped to Whole Genome Laboratory (WGL) of Baylor College of Medicine (Houston, TX, USA). Single WES was performed at WGL as previously reported [6]. Briefly, genomic DNA from patients was fragmented by sonication. The fragments were ligated to illumina multiplexing paired-end adapters, amplified by polymerase chain-reaction assay, and hybridized to biotin-labeled VCRome (at 47 °C for 64–72 h). Paired-end sequencing was performed on IlluminaHiSeq 2000 platform. The mean coverage of the exome was 100 – 120X; and 95 % of the exome was covered at >20X coverage. The output data from IlluminaHiSeqwere converted from bcl file to FastQ file by Illumina Consensus Assessment of Sequence and Variation Software and mapped by Burrows-Wheeler Aligner (BWA) program. Variants were filtered using stringencies of minor allele frequencies and mutation databases and disease specific databases. The results (with disease phenotype) were interpreted by molecular laboratory and medical directors certified by American board of Medical Genetics. Variant annotation and filtering were performed using Atlas-SNP and Atlas-indel.16, in addition to CASSANDRA software. Variants with a minor allele frequency of less than 5 % (according to either the 1000 Genomes Project18 or the ESP5400 data of the National Heart, Lung, and Blood Institute GO Exome Sequencing Project [http://evs.gs.washington.edu/EVS]) were retained. Variants were interpreted according to American College of Medical Genetics guidelines and patient phenotypes. Variants related to patient phenotypes were confirmed by Sanger sequencing for the patients, parents and affected relative if applicable. The interpretation of genes related to the patient’s clinical phenotype was based on the clinical information of the patient. WES reports included also In silico-predictions for non-synonymous (missense) changes by SIFT and PolyPhen-2 [6]. Table 1 lists variants of unknown significance that confirmed the clinical disorder after further investigations. Tables 2, 3 list pathogenic mutations that directly confirmed the clinical disorder without a need for further investigation.

Table 1 Variants of unknown significance, which were confirmed pathogenic or likely pathogenic
Full size table
Table 2 Confirmed IEM by WES. n = 14
Full size table
Table 3 Confirmed genetic diseases by WES, n = 17
Full size table

Results

Eighty-five patients (46 % females) had WES testing. WES results were divided into the following categories: [1] Variants of unknown significance, which were confirmed pathogenic (12 patients, Group A) or likely pathogenic (12 patients, Group B) based on consistent phenotypes, biochemical findings, familial (segregation) studies, or reported pathogenicity (Table 1); [2] WES diagnosis of IEM (14 patients, Table 2); [3] WES diagnosis of genetic diseases (17 patients, Table 3); [4] Variants of unknown significance, which were inconsistent with patients’ phenotypes or biochemical/ familial (segregation) studies (24 patients); and [5] No pathogenic variants related to patient phenotype were identified (six patients).

Patients who had variants of unknown significance were investigated using biochemical tests, familial (segregation) studies, and a review of the literature. Patients with consistent phenotypes, biochemical findings, familial (segregation) studies, or reported pathogenicity were considered “confirmed pathogenic” (Table 1A). Patients with consistent phenotypes or previously reported probable pathogenicity were considered “likely pathogenic” (Table 1B). The 24 patients who had ‘variants of unknown significance’ that were inconsistent with their phenotypes or biochemical/ familial (segregation) studies were not included in this report.

A. Patients’ demographics

Eighty of the 85 (93 %) patients were younger than 18 years and 44 (52 %) were younger than 5 years. Sixty-eight (80 %) patients had neurologic abnormalities. Patients had appropriate metabolic and genetic investigations (e.g., metabolic screening, microarray analysis, skin/ muscle biopsy, biochemical studies, and DNA testing) before WES testing. For most of patients, both parents DNA were tested. The cost of WES was billed to insurance companies in the UAE.

B. Molecular diagnoses

Fourteen (16 %) patients had IEM (Table 2), 17 (20 %) had genetic syndromes (Table 3), 12 (14 %) had variants of unknown significance confirmed to be pathogenic (Table 1A), and 12 (14 %) had variants of unknown significance likely to be pathogenic (Table 1B). Therefore, WES information was clinically useful in 55 (65 %) patients and diagnostic in 43 (50 %) patients.

