Diagnostic yield and genetic landscape of rare pediatric diseases in Vietnam identified by exome sequencing

Abstract

Childhood diseases, especially those involving genetic factors, sometimes present challenges in clinical confirmation due to their nonspecific manifestations and multisystem involvement. Understanding genetic etiology, the prevalence of these diseases, and establishing an appropriate genetic approach were key to timely intervention and effective disease management, particularly in low-resource countries. In Vietnam, a lack of data has been reported regarding the diagnostic yield and variant spectrum of rare and complicated diseases in pediatric patients. A retrospective study was conducted on nationwide pediatric patients presenting clinical abnormalities suspected of genetic causes, whose genetic data should be available. Data was extracted from medical reports and statistically summarized. About 369/958 (38.52%) probands were diagnosed with related genetic diseases; of them, 35.3% had single-gene disorders, and 5.1% had copy number variants. Over 70% of participants were 5 years old or younger (mean age of 3.13 ± 4.24 years). Copy number variant sequencing (CNV-seq) revealed 56 copy number variants (75.0% micro-deletions, 21% micro-duplications). Whole-exome sequencing (WES) as well as a 4503-gene panel detected 448 variants (90% of autosomal variants). In 338 single-gene cases, related phenotypes mainly affected the neurological system (31.1%). Using advanced genetic testing in a local setting, as WES and a 4503-gene panel integrated CNV-seq, could help achieve diagnostic yields comparable to established international programs. This study offers a preliminary approach to clinical diagnosis strategies and therapeutic awareness for genetic disorders in childhood illnesses in Vietnam.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-48963-9.

Introduction

More than six thousand rare diseases identified globally collectively affect children’s health¹. Due to their complexity, long-term impact, and life-threatening consequences, these diseases introduce substantial challenges for both the affected families and the broader healthcare system². Genetic components are evidenced to be involved in 80% of rare diseases. About 70% of them are present in childhood, which may constitute approximately 10% of the population^3,4.

The evolution of genetic sequencing technologies and their implementations in routine clinical practice have improved the diagnostic rates as well as medical management in children’s healthcare⁵. Moreover, diagnostic approaches could bring advantages in timely intervention and potential therapies, thus reducing intensive care rate and complications⁶. DNA microarray analysis and next-generation sequencing (NGS) have facilitated the discovery of numerous novel genes linked to rare genetic disorders, significantly enhancing diagnostic accuracy and efficiency⁷. Additionally, applying NGS in Clinical exome sequencing (CES) and whole exome sequencing (WES) provide a cost-effective and technically accessible strategy for comprehensive rare disease analysis and research⁸.

For years, genetic testing has been primarily recommended for practical situations, such as when explicit symptoms or a familial history suggest a genetic condition, especially if early treatment can improve health outcomes and prognosis⁵. In clinical practice, confirming a diagnosis of genetic diseases is challenging due to their nonspecific clinical manifestations, multisystem involvement, and limited physician familiarity—often resulting in a prolonged diagnostic odyssey. In a population-based study in China, diagnostic delays or misdiagnosis were evidenced to be more frequent in adults compared to non-adults⁹. Additionally, from 1-year-old, pediatric patients were 1.09 to 1.113 times more likely to be misdiagnosed compared to those under 1 year⁹. In the Netherlands, the neonatal intensive care unit (NICU) data showed that up to 37% of infants were positive with genetic disorders, including those without congenital abnormalities found¹⁰. The study in China demonstrated a high prevalence of deceased newborns in NICU responding to genetic disorders (19.7%) without appropriate clinical diagnoses as well as medical actions before death. Studies in the US cohorts showed consistently that genetic disorders were underlying causes of deaths in up to 41% of neonatal and infant patients in NICU¹¹. Apparently, the genetic conditions have induced children’s longer stay in the NICU and the increased burden of expense, which are promisingly enhanced by appropriate diagnosis and timely treatment, if available¹².

As genetic causes also play a main role in congenital anomaly in pediatric patients, sporadic studies in Vietnam reported the high prevalence of congenital abnormalities ranging from 6.2 per 1000 to 38.4 per 1000^13,14, depending on regions. Generally, the Ministry of Health in 2017 reported that up to 2% of children have birth defects in Vietnam¹⁵. Nevertheless, the reports from the Vietnamese database regarding genetic childhood diseases were lacking. To generate the first comprehensive report about genetic causes of children abnormalities in Vietnam, this study aims to investigate the prevalence and genetic characteristics of abnormal conditions in children in a large Vietnamese cohort.

