Advertisement

Genome sequencing is a sensitive first-line test to diagnose individuals with intellectual disability

  • Anna Lindstrand
    Correspondence
    Correspondence and requests for materials should be addressed to Anna Lindstrand, Department of Clinical Genetics, Karolinska University Hospital, Solna SE-17176, Stockholm, Sweden
    Affiliations
    Department of Molecular Medicine and Surgery, Karolinska Institutet, Stockholm, Sweden

    Department of Clinical Genetics, Karolinska University Hospital, Stockholm, Sweden
    Search for articles by this author
  • Marlene Ek
    Affiliations
    Department of Molecular Medicine and Surgery, Karolinska Institutet, Stockholm, Sweden

    Department of Clinical Genetics, Karolinska University Hospital, Stockholm, Sweden
    Search for articles by this author
  • Malin Kvarnung
    Affiliations
    Department of Molecular Medicine and Surgery, Karolinska Institutet, Stockholm, Sweden

    Department of Clinical Genetics, Karolinska University Hospital, Stockholm, Sweden
    Search for articles by this author
  • Britt-Marie Anderlid
    Affiliations
    Department of Molecular Medicine and Surgery, Karolinska Institutet, Stockholm, Sweden

    Department of Clinical Genetics, Karolinska University Hospital, Stockholm, Sweden
    Search for articles by this author
  • Erik Björck
    Affiliations
    Department of Molecular Medicine and Surgery, Karolinska Institutet, Stockholm, Sweden

    Department of Clinical Genetics, Karolinska University Hospital, Stockholm, Sweden
    Search for articles by this author
  • Jonas Carlsten
    Affiliations
    Department of Molecular Medicine and Surgery, Karolinska Institutet, Stockholm, Sweden

    Department of Clinical Genetics, Karolinska University Hospital, Stockholm, Sweden
    Search for articles by this author
  • Jesper Eisfeldt
    Affiliations
    Department of Molecular Medicine and Surgery, Karolinska Institutet, Stockholm, Sweden

    Department of Clinical Genetics, Karolinska University Hospital, Stockholm, Sweden

    Science for Life Laboratory, Department of Molecular Medicine and Surgery, Karolinska Institutet, Solna, Sweden
    Search for articles by this author
  • Giedre Grigelioniene
    Affiliations
    Department of Molecular Medicine and Surgery, Karolinska Institutet, Stockholm, Sweden

    Department of Clinical Genetics, Karolinska University Hospital, Stockholm, Sweden
    Search for articles by this author
  • Peter Gustavsson
    Affiliations
    Department of Molecular Medicine and Surgery, Karolinska Institutet, Stockholm, Sweden

    Department of Clinical Genetics, Karolinska University Hospital, Stockholm, Sweden
    Search for articles by this author
  • Anna Hammarsjö
    Affiliations
    Department of Molecular Medicine and Surgery, Karolinska Institutet, Stockholm, Sweden

    Department of Clinical Genetics, Karolinska University Hospital, Stockholm, Sweden
    Search for articles by this author
  • Hafdís T. Helgadóttir
    Affiliations
    Department of Molecular Medicine and Surgery, Karolinska Institutet, Stockholm, Sweden

    Department of Clinical Genetics, Karolinska University Hospital, Stockholm, Sweden
    Search for articles by this author
  • Maritta Hellström-Pigg
    Affiliations
    Department of Molecular Medicine and Surgery, Karolinska Institutet, Stockholm, Sweden

    Department of Clinical Genetics, Karolinska University Hospital, Stockholm, Sweden
    Search for articles by this author
  • Ekaterina Kuchinskaya
    Affiliations
    Department of Molecular Medicine and Surgery, Karolinska Institutet, Stockholm, Sweden

    Department of Clinical Genetics, Karolinska University Hospital, Stockholm, Sweden
    Search for articles by this author
  • Kristina Lagerstedt-Robinson
    Affiliations
    Department of Molecular Medicine and Surgery, Karolinska Institutet, Stockholm, Sweden

