If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Exome sequencing has recently been proved to be a successful diagnostic method for complex neurodevelopmental disorders. However, the diagnostic yield of exome sequencing for autism spectrum disorders has not been extensively evaluated in large cohorts to date.
Materials and Methods
We performed diagnostic exome sequencing in a cohort of 163 individuals with autism spectrum disorder (66.3%) or autistic features (33.7%).
Results
The diagnostic yield observed in patients in our cohort was 25.8% (42 of 163) for positive or likely positive findings in characterized disease genes, while a candidate genetic etiology was reported for an additional 3.3% (4 of 120) of patients. Among the positive findings in the patients with autism spectrum disorder or autistic features, 61.9% were the result of de novo mutations. Patients presenting with psychiatric conditions or ataxia or paraplegia in addition to autism spectrum disorder or autistic features were significantly more likely to receive positive results compared with patients without these clinical features (95.6% vs 27.1%, P < 0.0001; 83.3% vs 21.2%, P < 0.0001, respectively). The majority of the positive findings were in recently identified autism spectrum disorder genes, supporting the importance of diagnostic exome sequencing for patients with autism spectrum disorder or autistic features as the causative genes might evade traditional sequential or panel testing.
Conclusions
These results suggest that diagnostic exome sequencing would be an efficient primary diagnostic method for patients with autism spectrum disorders or autistic features. Moreover, our data may aid clinicians to better determine which subset of patients with autism spectrum disorder with additional clinical features would benefit the most from diagnostic exome sequencing.
Autism spectrum disorders (ASDs) are highly heritable neurodevelopmental disorders seen in 1% to 2% of children with varying degree of symptoms and severity.
Single gene testing in such a heterogeneous group of disorders is challenged by profound locus and clinical heterogeneity. Since its introduction in 2011, diagnostic exome sequencing (DES) has proven instrumental in providing a molecular diagnosis for many patients with a broad spectrum of previously undiagnosed genetic disorders
Traditionally, karyotyping, chromosomal microarray (CMA), and fragile X testing are performed as first-tier tests for ASDs with varying diagnostic yields (2.2% to 2.5%, 9.3 to 24%, and 0.4% to 8%, respectively).
Confirmation of chromosomal microarray as a first-tier clinical diagnostic test for individuals with developmental delay, intellectual disability, autism spectrum disorders and dysmorphic features.
where a broad search for causal variants across the genome is needed after traditional approaches have proved unsuccessful. Multiple studies have reported the results of exome sequencing in different large ASD cohorts, focusing mainly on simplex patients with ASD and the identification of de novo causative mutations
The estimates vary between 3.1% and 28.6% depending on the complexity of the ASD phenotype. To better elucidate the diagnostic yield of DES for patients with ASD or autistic features, we evaluated the results in a cohort of 163 patients with reported ASD or autistic features referred for DES. For a clinical diagnostic reference laboratory, the evaluation records for ASD are often unavailable, and thus the cohort reported here consists of individuals with either clinician-reported ASD diagnosis or autistic features provided in the clinical evaluations.
Patients and Methods
Patient cohort
The study sample consisted of the first 1200 consecutive samples sent for DES to Ambry Genetics Laboratory, which is a for-profit laboratory. All the authors are employed by the Ambry Genetics Laboratory. Clinicians were encouraged to also refer all first-degree and other informative family members for testing. Solutions Institutional Review Board determined the study to be exempt from the Office for Human Research Protections Regulations for the Protection of Human Subjects (45 CFR 46) under category 4. Being a retrospective data analysis of anonymized data, it was exempted from the requirement to receive consent from patients. Detailed clinical evaluations, records of prior genetic testing, and pedigree information were provided by the referring physicians. The majority of the referring physicians were geneticists (67.8%), while the remainder were specialists in neurology (13.5%), pediatrics (6.1%), or other clinical departments (12.6%). All patient information was carefully reviewed and summarized by the American College of Medical Genetics and Genomics (ACMG) board-certified genetic counselors, licensed in their respective states. Their combined experience in a pediatric clinic totals 47 years, and one of them (Z.P.) has ten years' experience in an autism-specific clinic and is an expert resource for our clinical exome group.
Whole exome sequencing
Genomic DNA was isolated from whole blood from all probands and accompanying family members. Exome library preparation, sequencing, bioinformatics, and data analysis were performed as previously described.