Results

Among 958 patients, 369 probands were diagnosed with genetic causes, leading to an overall genetic diagnostic rate of 38.52% in our study population. Accordingly, single-gene disorders were detected in 338 cases and copy number variants were in 49 cases, accounting for 35.3% and 5.1%, respectively (Table 1). Among them, 18 cases were confirmed with both monogenic and copy number variants. While the G4500 panel yielded 37.3% cases with positive reports, the WES was shown to have a higher rate of 45.2%, both integrated with CNV-seq (Table 1).

Table 1.

Genetic diagnosis of 958 cases.

	Diagnostic rate (%)
	Overall	G4500 + CNV-seq	WES + CNV-seq
Number of cases	958	812	146
Negative cases	589	509	80
Positive cases	369 (38.52%)	303 (37.3%)	66 (45.2%)
Single-gene variants	320 (33.4%)
Copy number variants	31 (3.2%)
Both	18 (1.9%)

WES: Whole exome sequencing. G4500: Genetic panel of 4503 genes. All G4500 and WES were integrated with Copy number variant sequencing (CNV-seq).

In Table 2, age and sex of study participants were described, stratified by genetic abnormalities (copy number variants or monogenic anomalies). Over 70% of participants were 5 years old or younger, with a mean age of 3.13 ± 4.24 years. A similar age trend was shown in all positive cases as well as in the two subgroups. In detail, the single-gene disorders had a higher mean of age (3.20 ± 4.37) compared to the latter group (2.78 ± 4.40), copy number variants. In the context of 26.2% of data missing for sex information, males were the dominant group in 39.8% of the total participants, which increased to 53.8% of single-gene cases.

Table 2.

Characteristics of study participants.

Characteristics	Total (N = 958)	Positive cases (N = 369)	Copy number variants* (+) (n = 49)	Single-gene variants* (+) (n = 338)
Sex (n, %)
Male	382 (39.8)	171 (46.3)	19 (38.8)	181 (53.55)
Female	326 (34.0)	128 (34.7)	21 (42.9)	122 (36.09)
missing	251 (26.2)	70 (19.0)	9 (18.4)	35 (10.4)
Age (mean, SD) (years)	3.13 ± 4.24	3.16 ± 4.39	2.78 ± 4.40	3.20 ± 4.37
Age groups (n, %)
≤ 5 years old	709 (73.9)	275 (74.3)	38 (77.6)	252 (74.56)
> 5 years old	209 (21.8)	88 (23.9)	11 (22.4)	82 (24.26)
missing	41 (4.3)	6 (1.9)	0 (0.0)	4 (1.2)

*Positive reports (+) by CNVs (copy number variants sequencing) and WES/ G4500 (Whole exome sequencing or Genetic panel of 4503 genes).

While most of the chromosomal cases carried structural abnormalities with multisystem related phenotypes, phenotypes related to single-gene variants were varied, being described in Table 3. The five most phenotypes belong to neurology (31.10%, 105/338), metabolic disorders (16.86%, 57/338), multisystem affected (16.57%, 56/338), hematology (8.9%, 30/338), and cardiology (7.1%, 24/338).

Table 3.

Phenotypes related to single-gene variants detected in 338 positive cases with top related diseases.