    Department of Clinical Genetics, Karolinska University Hospital, Stockholm, Sweden
    Search for articles by this author
  • Lars-Åke Levin
    Affiliations
    Department of Health, Medicine and Caring Sciences, Linköping University, Linköping, Sweden
    Search for articles by this author
  • Agne Lieden
    Affiliations
    Department of Molecular Medicine and Surgery, Karolinska Institutet, Stockholm, Sweden

    Department of Clinical Genetics, Karolinska University Hospital, Stockholm, Sweden
    Search for articles by this author
  • Hillevi Lindelöf
    Affiliations
    Department of Molecular Medicine and Surgery, Karolinska Institutet, Stockholm, Sweden

    Department of Clinical Genetics, Karolinska University Hospital, Stockholm, Sweden
    Search for articles by this author
  • Helena Malmgren
    Affiliations
    Department of Molecular Medicine and Surgery, Karolinska Institutet, Stockholm, Sweden

    Department of Clinical Genetics, Karolinska University Hospital, Stockholm, Sweden
    Search for articles by this author
  • Daniel Nilsson
    Affiliations
    Department of Molecular Medicine and Surgery, Karolinska Institutet, Stockholm, Sweden

    Department of Clinical Genetics, Karolinska University Hospital, Stockholm, Sweden

    Science for Life Laboratory, Department of Molecular Medicine and Surgery, Karolinska Institutet, Solna, Sweden
    Search for articles by this author
  • Eva Svensson
    Affiliations
    Department of Pediatric Neurology, Karolinska University Hospital, Huddinge, Sweden
    Search for articles by this author
  • Martin Paucar
    Affiliations
    Department of Molecular Medicine and Surgery, Karolinska Institutet, Stockholm, Sweden

    Department of Clinical Genetics, Karolinska University Hospital, Stockholm, Sweden
    Search for articles by this author
  • Ellika Sahlin
    Affiliations
    Department of Molecular Medicine and Surgery, Karolinska Institutet, Stockholm, Sweden

    Department of Clinical Genetics, Karolinska University Hospital, Stockholm, Sweden
    Search for articles by this author
  • Bianca Tesi
    Affiliations
    Department of Molecular Medicine and Surgery, Karolinska Institutet, Stockholm, Sweden

    Department of Clinical Genetics, Karolinska University Hospital, Stockholm, Sweden
    Search for articles by this author
  • Emma Tham
    Affiliations
    Department of Molecular Medicine and Surgery, Karolinska Institutet, Stockholm, Sweden

    Department of Clinical Genetics, Karolinska University Hospital, Stockholm, Sweden
    Search for articles by this author
  • Johanna Winberg
    Affiliations
    Department of Molecular Medicine and Surgery, Karolinska Institutet, Stockholm, Sweden

    Department of Clinical Genetics, Karolinska University Hospital, Stockholm, Sweden
    Search for articles by this author
  • Max Winerdal
    Affiliations
    Department of Molecular Medicine and Surgery, Karolinska Institutet, Stockholm, Sweden

    Department of Clinical Genetics, Karolinska University Hospital, Stockholm, Sweden
    Search for articles by this author
  • Josephine Wincent
    Affiliations
    Department of Molecular Medicine and Surgery, Karolinska Institutet, Stockholm, Sweden

    Department of Clinical Genetics, Karolinska University Hospital, Stockholm, Sweden
    Search for articles by this author
  • Maria Johansson Soller
    Affiliations
    Department of Molecular Medicine and Surgery, Karolinska Institutet, Stockholm, Sweden

    Department of Clinical Genetics, Karolinska University Hospital, Stockholm, Sweden
    Search for articles by this author
  • Maria Pettersson
    Affiliations
    Department of Molecular Medicine and Surgery, Karolinska Institutet, Stockholm, Sweden

    Department of Clinical Genetics, Karolinska University Hospital, Stockholm, Sweden
    Search for articles by this author
  • Ann Nordgren
    Affiliations
    Department of Molecular Medicine and Surgery, Karolinska Institutet, Stockholm, Sweden

    Department of Clinical Genetics, Karolinska University Hospital, Stockholm, Sweden
    Search for articles by this author
Open AccessPublished:September 05, 2022DOI:https://doi.org/10.1016/j.gim.2022.07.022

      Abstract

      Purpose

      Individuals with intellectual disability (ID) and/or neurodevelopment disorders (NDDs) are currently investigated with several different approaches in clinical genetic diagnostics.