Enhanced utility of family-centered diagnostic exome sequencing with inheritance model-based analysis: results from 500 unselected families with undiagnosed genetic conditions.
Roche NimbleGen EZ Exome System (Madison, WI, USA), or the IDT xGen Exome Research Panel V1.0 (Integrated DNA Technologies, Coralville, IA, USA) and sequenced using paired-end, 100- or 150-cycle chemistry on the Illumina HiSeq (Illumina, San Diego, CA, USA). Stepwise filtering included the removal of common single nucleotide polymorphisms, intergenic and 3′/5′ untranslated region variants, non–splice-related intronic variants, and non–splice-related synonymous variants. Alterations were filtered further based on family history and possible inheritance models. Identified candidate alterations were confirmed, and cosegregation studies were performed using automated fluorescence dideoxy sequencing.
Data analysis
Genes were classified as either candidate or characterized as Mendelian disease causing based on Ambry's clinical validity assessment criteria.
Standards and guidelines for the interpretation of sequence variants: a joint consensus recommendation of the American College of Medical Genetics and Genomics and the Association for Molecular Pathology.
The majority of the variants have been deposited in ClinVar (http://www.ncbi.nlm.nih.gov/clinvar/submitters/61756). Secondary or incidental findings unrelated to the current clinical indication of the probands were excluded from this study. Each gene was assessed for the level of phenotypic overlap leading to one of the following overall primary DES results: positive or likely positive, uncertain, or negative for characterized genetic etiologies and candidate or suspected candidate and negative for candidate genetic etiologies. Different DES testing strategies were available and are illustrated in Fig 1. The calculation of diagnostic rates among characterized genetic etiologies was based on all probands, regardless of the analysis strategy. The calculation of detection rates for candidate genetic etiologies was based on the number of probands in whom analysis of both characterized and candidate genetic etiologies was performed. The result interpretation process is described in detail in Farwell et al., 2014.
Enhanced utility of family-centered diagnostic exome sequencing with inheritance model-based analysis: results from 500 unselected families with undiagnosed genetic conditions.
The study sample was composed of 1200 individuals sequentially ascertained for DES. A total of 163/1200 patients (13.6%) had clinician-reported ASD diagnosis or autistic features reported in the clinical evaluations provided (“ASD cohort”). The majority of the 163 patients (108, 66.3%) had a clinician-reported ASD in the clinical evaluations, while 55 (33.7%) patients had autistic features reported in the clinical evaluations (Table 1). The average age at testing was 9.0 ± 6.7 years in the ASD cohort. Of all the 163 probands in the ASD cohort, 44 were females (27%) and 119 were males (73%) (Supplementary Table 1). Additional clinical features such as intellectual disability (ID)/developmental delays (DDs) (92.6%), epilepsy/seizures (38.7.0%), and psychiatric condition (27.6%) were reported (Table 1).
Table 1Diagnostic Characteristics of the ASD Cohort
Our data revealed that patients with psychiatric conditions or ataxia and/or paraplegia in addition to ASD or autistic features were significantly more likely to receive positive results compared with patients in the ASD cohort without these comorbidities (95.6% vs 27.1 %, P < 0.0001; 83.3% vs 21.2%, P < 0.0001, respectively). These results together with additional clinical characteristics and other organ system involvements are listed in Table 1.
Ninety-six percent of the probands in the ASD cohort had one or more “first-tier” tests (karyotyping, CMA, and fragile X) performed, and of these patients, over 90% had some form of an array performed, while only 33% underwent testing for specific ASD genes before exome sequencing (Supplementary Table 2). Positive family history (first- or second-degree relatives) for ASD, DD, or ID was reported for 25.2% of the ASD cohort and for 19% of the ASD cases with positive or likely positive findings in characterized genetic etiologies.
Overall positive rate in characterized and candidate genetic etiologies in the ASD cohort
We identified positive or likely positive findings in characterized genes in 42/163 (25.8%) patients within the ASD cohort (Fig 2). Among the uncertain and negative findings, proactive reanalysis based on new literature and further cosegregation analysis provided a definitive molecular diagnosis in five genes (FTSJ1, KCNH1, NR2F1, MTRR, and ZBTB20) among five patients.