Classification	Single-gene cases	%
	n = 338
Neurology	105	31.10%
Developmental and epileptic encephalopathy (26), Dravet syndrome (10), Charcot-Marie-Tooth disease type 2 S (7), Familial infantile convulsions with paroxysmal choreoathetosis (6), Neurofibromatosis (4)
Metabolic disorders	57	16.86%
Citrullinemia (11), Glycogen storage disease (10), Mitochondrial DNA depletion syndrome (4), Methylmalonic aciduria (3), Ornithine transcarbamylase deficiency (2)
Multisystem	56	16.57%
Hematology	30	8.90%
Beta thalassemia (4), G6PD deficiency (8), Spherocytosis (7), Thrombophilia (4), Hemophilia A (2)
Cardiology Cardiomyopathy (16), Hypercholesterolemia (3), Atrial septal defect (2), Brugada Syndrome (1), Pulmonary arterial hypertension (1), and Left ventricular noncompaction (1)	24	7.10%
Immunology	19	6.09%
Immunodeficiency (6), Autoimmune lymphoproliferative syndrome (3), Autoimmune polyendocrinopathy syndrome (3), Hemophagocytic lymphohistiocytosis (2), X-linked immunodysregulation, polyendocrinopathy, and enteropathy (2)
Gastroenterology	15	4.35%
Top liver diseases: Cholestasis (5), Dubin-Johnson syndrome (5), Gilbert syndrome (2)
Endocrinology	10	2.90%
Hereditary pancreatitis (6), Diabetes mellitus (2), Thyroid dyshormonogenesis (2), Familial hyperinsulinemic hypoglycemia (2)
Skeletal disorder Osteogenesis imperfecta (2), Hypophosphatemic rickets (2), Mucopolysaccharidosis II (1), Craniofacial-skeletal-dermatologic dysplasia (1), and Achondroplasia (1)	7	2.03%
Pulmonology Cystic fibrosis (3), Pulmonary alveolar microlithiasis (1)	4	1.16%
Dermatology	3	0.87%
Oculocutaneous albinism (1), Dyschromatosis symmetrica hereditaria (1), Dermatitis (1)
Musculoskeletal disorders	2	0.58%
Osteogenesis imperfecta (2), Marshall syndrome (1), Achondroplasia (1), Kaufman oculocerebrofacial syndrome (1), Frontometaphyseal dysplasia (1)
Cardioskeletal myopathy	1	0.29%
Nephrology	1	0.29%
Neuromuscular	1	0.29%
Others	3	0.87%

Copy number variants

Copy number variants were confirmed to be the genetic etiology of 49 probands (Supplementary table S1. Description of 49 cases with copy number variants). From those patients, a total of 56 chromosomal variants were found by CNV-seq (Fig. 2). The majority were structural abnormalities, including 75.0% of micro-deletions, 21% of micro-duplications, (top phenotypes are Angelman/Prader-Willi syndrome (n = 9), Williams-Beuren Syndrome (n = 4), DiGeorge syndrome (n = 2), Wolf-Hirschhorn Syndrome (n = 2), and 22q11 duplication syndrome (n = 2)), and only 4% belonged to trisomy (01 trisomy 21 and 01 triple X) (Fig. 2. A). Chromosome 15 (Chr15) hosting 10 variants was the common location with most aberrations. The remaining were commonly located in Chr7 (n = 6), Chr17 (n = 5), and Chr22 (n = 5). In our cases, no copy number variants were found in Chr10, Chr12, Chr20, and ChrY (Figs. 1 and 2.B).

Fig. 1.

Study flowchart.

Fig. 2.

Chromosomal abnormalities in 49 cases. A. Fifty-six genetic variants reported from CNV-seq. B. Distribution of 56 genetic variants across the 23 chromosome pairs.

There were some clustering hotspots for copy number variants such as del(15)(q11.2-q13.1) (9 cases), del(7)(q11.23) (4 cases), del/dup(22)(q11.21) (4 cases), and del(4)(p16.3) (2 cases).

Single-gene disorders

Single-gene causality was confirmed in 338 probands by WES/G4500. Significantly, a total of 448 single-gene variants were associated with probands’ disorders (Supplementary table S2. Description of 338 cases with single-gene abnormalities). Most of them are pathogenic and likely pathogenic variants. About 154 variants, though VUS, were reported with a sufficient explanation of the patient’s manifestation by physicians. The variants’ mode of inheritance was reported as 44% autosomal dominant (AD) and 46% autosomal recessive (AR) variants. The remaining 10% were X-linked conditions, highlighted with 8% of recessive X-linked and 2% of dominant X-linked variants (Fig. 3.A).

Fig. 3.

Single-gene variant findings among 338 positive cases by G4500/WES. (A) Inheritance mode of 448 variants. (B) Type of single-gene variants.

These monogenic variants led to different types of mutations: missense (54%), frameshift (15%), and nonsense mutations (14%). A small proportion of the variants were splice site (8%), in-frame (3%), and intron (4%) mutations. Other types, such as downstream, promoter, silent, and start-lost, were occasionally observed, making up 2% of all variants.