      Methods

      We compared the results from 3 diagnostic pipelines in patients with ID/NDD: genome sequencing (GS) first (N = 100), GS as a secondary test (N = 129), or chromosomal microarray (CMA) with or without FMR1 analysis (N = 421).

      Results

      The diagnostic yield was 35% (GS-first), 26% (GS as a secondary test), and 11% (CMA/FMR1). Notably, the age of diagnosis was delayed by 1 year when GS was performed as a secondary test and the cost per diagnosed individual was 36% lower with GS first than with CMA/FMR1. Furthermore, 91% of those with a negative result after CMA/FMR1 analysis (338 individuals) have not yet been referred for additional genetic testing and remain undiagnosed.

      Conclusion

      Our findings strongly suggest that genome analysis outperforms other testing strategies and should replace traditional CMA and FMR1 analysis as a first-line genetic test in individuals with ID/NDD. GS is a sensitive, time- and cost-effective method that results in a confirmed molecular diagnosis in 35% of all referred patients.

      Graphical abstract

      Keywords

      Introduction

      Intellectual disability (ID), defined as limitations in both intellectual function and adaptive behavior, affects approximately 1% of the world population.
      • Maulik P.K.
      • Mascarenhas M.N.
      • Mathers C.D.
      • Dua T.
      • Saxena S.
      Prevalence of intellectual disability: a meta-analysis of population-based studies.

      American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders. 5th ed. American Psychiatric Association Publishing.

      • Dragojlovic N.
      • Elliott A.M.
      • Adam S.
      • et al.
      The cost and diagnostic yield of exome sequencing for children with suspected genetic disorders: a benchmarking study.
      Depending on the cognitive ability (as measured by an IQ test), individuals are subdivided into 4 groups: mild (IQ 50-70), moderate (IQ 35-50), severe (IQ 20-35), and profound (IQ < 20).

      Committee to Evaluate the Supplemental Security Income Disability Program for Children with Mental Disorders, Board on the Health of Select Populations, Board on Children, Youth, and Families, Institute of Medicine; Division of Behavioral and Social Sciences and Education National Academies of Sciences, Engineering, and Medicine. Boat TF, Wu JT, eds. Mental Disorders and Disabilities among Low-Income Children. National Academies Press; 2015.