Figure 2Diagnostic yield for characterized genetic etiologies and reanalysis results. *Candidate or suspected candidate genetic etiology was reported in seven families and in three cases; clinical significance was corroborated by new pertinent information. **Five more cases were reclassified as positive due to proactive reanalysis based on new publications and further familial cosegregation analysis. (The color version of this figure is available in the online edition.)
Originally, three of 120 (2.5%) patients in the ASD cohort received a report with a candidate genetic etiology (TRIP12, HDAC1, and SETD5). Suspected candidate genetic etiologies were reported for an additional four patients (MTOR, HTR2C, RYR3, and MN1) (four of 120; 3.3%). Further functional evaluation, subsequent publications, or additional cases with similar clinical presentation sent to our laboratory confirmed clinical significance in three cases with candidate genetic etiology (MTOR,
De novo loss-of-function mutations in SETD5, encoding a methyltransferase in a 3p25 microdeletion syndrome critical region, cause intellectual disability.
and TRIP12), which were later reclassified as positive characterized gene findings (Fig 2).
Interestingly, of the 108 individuals with ASD diagnosis described in clinical evaluations, 20 (18.5%) received a positive diagnosis in characterized genetic etiology, while eight (14.5%) patients with autistic features received a positive diagnosis (P = 0.05) (Supplementary Figure 1). Moreover, all four individuals with a candidate genetic etiology (3.7%) reportedly had ASD diagnosis.
Positive gene findings
Among the 42 total positive or likely positive findings in characterized genetic etiologies, 42 unique genes with pathogenic or likely pathogenic alterations were identified (Table 2). Pathogenic alterations in four genes (SETD5, ADNP, MECP2, WDR45) were identified in two individuals each, whereas the rest of the genes occurred only once in our cohort. Of the 42 positive cases, 40 have pathogenic or likely pathogenic alteration(s) in genes previously reported with mutations in autistic individuals or implicated in ID, DD, or syndromes such as Pitt-Hopkins syndrome, where autism is a common presentation of the syndrome. Mutations in five genes do not explain the autistic features in the proband. Two of these genes (GRHL3 and FHL1) are an underlying cause for other features in the proband, while mutations in three of these genes (CYP21A2, NOTCH1, and PKD1) were discovered in patients with dual diagnosis. In total, four (9.5%) patients received a dual molecular diagnosis (Table 3). Some of the individuals described here have previously been reported.
Enhanced utility of family-centered diagnostic exome sequencing with inheritance model-based analysis: results from 500 unselected families with undiagnosed genetic conditions.
Positive or likely positive results in characterized genetic etiologies
Of the 42 positive gene findings, 26 (56.5%), eight (17.4%), five (10.9%), and six (13.0%) were associated with autosomal dominant, autosomal recessive, X-linked dominant, and X-linked recessive conditions, respectively. One patient had a de novo pathogenic alteration in a mitochondrial gene, MT-ATP6. The inheritance patterns for pathogenic or likely pathogenic molecular findings in the ASD cohort are summarized in Supplementary Table 4.
Altogether, we identified 51 alterations in 42 unique genes. Twenty-six of the total unique alterations were de novo (51.0%), 22 were inherited (43.1%), and three (5.9%) were of uncertain origin that could not be confirmed due to one or both parents being unavailable for testing. In total, confirmed de novo events contributed to 61.9% (26 of 42) of the positive or likely positive findings in characterized genetic etiologies in the ASD cohort (Supplementary Table 4).
Candidate genetic etiologies
Of the four candidate genetic etiologies reported (Table 4), three (75.0%) were proposed to be autosomal dominant alleles (two de novo and one apparently de novo due to likely gonadal mosaicism) and one (25.0%) was associated with inherited biallelic changes (Supplementary Table 4).
Table 4Genes Identified Among Candidate Genetic Etiologies From Patients in the ASD Cohort
We examined the rate of positive findings of DES in 163 patients from the ASD cohort and observed 25.8% diagnostic yield in characterized genetic etiology. Seven patients (seven of 120, 5.8%) received a report with a candidate genetic etiology (TRIP12, HDAC1, SETD5, MTOR, HTR2C, RYR3, and MN1). Three of these genes were later reclassified to characterized findings based on corroborating evidences from the literature (SETD5, TRIP12, and MTOR).
De novo loss-of-function mutations in SETD5, encoding a methyltransferase in a 3p25 microdeletion syndrome critical region, cause intellectual disability.