Top five related phenotypes of single-gene disorders were listed: Citrullinemia (n = 11), Dravet syndrome (n = 10), Glycogen storage disease (n = 10), and G6PD deficiency (n = 8), and KCNQ2-related developmental and epileptic encephalopathy (n = 7). There were some clustering hotspots for sing-gene variants such as SLC25A13 c.852_855del (p.Met285ProfsTer2) (8 cases), PRRT2 c.649dup (p.Arg217ProfsTer8) (4 cases), IGHMBP2 c.1334 A > C (p.His445Pro) (4 cases), HBB c.52 A > T (p.Lys18Ter) (4 cases), and G6PD c.1376G > T (p.Arg459Leu) (3 cases).

Discussion

Over the years, numerous studies have explored the utility of genetic sequencing modalities, including whole genome sequencing (WGS), WES, as well as specific clinical genetic panels, in determining the genetic etiology of pediatric conditions¹¹. However, the diagnostic yield has been variably reported across studies, ranging from 20% in the Netherlands to 72.2% in a Polish study^16,17. Furthermore, the emerging rapid WES (rWES) reported its ability in defining 31% of cases with suspected monogenic disorders¹⁸. In a meta-analysis with 24,631 probands, genome sequencing (GS) was proven to have a higher diagnostic yield than exome sequencing (ES), as 30.6% and 23.2% of pooled estimates, respectively¹⁹. The current study revealed a 38.52% diagnostic rate for genetic disorders among high-risk Vietnamese pediatric patients. Using NGS for DNA sequencing, the panel using 4,503 genes yielded a diagnostic rate of 37.3%, while WES achieved 45.2%.

The inconsistent rates between studies are likely attributable to variations in study subject inclusion: differences in age ranges, clinical dysfunctions, severity of illness, and the inherent characteristics of study populations, such as race. In those studies, patient ages ranged from neonatal to adolescence. Evidenced by a Chinese study, genetic illnesses appeared to be more frequent in younger ages, which was shown as a higher diagnostic yield in neonates (43.2%) compared to overall subjects (36.6%)²⁰. Moreover, considering what kinds of dysfunctions and how severity of the conditions, the study in Poland showed the largest genetic diagnosis rate (72.2%) in the infants with metabolic problems, and with 100% of NICU stay¹⁷. Indeed, some other studies also observed different diagnostic rates by groups of clinical abnormalities, ranging highest in skeletal dysplasias, neurological or metabolic phenotypes^21,22. Furthermore, the analysis modality used also significantly affects the yield. Most of the previous studies described WGS or WES rather than target gene panels in confirming genetic disorders, most considering a rapid or ultra-rapid genomic sequencing process, which was proven to enhance diagnostic rates as well as the cost of care for newborns in intensive care units²³. Interestingly, two studies used target gene panels (3,426 and 4,813 genes)^24,25, in a limited number of participants (n = 23 and n = 20), showing diagnostic rates of 30% and 40%, respectively. It appears that using WGS/WES rather than disease-related gene panels, subjecting newborns or infants rather than older children, and observing in NICU patients rather than non-NICU patients, reveals higher diagnostic yields of genetic conditions¹¹. In discussion with the current study, the study involved two-thirds of the probands as 5 years old or younger. The most phenotypes in single-gene cases were related to neurology (31.10%, 105/338), metabolic disorders (16.86%, 57/338), multisystem affected (16.57%, 56/338), hematology (8.9%, 30/338), and cardiology (7.1%, 24/338); however, the severity of conditions was not evaluated and reported. Considering the varying inputs from the aforementioned studies, including our own, the sample size of each study also influences the diagnostic yield when analyzed separately. Fortunately, a meta-analysis of over 20,000 pediatric patients established pooled diagnostic rates for WES at 36% and WGS at 41% for detecting genetic disorders in ill infants²⁶. The diagnostic yield observed in our study aligns with the average values reported in prior research, showing comparable performance of WES and a 4503-gene NGS-based panel within a low-resource context. Collectively, our findings highlighted that advanced genetic testing can be implemented effectively, achieving significant diagnostic success even in settings with limited resources, such as Vietnam.