      The etiology of ID is heterogeneous and includes environmental factors, such as congenital infections or hypoxic encephalopathy, and genetic factors. Genetic studies have suggested that among individuals with IQ of <50, up to 70% have a genetic background consisting of monogenic disorders and/or deletion (DEL)/duplication (DUP) syndromes.
      • Vissers L.E.L.M.
      • Gilissen C.
      • Veltman J.A.
      Genetic studies in intellectual disability and related disorders.
      ID cases may also be grouped, depending on the presence of comorbid features, as syndromic (one or more clinical features in addition to ID) and nonsyndromic. Finally, comorbidity for other diagnoses within the neurodevelopment disorders (NDDs), such as autism, attention deficit hyperactivity disorder, and speech development disorders, is common in all ID groups.
      The most common diagnoses in ID are trisomy 21 (Down syndrome; OMIM 190685), 22q11 deletion syndrome (DiGeorge syndrome; OMIM 188400), and Fragile X syndrome (OMIM 300624, caused by a CGG expansion in FMR1). However, these diagnoses only account for a small fraction of ID etiology, which is extremely heterogeneous. Both chromosomal abnormalities and submicroscopic DELs and DUPs as well as pathogenic variants in >1000 genes
      • Martin A.R.
      • Williams E.
      • Foulger R.E.
      • et al.
      PanelApp crowdsources expert knowledge to establish consensus diagnostic gene panels.
      have been linked to ID. Most of the pathogenic variants are single nucleotide variants (SNVs) or insertion/DELs (INDELs), but other variant types also play an important role. In addition to FMR1, a number of other repeat expansion loci have been associated with ID.
      • van der Sanden B.P.G.H.
      • Corominas J.
      • de Groot M.
      • et al.
      Systematic analysis of short tandem repeats in 38,095 exomes provides an additional diagnostic yield.
      Finally, both simple structural variants (SVs) (translocations, inversions, DELs, and DUPs) as well as complex chromosomal rearrangements (CCRs) may affect single genes and cause monogenic diseases.
      • Plesser Duvdevani M.
      • Pettersson M.
      • Eisfeldt J.
      • et al.
      Whole-genome sequencing reveals complex chromosome rearrangement disrupting NIPBL in infant with Cornelia de Lange syndrome.
      • Hammarsjö A.
      • Pettersson M.
      • Chitayat D.
      • et al.
      High diagnostic yield in skeletal ciliopathies using massively parallel genome sequencing, structural variant screening and RNA analyses.
      • Bramswig N.C.
      • Lüdecke H.J.
      • Pettersson M.
      • et al.
      Identification of new TRIP12 variants and detailed clinical evaluation of individuals with non-syndromic intellectual disability with or without autism.
      • Grigelioniene G.
      • Nevalainen P.I.
      • Reyes M.
      • et al.
      A large inversion involving GNAS exon A/B and all exons encoding Gsα is associated with autosomal dominant pseudohypoparathyroidism type Ib (PHP1 billion).
      As a consequence, the clinical genetic investigation needs to capture many different variant types, often requiring multiple genetic tests.
      As novel genetic screening technologies have emerged, the standardized primary genetic analysis of individuals with ID has evolved to maximize the number of individuals receiving a diagnosis. When G-banded karyotyping was replaced by chromosomal microarray (CMA) analysis the number of identified abnormalities increased dramatically from 3% to 10% to 15%.
      • Miller D.T.
      • Adam M.P.
      • Aradhya S.
      • et al.
      Consensus statement: chromosomal microarray is a first-tier clinical diagnostic test for individuals with developmental disabilities or congenital anomalies.
      At present, it is becoming apparent that to capture the high number of monogenic causes of ID, a state-of-the-art genetic investigation of individuals with ID has to include an ID gene panel or in silico ID gene panel filtered from exome sequencing (ES) or genome sequencing (GS).
      • Miller D.T.
      • Adam M.P.
      • Aradhya S.
      • et al.