Forty-one (75 %) of the 55 patients had consanguineous families. All patients with autosomal recessive diseases had consanguineous families (Tables 1–3); except the patient with MTHFR mutation and the patient with TBX19 mutation (Table 2) who had families of the same tribe/ region; the patient with compound heterozygous mutations in MPDZ (Table 1) and the patients with PCSK1, AHI1, and DOCK7 mutations (Table 3). The patients with autosomal dominant mutation in CHRNB1 and FOXG1 (Table 1), the patient with mitochondrial mutation in tRNA Ala (Table 2), and the patient with X-linked mutation in PDHA1 (Table 2) had non-consanguineous families. In Table 3, the patients with X-linked or autosomal dominant conditions had non-consanguineous families, except for the patient with MYBPC3 mutations.

One hundred twenty-eight variants were identified. Forty-four (34 %) variants were pathologic. Twenty (15 %) of these variants were previously reported and 24 (19 %) were novel.

C. Autosomal dominant mutations

Eleven (13 %) patients had autosomal dominant disorders. Five patients had truncating or missense mutations which were previously reported. Three of these mutations were de novo (Table 3). Six patients had autosomal dominant (Table 1); three were de novo variant, one was inherited from symptomatic parent, and two were inherited from parents of unknown phenotype.

D. Autosomal recessive mutations

Forty (47 %) patients had autosomal recessive diseases. Thirty-five patients had homozygous variants and five patients had compound heterozygous variants. Eleven pathogenic mutations involved inborn errors of metabolism (Table 2). Five of these mutations were novel, including a compound heterozygous mutation (Table 2). Eleven pathogenic homozygous mutations involved genetic syndromes (Table 3). Five of these mutations were previously reported and six were novel (Table 3).

Of the reported variants of unknown significance, nine were confirmed pathogenic homozygous mutations on further testing, six of which were novel (Table 1A). Seven of these disorders were IEM (Table 1A). In addition, six homozygous and three compound heterozygous variants were likely pathogenic; only one of these variants was previously reported (Table 1B).

E. X-linked mutations

The three patients with X-linked disorders had de novo mutations; one had pyruvate dehydrogenase e1-alpha deficiency (Table 2) and one had mental retardation, X-linked 30 (Table 3). The patient with epileptic encephalopathy, early infantile had mosaic de novo variant (Table 1A).

F. Mitochondrial mutations

Two female patients had known pathogenic mutations. The patient with Leigh disease had 56 % heteroplasmic m.13513G > A (p. D393N, ND5) mutation in the mitochondrial ND5 gene. The second patient had mitochondrial myopathy with 59 % heteroplasmic m.5591G > A (tRNAAla), Table 2.

G. Carrier status

Carrier status for Mendelian recessive disorders was detected in nine patients. Two patients had sickle cell trait, two had β-thalassemia trait, one had CFTR gene mutation (cystic fibrosis heterozygosity), and four females had G6PD gene mutations (glucose-6-phosphatase dehydrogenase heterozygosity).

H. Medically actionable genetic variants

Medically actionable genetic variants unrelated to the clinical phenotype were identified in eight patients. Seven patients had mutations in the G6PD gene (c.563C > T in five patients, c.1003G > A in one patient, andc.634A > G in one patient) and one had mutation in the MYBPC3 gene (c.2148 + 1G > A) that causes dilated cardiomyopathy.

Discussion

In this study, WES provided useful clinical information in most patients in whom several other tests were not revealing. It attained a diagnostic rate of 50 % compared with25–42%in other studies [6–11]. This higher yield is a result of the frequent autosomal recessive diseases (mostly homozygous mutations) in the region and the multiple affected children in the same family.

Thirty-one of the 43 (72 %) patients with confirmed molecular diagnoses had autosomal recessive disorders compared with 34 % in another study [8]. Twenty-nine of these patients were homozygous for the rare allele, indicating consanguinity played a major role in making autosomal recessive disorders frequent in this culture.