The study cases were dominated by single-gene conditions (35.3%) and minorly contributed by chromosome disorders (5.1%). Overall, 96% of the chromosomal abnormalities were structural aberrations and 90% of single-gene variants were associated with autosomal disorders. The genetic pattern in this study was in line with the study by Guihua Lai et al.²⁷, of which, deletion aberrations ranked first in CNVs reports (77.3%, 17/22 variants) and autosomal inheritance accounted the largest proportion in monogenic diseases (70.49% of AD and 13.12% of AR). X-linked conditions were also observed in our cohort (10%, with 8% being X-linked recessive disorders). The X-linked observation is slightly higher in the report by Guihua Lai et al. (13.12%), potentially explained by a higher male-to-female child ratio (85/59) compared to our study subjects (381/326). However, it is important to note that 26.2% of our cases were missing information on sex information, which may impact the accuracy of the comparison. The chromosomal abnormalities were predominantly located on chromosomes 7, 15, 17, and 22, whose regions are associated with well-characterized syndromes such as 22q11.2 deletion syndrome, Williams syndrome, and Angelman/Prader-Willi syndromes. Such abnormalities are frequently overlooked with WES/CES alone. Our approach addresses this by incorporating CNV-seq into WES/CES, enhancing detection capabilities.

Our study was lit with several strengths. It was the first in Vietnam to offer comprehensive data on pediatric genetic diseases using modern sequencing technologies. In a low-resource setting, no specific guidelines or subsidies for genetic analysis, our selective approach, integrating CNV-seq with either WES or the G4500 panel, allowed us to provide a lower-cost test with an affordable turnaround time, significantly improving accessibility for patients and their families. Moreover, in a practical setting, our study’s results would raise awareness of available genetic approaches for high-risk children in Vietnam, leading to the next step of developing healthcare planning, optimizing resource allocation, and setting realistic expectations for both clinicians and families managing diagnosed cases. Crucially, our findings also addressed the underrepresentation of Southeast Asian populations in global genomic databases. This vital contribution may even lead to the discovery of population-specific genetic factors in disease causation.

Several limitations warranted discussion. As a retrospective study, data extraction was constrained by its secondary assessment nature. Some information may add further analysis to our study, such as sex information for 26.2% of our subjects was lacking. Furthermore, clinical data were collected from multiple pediatric centers without standardized phenotypic descriptions, such as those provided by systems like Human Phenotype Ontology (HPO). This lack of uniformity might consequently limit the analysis of genotype-phenotype correlations and hinder the accurate identification of pathogenic variants causing diseases. Our analysis included only cases that underwent genetic testing at our center. However, it’s important to note that our laboratory is one of Vietnam’s largest-scale facilities offering cutting-edge genetic sequencing methods, and our sample was recruited from five national children’s hospitals across the country. We believe that the results could generate the most updated and comprehensive data that would be a representation of Vietnamese children. Currently, WES/WGS offers broader coverage and improved detection of various types of genetic variants. Even so, a great number of our cases underwent NGS-based panel (n = 812) rather than WES (n = 146) because of limited resources, leading to the possibility of overlooking some aberrations. Fortunately, our study reported a quite high diagnostic yield of a 4,504-gene panel of 37.3% compared to previous evidence.

In conclusion, this pioneering study in Vietnamese children demonstrated that advanced genetic testing in a local setting can achieve diagnostic yields comparable to established international programs. The genetic profile observed in Vietnamese children contributed valuable data to global rare disease knowledge and supported the continued development of genomic medicine in Vietnam. Future directions should include longitudinal follow-up to assess clinical impact, expand modalities to increase diagnostic yield further, and create Vietnamese-specific databases to enhance variant interpretation.

Materials and methods

Data collection

This is a retrospective study collecting 2-year period data from January 2023 to December 2024 across 05 nationwide children’s hospitals in Vietnam. Study subjects belong to a cohort of pediatric patients presenting with clinical abnormalities suspected of genetic causes, especially those who were unexplained by any non-genetic disorders and were indicated for genetic analysis by their physicians.

Pediatric patients were indicated for WES with CNV-seq, or for a 4,503-gene panel integrated with CNV-seq, when they presented with a complex clinical phenotype involving multiple possible differential diagnoses.

Only individuals who had undergone genetic analysis and whose genetic reports were available were eligible for study involvement. Study information at baseline such as children age at diagnosis, biological sex, major clinical symptoms observed by physicians were obtained from medical records. The study flow chart is shown in Fig. 1.