      Consensus statement: chromosomal microarray is a first-tier clinical diagnostic test for individuals with developmental disabilities or congenital anomalies.
      • Beaudet A.L.
      The utility of chromosomal microarray analysis in developmental and behavioral pediatrics.
      • Waggoner D.
      • Wain K.E.
      • Dubuc A.M.
      • et al.
      Yield of additional genetic testing after chromosomal microarray for diagnosis of neurodevelopmental disability and congenital anomalies: a clinical practice resource of the American College of Medical Genetics and Genomics (ACMG).
      In fact, exomes are increasingly being used as a tier 1 test in ID. However, local policies differ and even in western countries (such as Sweden), multiple strategies for genetic investigations of ID are still in use. In many centers, the specific tests performed in each case are decided from the referral and both CMAs and testing for common single-gene disorders such as Fragile X are still frequently requested.
      GS is increasingly applied in genetic diagnostics with impressive results and has the ability to capture a broad range of different genetic variants.
      • Stranneheim H.
      • Lagerstedt-Robinson K.
      • Magnusson M.
      • et al.
      Integration of whole genome sequencing into a healthcare setting: high diagnostic rates across multiple clinical entities in 3219 rare disease patients.
      • Lindstrand A.
      • Eisfeldt J.
      • Pettersson M.
      • et al.
      From cytogenetics to cytogenomics: whole-genome sequencing as a first-line test comprehensively captures the diverse spectrum of disease-causing genetic variation underlying intellectual disability.
      • Smedley D.
      • Smith K.R.
      • et al.
      100,000 Genomes Project Pilot Investigators
      100,000 Genomes pilot on rare-disease diagnosis in health care – preliminary report.
      A comprehensive GS analysis may be used to screen for SNVs/INDELs,
      • McKenna A.
      • Hanna M.
      • Banks E.
      • et al.
      The Genome Analysis Toolkit: a MapReduce framework for analyzing next-generation DNA sequencing data.
      SVs,
      • Abyzov A.
      • Urban A.E.
      • Snyder M.
      • Gerstein M.
      CNVnator: an approach to discover, genotype, and characterize typical and atypical CNVs from family and population genome sequencing.
      • Chen X.
      • Schulz-Trieglaff O.
      • Shaw R.
      • et al.
      Manta: rapid detection of structural variants and indels for germline and cancer sequencing applications.
      • Eisfeldt J.
      • Vezzi F.
      • Olason P.
      • Nilsson D.
      • Lindstrand A.
      TIDDIT, an efficient and comprehensive structural variant caller for massive parallel sequencing data.
      short tandem repeats (STRs),
      • Dolzhenko E.
      • Deshpande V.
      • Schlesinger F.
      • et al.
      ExpansionHunter: a sequence-graph-based tool to analyze variation in short tandem repeat regions.
      SMN1/SMN2 copy number,
      • Chen X.
      • Sanchis-Juan A.
      • French C.E.
      • et al.
      Spinal muscular atrophy diagnosis and carrier screening from genome sequencing data.
      and loss of heterozygosity.
      • Boeva V.
      • Popova T.
      • Bleakley K.
      • et al.
      Control-FREEC: a tool for assessing copy number and allelic content using next-generation sequencing data.
      Ultimately, to maximize the diagnostic yield, all those variant types need not only be called but also be assessed, highlighting the demands for advanced data analysis as well as advanced clinical interpretation.
      In this article, we summarize the genetic results from 624 individuals with ID, investigated as singletons at the Karolinska University Hospital (Stockholm, Sweden) during 2020 and 2021 and compare the outcome of the different testing strategies. Altogether, our results show that GS as the first-line genetic analysis is feasible and effective, providing a molecular diagnosis for one-third individuals with ID. Even though GS as a secondary test captures a similar number of causative variants, this approach is not cost-effective, the time to diagnosis is delayed, and patients are lost to follow up.