One originally negative and four uncertain patients were subsequently provided with a definitive diagnosis, highlighting the value of a proactive reanalysis process that is based on new literature as well as further cosegregation analysis. Among all 42 positive or likely positive characterized genetic etiologies reported, five of the 42 unique genes were not related to ASDs based on the known clinical spectrum of the gene. One patient with an alteration in FHL1 gene received a diagnosis of Emery-Dreifuss muscular dystrophy, and another patient with an alteration in GRHL3 gene received a diagnosis of Van der Woude syndrome characterized by cleft lip and/or palate. The remaining three genes were identified in probands receiving dual diagnoses. The fourth patient with a dual diagnosis received two ASD-related diagnoses. ASDs are fairly common disorders and may be unrelated to the other diagnosis in probands with dual diagnoses. Alternatively, the additional diagnosis may represent an expansion of the clinical phenotype that has not yet been reported. The rest of the genes have previously been identified in patients with ASD, ID, and/or DD, or implicated in syndromes in which autism is a common phenotypic feature.
Interestingly, patients presenting with psychiatric conditions or ataxia and/or paraplegia in addition to autistic features were significantly more likely to receive positive results compared with patients without these features, and this may imply that patients with these combinations of phenotypes may benefit more from undergoing DES.
A few of the genes in positive findings, such as MECP2 and SHANK3, are well-established genes for ASDs, whereas the vast majority (92.5%) of the unique ASD-related genes among positive or likely positive characterized and candidate genetic etiologies were recently identified (after January 2012) as “nonclassical” ASD genes with novel pathogenic sequence variants. This fact indicates that alterations in newly characterized genes may account for a significant number of individuals with ASD or autistic features. Indeed, three of the newly characterized ASD genes were recurrent (SETD5, ADNP, and WDR45) among the positive or likely positive findings in characterized and candidate genetic etiologies in the ASD cohort. The identification of recurrent mutations in newly characterized genes from an unselected laboratory cohort corroborates their role in ASDs. However, the phenotypic spectrum and underlying disease mechanism of the recently implicated genes remain to be clarified. Overall, these observations highlight the utility of an unbiased screening method interrogating all genes in the human genome as a molecular diagnostic tool for individuals with neurodevelopmental disorders, including ASDs. In comparison, the gene content included in a gene panel may not be flexible and likely does not include the analysis of the most recently characterized genes.
The advent of cost-effective, trio-based exome sequencing has shed light on the contribution of de novo alterations in ASD incidence. It is estimated that de novo loss-of-function mutations contribute to a minimum of 10% of simplex ASD cases, and de novo missense mutations, up to 10% of affected children.
In our study, the majority of positive or likely positive findings in the ASD cohort were the result of de novo mutations, supporting the de novo paradigm demonstrated in ASDs. Of all the de novo alterations, 53.8% were loss-of-function mutations, and the rest 46.2%, missense alterations.
The majority of the proteins encoded by the genes reported here function in synaptic, transcriptional, and chromatin-remodeling pathways, which have been repeatedly implicated in ASDs.
Moreover, pathogenic mutations were identified in well-established ASD genes as well as in genes associated with other conditions such as ID (CTNNB1, CUL4B, DYNC1H1, ELP2, FTSJ1, IQSEC2, KCNH1, KMT2A, MECP2, MTOR, SETD5, SHANK3, TCF4, UBE3A, WDR45, ZBTB20) and epilepsy (CACNA1A, CHD2, CDKL5, SCN8A, SYN1), supporting the broad phenotypic spectrum and shared molecular pathways. Although the genetic heterogeneity poses a great challenge for developing therapies for ASDs, the few major signaling pathways implicated in ASDs, such as the PI3K-mammalian target of rapamycin signaling cascade, have been targets of active research for therapeutic implications.
Molecular diagnosis through DES can aid in the stratification for clinical trials and inform target identification.
Previous reports assessing the diagnostic yield of DES in ASDs have utilized relatively small samples and reported the highest diagnostic yields among the more complex ASD phenotypes.
reported a diagnostic yield of 16.7% among the most complex cases and 28.6% for those with less evidence of dysmorphology, whereas cases with the least complex phenotype received a molecular diagnosis 3.1% of the time. In our ASD cohort, multiple congenital anomalies were observed in 13.5% of patients and positive findings were slightly more frequent among patients with multiple congenital anomalies when compared with patients without these clinical features (45.5% vs 22.7%; P = 0.03). These results support the findings by Tammimies et al.
on the utility of whole exome sequencing as a first-tier ASD test especially for children with physical and congenital anomalies. The majority of the patients in the ASD cohort have an additional neurological phenotype, such as an epilepsy and/or seizure phenotype or a psychiatric condition in addition to autistic features. Moreover, we observed a higher rate of ID in our ASD cohort compared with the varying estimates (16.7% to 84%) generally reported.