All participants’ legal guardians provided informed consent for the use of their de-identified information and reports of their genetic analysis for the study. The study protocol was reviewed and approved by the Medical Genetics Institute Committee (Approval number: 07/2025/CT-VDTYH). The study was conducted in accordance with the Declaration of Helsinki.

DNA sequencing

Whole exome sequencing (WES) and clinical exome sequencing (CES)

For WES, a DNA library was constructed by using a Nextera Rapid Capture Exome Kit (Illumina, USA) capturing over 98% of the exonic contents. The library concentration was quantified by a Qubit dsDNA Broad Range Assay Kit (Invitrogen, USA). Library size was measured by a Lab Chip 3 K Hisense Kit (Perkin Elmer, USA). Paired-end exome sequencing with a read length of 75 × 2 bp was performed on a HiSeq. 4000 (Illumina, USA).

For CES, a panel of 4503 genes related to most common genetic disorders in children was applied. The libraries were prepared using a New England Biolabs (MA, USA) kit, enriched using targeted xGen probes from Integrated DNA Technologies and sequenced in the Illumina Nextseq System with a minimum mean coverage of 100x and at least 95% of target genes more than 10×. The sequencing result was aligned to reference genome GRCh38 to identify variants. Based on the genetic database for Vietnamese (including the 1000 Exome Sequencing Project), variants whose frequencies are lower than 1% in the population would be classified.

The region of interest (ROI) in this study was restricted to coding regions and exon–intron boundaries of the analyzed genes. Both the G4500 targeted panel and whole-exome sequencing (WES) were designed to detect single-nucleotide variants and small insertions/deletions (< 10 nucleotides) located within exons and the immediately flanking splice-site regions (± 10 nucleotides from exon–intron junctions). Reported variants underwent comprehensive quality assessment and were confirmed by Sanger sequencing when potential quality concerns were identified, including low read depth, pseudogene interference, or sequence complexity–related artifacts.

Copy number variant sequencing (CNV-seq)

CNV-seq is a copy number analysis based on NGS under NextSeq 500 platform (Illumina, San Diego, USA) that can detect all chromosomal aneuploidies and abnormalities in copy number variants with a size of at least 400,000 nucleotides.

Sequencing data quality control was conducted with the FastQC (version 0.11.9). DNA samples were sequenced on the NextSeq 550 platform using a paired-end 2 × 75 bp Reagent Kit (Illumina). Raw sequences from each sample were aligned to human reference NCBI build GRCh38 using BWA and SAM tools packages. Duplicate reads were removed with Picard tools (Broad Institute). The final output data were collated on OMIM and DECIPHER databases to classify CNV variants. The CNV-seq results were sub-grouped by structural (deletion, duplication) and numerical (aneuploidy, monoploidy) chromosomal abnormalities.

All detected variants were then categorized as pathogenic, likely pathogenic, likely benign, benign variants, or variant of uncertain significance (VUS) according to the ACMG²⁸. Pathogenic, likely pathogenic, and VUS variants were reported for copy number and single gene variants.

In this study, the diagnostic yield was determined based on pathogenic or likely pathogenic variants, or highly suspicious VUS that were considered disease-causing after clinical correlation.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Author contribution statement: Conceptualization : H.G, H.S.T, HN.N, YT.L; Investigation: TA.T, TH.L.N, T.M.N, CD.V, TH.T.N, LP.H, AH.N.P, P.M.V, P.N.C, T.D.T; Methodology: H.G, TT.T.D, HN.N, U.V.T; Data Curation : TA.T, KH.T.N, TH.L.N, KV.T.T, K.V.N, TP.N.H, Y.N.N, L.T.L, U.V.T; Formal Analysis: HT.T.T, YT.L, H.S.T; Writing-original draft: HT.T.T, YT.L, U.V.T; Supervision: H.S.T, H.G; All authors approved the manuscript before submission.

Funding

This study did not receive any funding in any form.

Data availability

The datasets generated and/or analyzed during the current study are available in the SRA database repository [ACCESSION NUMBER: PRJNA1333793]. All variants reported in this study are deposited in ClinVar with reserved accessions [SCV006550975-SCV006551383].

Declarations

Competing interests

The authors declare no competing interests.

Conflict of interest

All the authors declare no conflict of interest.