      Materials and Methods

      Study subjects

      In this retrospective study, included individuals were referred for diagnostic genetic testing at the Department of Clinical Genetics, Karolinska University Hospital, Stockholm, Sweden, from February 1st, 2020 to March 4th, 2021. In total, 229 unrelated individuals with an ID diagnosis or a strong clinical suspicion of ID underwent genetic testing using GS. In 100 individuals, GS was the first-line genetic analysis (cohort 1) and in 129 individuals, GS was the secondary/tertiary genetic test, most commonly after CMA/FMR1 testing when that analysis could not find a cause for the clinical phenotype (cohort 2). Finally, 421 individuals were analyzed using CMA (of which 212 [50%] also had FMR1 expansion testing performed) (cohort 3). Common comorbidities in all cases were dysmorphic features, congenital malformations, and epilepsy (Table 1).
      Table 1Clinical parameters of included cases
      Cohort CharacteristicsCohort 1 (N = 100)Cohort 2 (N = 129)Cohort 3 (N = 421)
      Median age, y676
      Male, n (%)67 (67)84 (65)292 (69)
      Consanguinity, n (%)12 (12)9 (7)7 (2)
      Main phenotypic features
       ID, n (%)97 (97)125 (97)309 (73)
       Autism, n (%)39 (39)61 (47)242 (57)
       ADHD, n (%)14 (14)20 (16)69 (16)
       Epilepsy, n (%)8 (8)16 (12)20 (5)
       Malformations, n (%)5 (5)13 (10)18 (4)
      Genetic tests performed (within 13 months from referral)
       CMA, n (%)0124 (96)421 (100)
      FMR1 STR, n (%)055 (44)212 (50)
       GS, n (%)100 (100)129 (100)37 (9)
       Parental follow up, n (%)29 (29)29 (22)36 (9)
      ADHD, attention deficit hyperactivity disorder; CMA, chromosomal microarray; GS, genome sequencing; ID, intellectual disability; STR, short tandem repeat.
      All cases were analyzed as singletons. Follow-up analyses of parental samples were performed in 29 (29%) (cohort 1), 29 (22%) (cohort 2), and 36 (9%) (cohort 3) cases (Figure 1; Supplemental Tables 1, 2, 3, and 4).
      Figure thumbnail gr1
      Figure 1Study overview. Flowchart showing results and samples analyzed in the 3 cohorts before (top) and after parental follow-up analysis (bottom). Pathogenic variants refer to those classified as American College of Medical Genetics and Genomics class 4 and class 5. CCR, complex chromosomal rearrangement; CNV, copy number variation; het, heterozygous; transl, translocation; unbal, unbalanced; VUS, variant of uncertain significance.

      GS

      The detailed workflow for clinical GS at our center was described previously.
      • Stranneheim H.
      • Lagerstedt-Robinson K.
      • Magnusson M.
      • et al.
      Integration of whole genome sequencing into a healthcare setting: high diagnostic rates across multiple clinical entities in 3219 rare disease patients.
      ,
      • Lindstrand A.
      • Eisfeldt J.
      • Pettersson M.
      • et al.
      From cytogenetics to cytogenomics: whole-genome sequencing as a first-line test comprehensively captures the diverse spectrum of disease-causing genetic variation underlying intellectual disability.
      For all analysis, mapping to GRCh37 (hg19) was used. In brief, large copy number variations (CNVs) were assessed using the vcf2cytosure pipeline in which detected SVs were visualized in the CytoSure Interpret Software (Oxford Gene Technology).
      • Lindstrand A.
      • Eisfeldt J.
      • Pettersson M.
      • et al.
      From cytogenetics to cytogenomics: whole-genome sequencing as a first-line test comprehensively captures the diverse spectrum of disease-causing genetic variation underlying intellectual disability.
      SNVs/INDELs and intragenic DELs/DUPs were filtered using an ID gene panel including 1097 genes (Supplemental Document 1), and only coding variants and variants theoretically affecting splicing were considered. STRs were assessed in 7 genes (ATN1, CNBP, CSTB, DIP2B, FMR1, GLS, and DMPK). Remaining variants were ranked and assessed in our in-house developed analysis tool Scout,
      • Stranneheim H.
      • Lagerstedt-Robinson K.
      • Magnusson M.
      • et al.
      Integration of whole genome sequencing into a healthcare setting: high diagnostic rates across multiple clinical entities in 3219 rare disease patients.
      and variants with a high ranking that matched the clinical phenotype were classified according to American College of Medical Genetics and Genomics (ACMG)/Association for Molecular Pathology (AMP) guidelines.
      • Richards S.
      • Aziz N.
      • Bale S.
      • et al.
      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.
      Pathogenic and likely pathogenic variants were reported, ie, the variant was listed in the GS report. Variants of uncertain significance (VUS) were reported when the likelihood of them being pathogenic was high according to local praxis—depending on both variant characteristics (ie, not present in normal (NML) variation databases, high local ranking, compound variant present for autosomal recessive genes) and overlap between the phenotype described for the affected gene and the clinical symptoms present in the affected individuals. In some VUS cases, parental samples were requested to determine de novo status.
      The final diagnostic yield solely included variants scored as ACMG/AMP class 4 and 5. Class 3 variants that in combination with inheritance pattern and clinical phenotype of the patient (ID/NDD) rendered a strong suspicion of pathogenicity were considered as clinically relevant findings, but were not part of the reported overall yield.

