Patient demographics

A total of 1436 Iranian index cases were referred for diagnostic NGS testing, with 1075 (74.9%) submitted for exome sequencing and 361 (25.1%) for targeted gene panel sequencing. NGS was performed on 1315 probands only (solo, 91.6%), 42 probands plus parents (trio, 2.9%), and 79 one or both healthy parents of the deceased affected child(ren) (parents, 5.5%). The majority of patient referrals (62.3%), encompassing 76.4% of ES cases (821/1075), were between 2018 and 2020. The breakdown of clinical indications in ES cases over time is presented in Supplementary Fig. 1.

The patient population comprised 777 males (54.1%) and 614 females (42.8%), in addition to 45 (3.1%) fetal cases from terminated pregnancies. The age of probands at testing, categorized into three age groups, ranged between prenatal to 72 years. Notably, parental consanguinity was present in the majority of cases (72.4%), while most patients (65%) had a negative family history (Table 1). Overall, 56% (805/1436) of patients presented with neurological conditions (Supplementary Fig. 2).

Table 1 Patient demographics and testing indications

The distribution of patients’ major clinical indications based on HPO is presented in Table 1. The two most common testing indications, collectively accounting for more than half of all referrals, were abnormality of the nervous system (29.5%) and abnormality of the musculature (26.9%). Developmental delay and/or intellectual disability and seizures were the most frequent indications in patients grouped as an abnormality of the nervous system. Patients within the abnormality of the musculature group mainly presented with muscular dystrophy, myopathy, and myasthenic syndrome. The next most frequent indications for referral were abnormalities of the ear (8.1%), metabolism/homeostasis (6.4%), central motor function (6.2%), peripheral neuropathy (6.1%), skin (4.9%), and eye (3.8%). A complete list of the patient’s phenotypic and genotypic information is provided in Supplementary Table 2.

Spectrum of identified variants

A total of 1115 unique variants in phenotype-related genes were reported across 1005 cases with non-negative results, including 286 (25.7%) pathogenic variants, 374 (33.5%) likely pathogenic variants, and 455 (40.8%) variants of uncertain significance, classified according to the ACMG/AMP guidelines (Supplementary Table 2). Remarkably, about 55.6% (620/1115) of all the reported variants, including 241 (36.5%, 241/660) of the pathogenic and likely pathogenic variants, were novel as defined by not being previously reported in patients described in the literature. A broad spectrum of variant types was observed among all the reported variants, including 612 missense, 186 nonsense, 178 frameshift, 102 splice-site, 24 in-frame insertion or deletions, 8 initiation codon, 2 stop-loss, 2 promoter region, and 1 synonymous. Of 660P/LP variants, 406 (61.5%) were null variants (nonsense, frameshift, ±1 or 2 splice sites, initiation codon), 241 (36.5%) were missense variants, and 13 (2.0%) were other variant types (in-frame insertion or deletions, stop-loss, promoter region, synonymous) (Fig. 1a).

Fig. 1: Characteristics of the reported variants and test yield based on patients’ phenotypes.

a Distribution of the identified variants classified according to the ACMG guidelines and variant types. VUS variants of uncertain significance, P/LP pathogenic and likely pathogenic variants, Null nonsense, frameshift, ±1 or 2 splice sites, and initiation codon variants; Other, in-frame insertions or deletions, stop-loss, synonymous and promoter region variants. b Diagnostic yield based on major clinical indications (top-level HPO term).

Diagnostic yield

Overall, across the 1436 cases, a positive molecular diagnosis was reported in 46.7% (n = 670), a VUS result was reported in 22.1% (n = 318), an unclear result was reported in 1.2% (n = 17), and 30% (n = 431) received a negative result. There was a higher diagnostic yield from patients referred for targeted testing (54.3%, 196/361) compared to ES patients (44.1%, 474/1075). The frequency of positive results was the highest among cases performed as parents (49.4%, 39/79), followed by solo (47%, 618/1315) and trio cases (31%; 13/42) (Table 1). In the analysis of the diagnostic outcome in phenotype groups, the yield was highest for abnormalities of the skin (67.6%), blood and blood-forming tissues (64.7%), the musculature (54.5%), the skeletal system (51.4%), the ear (50%), central motor function (46.1%), and metabolism/homeostasis (45.6%). In contrast, abnormalities of the cardiovascular system (25%) and prenatal development/birth (24.2%) had the lowest diagnostic yield (Fig. 1b).