Of the patients with autosomal dominant disorders, 66 % had de novo mutations (Table 1 and Table 3). In addition, all patients with X-linked disorders had de novo mutations (Table 1A and Tables 2, 3). WES detected a low-level mosaicism in one patient with X-linked de novo variant in CDKL5, which was consistent with his phenotype (Table 1). WES identified a mutation in G6PC causing GSD1a (Table 1) in a patient who was clinically suspected to have GSD III or IX due to cardiomyopathy and hepatic manifestations but had normal leukocyte enzymes for GSD III and IX. Establishing GSD1a diagnosis in this patient enabled appropriate counseling and genetic screening of the family members.

WES was especially helpful in a few patients with suspected mitochondrial disorders in whom extensive mitochondrial work-up was not revealing. Mutations in FBXL4 and C10orf2 causing mitochondrial DNA depletion syndrome were identified (Table 2). WES facilitated a rapid diagnosis in these patients with mitochondrial disorders, avoiding a need for invasive muscle or skin biopsies.

WES also provided a relatively rapid diagnosis in patients with treatable conditions, such as glycogen storage diseases (Table 1B). In addition, it provided a diagnosis in patients with atypical features of rare diseases, such as the patient with global developmental delay, repetitive hand movements, and brain atrophy without the molar tooth sign. This patient was confirmed to have Joubert syndrome due to a homozygous mutation in AHI1, c.1051C > T (p.R351X) (Table 3). In addition, WES revealed pathological variants in ARID1B in two patients (Table 3); one of them was suspected to have mucopolysaccridosis. Founder mutations were identified in this tribal population, such as the pathologic variant causing transaldolase deficiency (Table 1). This finding facilitated further diagnoses in our community using single gene sequencing.

It is important to note that the results of WES should be appraised carefully. In one patient, WES reported a deleterious hemizygous mutation in SLC35A2, c.617_620del (p.Q206fs),congenital disorder of glycosylation, TYPE IIm (CDG2M). However, glycosylation screening was normal in this patient and three healthy brothers had the same mutation (not listed in the table). In another patient, WES reported a pathogenic mutation in MCCC2, c.1015G > A (p.V339M), but urine organic acids were normal and the patient had Sandhoff disease (confirmed by enzyme studies), Table 1.

In most cases, extensive and expensive investigations (including biochemical, genetic, and segregation studies) were necessary after the results of WES to determine the pathogenicity of the variants. In UAE, most families are consanguineous and have multiple affected individuals. Therefore, it was somewhat possible to stratify many of the variants of unknown significance as pathogenic vs non-pathogenic. Nevertheless, challenges remain in the interpretation of the 12 likely pathogenic variants (Table 1b). The availabilities of functional studies, improved knowledge of rare allele frequencies of healthy individuals, and more routine use of WES will improve the yield of this powerful analytical tool.

Competent staff is needed to explain WES results and to counsel patients and their families. The metabolic centers should develop and adhere to appropriate guidelines for utilizing WES as a diagnostic tool. Position statements and best practice guidelines are available from the American college of Medical Genetics, Canadian college of Medical Genetics and European society of Human Genetics [12–14].

Many variants are confined to ethnic populations and are, thus, difficult to diagnose by conventional genetic methods [2, 15]. The cost of diagnosing genetic disorders using conventional testing was estimated to be about $25,000 per patient [4]. However, it is unknown whether WES would increase or decrease this cost. In this study, the cost of WES was covered by local insurance companies for UAE citizens; the cost is a critical issue for the majority of patients who are not citizens of this country.

Conclusion

In this study, with the majority of patients had consanguineous families, WES identified pathogenic mutations in 50 % of cases. WES also provided population-specific information, which should facilitate molecular diagnosis of metabolic and genetic disorders in the region. However, further studies, including cost-effective assessments are needed before routine application of this test in IEM and genetic clinics.

Abbreviations

BWA, burrows-wheeler aligner; CDG2M, congenital disorder of glycosylation, TYPE IIm; CFTR, cystic fibrosis transmembrane conductance regulator; DNA, deoxyribonucleic acid; EDTA, ethylenediaminetetraacetic acid; G6PD, glucose-6-phosphatase dehydrogenase; GSD1A, glycogen storage disease Ia; IEM, inborn errors of Metabolism; MIM, mendelian Inheritance in man; RNA, ribonucleic acid; SNP, single nucleotide polymorphism; TX, Texas; UAE, United Arab Emirates; USA, United States of America; WES, whole exome sequencing; WGL, whole genome laboratory