      CMA

      The array experiments were performed according to the manufacturer’s protocol with minor modifications
      • Lindstrand A.
      • Eisfeldt J.
      • Pettersson M.
      • et al.
      From cytogenetics to cytogenomics: whole-genome sequencing as a first-line test comprehensively captures the diverse spectrum of disease-causing genetic variation underlying intellectual disability.
      using a 4 × 180,000 custom oligonucleotide microarray with even genome coverage and median probe spacing of approximately 18 kilobases (kb) (AMADID:031035, Oxford Gene Technology).

      FMR1 STR analysis

      FMR1 CGG repeat expansion analysis was performed according to the manufacturer’s protocol, using the AmplideX PCR/CE FMR1 Kit (Asuragen) and an ABI 3500xL Genetic Analyzer (Applied Biosystems).

      Fluorescence in situ hybridization

      Metaphase fluorescence in situ hybridization (FISH) follow up of parental samples was performed in individuals C3-P152 and C3-P348 using standardized protocols. For individual C3-P348, commercially available probes TelVysion 4p SpectrumGreen and Vysis CEP4 SpectrumAqua (Abbot) were used. For individual C3-P152, Vysis CEP18 SpectrumAqua (Abbot) combined with BAC/PAC probes were used and performed as previously described.
      • Lindstrand A.
      • Malmgren H.
      • Verri A.
      • et al.
      Molecular and clinical characterization of patients with overlapping 10p deletions.
      At least 10 metaphases per individual were analyzed.

      Statistics and cost calculations

      Descriptive statistics for age and turnaround time (TAT) in cohort 1, cohort 2, and cohort 3, and diagnostic yield between cohort 1 and cohort 3 are reported. Analysis of variance or Kruskal-Wallis nonparametric test and Wilcoxon rank-sum test were used to calculate differences in age and TAT and χ2 with Yates correction for differences in yield between cohort 1 and cohort 3. Statistical significance was set to P = .05. The χ2 tests were performed in Microsoft Excel, whereas analysis of variance or Kruskal-Wallis nonparametric test and Wilcoxon rank-sum test were performed in R Studio software (version 4.0.2).
      Costs for genetic analysis were calculated from the 2021 list prices at Clinical Genetics Karolinska University Hospital and converted from Swedish Krona to US Dollars using the average exchange rate of 2020 (9.2037 Swedish Krona = 1 US Dollars).
      The Swedish Riksbank. Search interest rates & exchange rates. The Swedish Riksbank.