The distribution of the presumed mode of inheritance in positive cases is presented in Table 2. Autosomal recessive inheritance accounted for the majority of positive cases (79.1%, 530/670), followed by autosomal dominant inheritance (14.6%, 98/670) and X-linked inheritance (6.3%; 42/670). Of the 530 autosomal recessive diagnosed conditions, 441 (83.2%) were linked to homozygous variants, 50 (9.4%) demonstrated compound heterozygosity of two distinct variants, and 39 (7.4%) were cases of heterozygous variants identified in the parent(s) of a deceased patient. Among the 98 autosomal dominant diagnosed conditions, 46 (46.9%) were linked to a de novo variant, 9 (9.2%) were inherited from a symptomatic parent, 4 (4.1%) resulted from parental gonadal mosaicism, and 39 (39.8%) remained undetermined due to unavailability of parental samples. Of the 42 X-linked diagnosed disorders, 36 (85.7%) occurred in males and 6 (14.3%) in females; 15 X-linked alleles were inherited from healthy mothers, 7 resulted from de novo variants, 3 were inherited from a symptomatic parent, and 17 remained with unknown parental origin.

Table 2 Inheritance patterns among 670 positive casesDual diagnoses

Among the 474 cases with a positive ES result, 11 patients (2.3%) received a dual molecular genetic diagnosis (targeted cases were excluded considering the limitations of panel sequencing to detect dual diagnosis comprehensively). The clinical and genetic information of these patients is provided in detail in Table 3 and Supplementary Table 3. The majority of patients in this group were considered to have overlapping phenotypes based on the observation of one or more clinical features associated with both molecular diagnoses, while patients for whom no phenotypic features were shared between molecular diagnoses were categorized as having two distinct phenotypes. Consanguinity was reported for the majority of these cases (81.8%, 9/11), and autosomal recessive phenotypes were expectably the most frequent observation in this group.

Table 3 Patients with dual molecular diagnosisImportant medical implications of diagnostic NGS

Medically actionable secondary findings in the ACMG-recommended list of 59 genes were analysed in 270 cases referred for ES testing. Reportable findings were identified in nine cases (3.3%), which include P/LP variants in MYBPC3 (2), DSG2 (2), MUTYH, OTC, TNNI3, BRCA1 and BRCA2 (Supplementary Table 4). The positive diagnostic results had the potential to impact the treatment and clinical management of several patients. Selected examples included children diagnosed with Wilson disease (MIM:277900), DOPA-responsive Dystonia (MIM: 128230), Pyruvate dehydrogenase E1-alpha deficiency (MIM:312170), Pyridoxine-dependent Epilepsy (MIM: 266100), and Cerebral creatine deficiency syndrome 2 (MIM:612736). Furthermore, predictive testing for subsequent pregnancies (prenatal or preimplantation genetic diagnosis) was performed in 16% (n = 108) of families with positive diagnoses.

Recurrent molecular findings

The majority of the positive cases (68.2%, 457/670) had P/LP variants in a gene at least twice observed in this series (117 genes). These genes were predominantly associated with neuromuscular disorders, sensorineural hearing loss, skin disorders and various neurodevelopmental phenotypes, providing insight into the prevalence of these genetic conditions among Iranian patients. The recurrently mutated genes in general and in each phenotypic group are listed in Supplementary Table 5 and Fig. 3. The most commonly diagnosed disorders in our cohort, each accounting for 1–5% of positive results, were muscular dystrophies related to CAPN3 (n = 34), DYSF (n = 15), LAMA2 (n = 12), SGCA (n = 11), and DMD (n = 22), NF1-related neurofibromatosis (n = 12), ATM-related ataxia (n = 11), MYO7A-related hearing loss (n = 10), CHRNE-related myasthenic syndrome (n = 9), RYR1-related congenital myopathy (n = 8), SPG11-related spastic paraplegia (n = 8), and TYR-related Albinism (n = 7).