These findings imply that our cohort represents the complex ASD cases that are referred for exome sequencing after long and extensive testing process, and therefore the diagnostic yield reported here may be closer to what is observed in genetics clinics for patients with no recognizable syndromes. The rate of ASD and autistic features is higher in our cohort than in the population (1% to 2%), which was expected because the indication for testing for most individuals was neurological dysfunction.
In addition to the “first-tier” testing (karyotype, CMA, and fragile X), the current ACMG practice guidelines recommend testing for single genes such as UBE3A, PTEN, and MECP2 among many others when suspecting ASDs.
Professional Practice Guidelines Committee. Clinical genetics evaluation in identifying the etiology of autism spectrum disorders: 2013 guideline revisions.
In our cohort, however, positive or likely positive findings in characterized genes and candidate genetic etiologies in all these genes were underrepresented (7.1%) likely because patients with mutations in these genes might already have a positive result through gene or gene panel sequencing and thus there was no need for DES.
Nearly all the patients in our cohort had a “first-tier” test performed, and of these, over 90% had some form of an array before testing in our laboratory. These findings imply that our cohort represents patients undiagnosed with the tests commonly recommended for the evaluation of neurodevelopmental disabilities and generally covered by insurance, whereas DES coverage varies widely between insurance policies. However, our study shows a significant increase in the diagnostic rate for neurodevelopmental disorders compared with traditional “first-tier” testing. This observation is of importance to the patients and their families because a diagnosis may alter the clinical management, help predict recurrence risks, inform prognosis, and end their long and invasive “diagnostic odyssey.” The information is significant to the clinicians who have been unable to provide answers to these patients. Moreover, the high diagnostic rate of DES reflects the potential medical-economic savings to both patients and insurance companies.
The main limitation of this study is the lack of phenotypic detail for many patients, a common issue for many laboratory-based studies. Therefore we were not able to confirm a definite ASD diagnosis in all patients, although the majority (66.3%) of the probands assigned to the ASD cohort had a physician-reported ASD diagnosis. Our findings did not differ significantly between probands with an ASD diagnosis and probands with autistic features, which provides support for our analysis of the phenotype. We also acknowledge that variation in clinic-specific guidelines and provider preferences regarding when to consider DES (i.e. early in the process after first-tier tests or only after exhausting all available traditional approaches), may also affect our estimate of the DES diagnostic yield in ASD. Although almost all our ASD cases underwent first-tier testing before DES, only 33 % had genetic testing for specific ASD genes, implying that the majority of the ASD cases included in this report may have been tested early in the diagnostic evaluation.
We are grateful to all the participating families and to their physicians and genetic counselors for providing samples and clinical evaluations.
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Appendix
Supplementary Table 1Demographic Characteristics of the ASD Cohort
Confirmation of chromosomal microarray as a first-tier clinical diagnostic test for individuals with developmental delay, intellectual disability, autism spectrum disorders and dysmorphic features.
Enhanced utility of family-centered diagnostic exome sequencing with inheritance model-based analysis: results from 500 unselected families with undiagnosed genetic conditions.
Standards and guidelines for the interpretation of sequence variants: a joint consensus recommendation of the American College of Medical Genetics and Genomics and the Association for Molecular Pathology.
De novo loss-of-function mutations in SETD5, encoding a methyltransferase in a 3p25 microdeletion syndrome critical region, cause intellectual disability.
Professional Practice Guidelines Committee. Clinical genetics evaluation in identifying the etiology of autism spectrum disorders: 2013 guideline revisions.
Conflict of interest: The manuscript summarizes data from Ambry Genetics' exome sequencing test that is among the commercially available tests through Ambry Genetics. The authors of this manuscript are all employed and receive a salary from Ambry Genetics. All authors have access to all relevant data.
30572-0&id=gr1.jpg){kind=link}
30572-0&id=gr2.jpg){kind=link}
30572-0&id=fx1.jpg){kind=link}