Furthermore, of the reported recessive P/LP variants, 59 were identified in two or more unrelated positive cases, and 35 were present in the heterozygous state in at least one healthy Iranian individual in the Iranome database17 (Supplementary Table 6). These variants, collectively accounting for 14.2% (94/660) of all the identified P/LP variants, suggest presumptive founder effects in the Iranian population that require additional haplotype and population-specific analysis for confirmation. Notably, among these presumed founder variants, 28 were absent from other population databases, including gnomAD and the Greater Middle East (GME) variome18. This subset encompasses 19 novel or rare recessive pathogenic variants exclusively reported in Iranian patients up to the present, which could be denoted as “Iranian-only” mutations. It is noteworthy that several of these variants were novel missense variants that could be classified as likely pathogenic only through observation in two unrelated cases with the same phenotypic presentation. These variants include MME c.499T>A (p.Trp167Arg), SGCA c.113A>G (p.His38Arg), FKRP c.1034G>C (p.Gly345Ala), ASAH1 c.109C>A (p.Pro37Thr)19, and OTOA c.1727T>C (p.Ile576Thr).

Unexpected inheritance patterns and molecular events

The predominance of consanguineous families in this cohort led to the identification of bi-allelic variants in 11 genes with only an autosomal dominant pattern of inheritance in the OMIM database at the time of reporting. These genes were either not previously associated with any AR disorders or were reported with recessive inheritance only in extremely rare cases in the literature (Table 4). The clinical presentations in the majority of cases in this series were either similar to or more severe than the established dominant disorder. For example, in three separate cases with variants in TBX4, GLI3 (both published elsewhere20,21), and BICC1 genes, a severe embryonic condition was observed in the homozygous fetuses, while the heterozygous parents were either mildly affected or unaffected. Other remarkable examples include recessive forms of DCTN1-related neurodegenerative disorder, KCNC3-related spinocerebellar ataxia, HARS1-related Charcot-Marie-Tooth disease22, and MITF-related hearing loss23. Unlike previous observations of biallelic PKD1 variants in severe pediatric Polycystic Kidney Disease24,25, a homozygous missense variant in this gene was identified in a 33-year-old patient presenting with bilateral cystic kidneys, hypertension, urinary issues, and an apparently negative family history. This finding is consistent with the published literature on biallelic hypomorphic PKD1 alleles leading to a mild adult-onset phenotype26. In addition, a distinct and only recently reported recessive phenotype for the UFSP2 gene was observed in a patient presenting with global developmental delay, intractable epilepsy, and brain atrophy27. Variants in this gene were previously associated with autosomal dominant skeletal system disorders without neurological dysfunction.

Table 4 Recessive inheritance in dominant disease genes

Furthermore, despite the initial assumption of a recessive genetic condition in 3 families with multiple similarly affected children and healthy parents, we identified heterozygous pathogenic variants in AD disease genes (SETBP1, CRYAA, SMCHD1) in the affected children but neither of the parents, indicating possible parental gonadal mosaicism. (patients 184, 794, 1003 in Supplementary Table 2). Consanguineous marriage suggestive of a typical AR pedigree was interestingly reported in two of these families. Identification of de novo variants in AR disease genes was another unusual molecular event observed in 3 patients who were each compound heterozygote for two P/LP variants in SGCA, ASPM, and ABCA4 genes. In these individuals, one of the variants was parentally inherited while the other had arisen de novo (patients 36, 253, 398 in Supplementary Table 2).

Expanding the mutational and phenotypic spectrum of less characterized genes

Variants in genes with limited mutational/phenotypic evidence in the literature or with OMIM phenotypes based on a single study were identified in 11 patients (MRPS34, FIBP, LRP4, ZBTB11, CRIPT, BCKDK, PLAA, CC2D1A, TUBB2A, SOD1). A detailed description of molecular findings and observed phenotypes in these patients are provided in Supplementary Table 7. These include the third so far described patients with FIBP-related Thauvin–Robinet–Faivre syndrome and LRP4-related congenital myasthenic syndrome, the third variant so far associated with SOD1-related spastic tetraplegia and axial hypotonia, and the first truncating variant associated with TUBB2A-related cortical dysplasia with other brain malformations. Notably, a possible expanded phenotype was observed in a patient with a homozygous MRPS34 variant presenting with chronic motor neuropathy, which deviates significantly from the delayed psychomotor development reported in a single study for this gene28. Similarly, a milder clinical presentation for PLAA-related disorder was observed in two siblings with developmental delay, seizures, and hypotonia.