      Results

      Detected genetic variants

      The combined diagnostic yield for all analyzed individuals in all 3 cohorts was 21% (130/624), and the per cohort yield was 35% for cohort 1, 26% for cohort 2, and 11% for cohort 3. Focusing on the 521 individuals who underwent primary testing either with GS-first (cohort 1, N = 100) or CMA/FMR1 analysis (cohort 3, N = 421), we found that the diagnostic yield was significantly higher with GS first (35% vs 11%, P < .001). An overview of the types of detected genetic variants and their inheritance patterns for all 3 cohorts are presented in Table 2 and Figure 2. Details for each cohort are given in the following, and all reported variants are listed in Supplemental Tables 1, 2, 3, and 4.
      Table 2Variant types, ACMG/AMP classification,
      • Richards S.
      • Aziz N.
      • Bale S.
      • et al.
      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.
      and inheritance patterns in the 3 cohorts after follow-up investigations
      Summary of Genetic ResultsCohort 1Cohort 2Cohort 3
      Total number of individuals with variants
      Genetic variant determined to be of clinical significance at the end of the study period.
      444651
      Total number of variants494851
      Variant types (% of total number of variants in the cohort)
      SNV/INDEL34 (71)46 (96)NA
       ACMG/AMP class 5714NA
       ACMG/AMP class 41418NA
       ACMG/AMP class 31314NA
      SV8 (17)1 (2)41 (82)
       Deletion6126
       Duplication2013
       Complex002
      Chromosomal
      Aneuploidy, XY female, unbalanced translocation.
      5 (8)0 (0)8 (16)
      UPD1 (2)0 (0)NA
      STR1 (2)1 (2)2 (4)
      Per variant inheritance pattern (% of total number of variants in the cohort)
      Autosomal dominant28 (57)23 (48)41 (80)
      Autosomal recessive homozygous7 (14)11 (23)0 (0)
      Autosomal recessive compound heterozygous4 (8)2 (4)0 (0)
      X-linked5 (10)12 (25)3 (6)
      Chromosomal
      Aneuploidy, XY female, unbalanced translocation.
      5 (10)0 (0)7 (14)
      ACMG, American College of Medical Genetics and Genomics; AMP, Association for Molecular Pathology; INDEL, insertion/deletion; NA, not applicable; SNV, single nucleotide variant; STR, short tandem repeat; SV, structural variant; UPD, uniparental disomy.
      a Genetic variant determined to be of clinical significance at the end of the study period.
      b Aneuploidy, XY female, unbalanced translocation.
      Figure thumbnail gr2
      Figure 2Detected genetic variants in the 3 cohorts. Variant types among the communicated findings in cohort 1 (A), cohort 2 (B), and cohort 3 (C). ANEU, aneuploidy; CCR, complex chromosomal rearrangement; INDEL, insertion/deletion; SNV, single nucleotide variant; STR, short tandem repeat; SV, structural variant; UPD, uniparental disomy; VUS, variant of uncertain significance.

      Cohort 1

      In the 100 unrelated individuals analyzed with GS first, 47 had at least 1 genetic variant reported, encompassing a total of 54 variants subdivided into SNV/INDELs (n = 39), SVs (n = 8), chromosomal (n = 5), uniparental disomy (UPD) (n = 1), and STR (n = 1) (Figure 1; Supplemental Table 1). The 8 detected SVs included 2 DELs (C1-P61, C1-P64, Supplemental Table 1) that were sized below or at borderline of the detection limit of the CMA used at our unit (3.2 kb and 39.2 kb respectively). In total, 4 aneuploidies, 2 47,XXY (Klinefelter syndrome), 1 47,XYY, and 1 mosaic 45,X/46,XX (estimated ratio 80/20) (Turner syndrome) were detected. In individual C1-P32, a 4-year-old girl, the GS analysis uncovered both an aberrant 46,XY karyotype and an SNV of uncertain significance (VUS) in AR, supporting the diagnosis androgen insensitivity (XY female; OMIM 300068), however, this does not explain the presence of developmental delay.
      Of the total 54 variants (SNV/INDELs [n = 39], SVs [n = 8], chromosomal [n = 5], UPD [n = 1] and STR [n = 1]) detected in 47 individuals, 33 were initially classified as pathogenic/likely pathogenic and 21 as VUS: 20 SNVs and 1 SV (a 2.8 megabase 10q26 DUP, C1-P101) (Supplemental Table 1). Parental samples were received for 29 individuals, of which 14 were VUS cases, and follow-up investigations resulted in a reclassification of 8 VUS variants: 5 were downgraded in pathogenicity from VUS to likely benign (inherited from a healthy parent) and 3 were upgraded from VUS to pathogenic (de novo), including the aforementioned 10q26 DUP. Parental follow up of pathogenic/likely pathogenic cases did not lead to any reclassification. After follow up, of the initial 54 genetic variants, 36 were classified as pathogenic/likely pathogenic (ACMG/AMP class 4 or 5), 5 as likely benign (ACMG/AMP class 2), and 13 remained as VUS (ACMG/AMP class 3) (Figure 1; Supplemental Table 1). However, 2 of these VUS were compound heterozygotes with a pathogenic/likely pathogenic variant in recessive genes that explain the disease phenotype of the patients, and these 2 patients were therefore included in the overall diagnostic yield (hyperlysinemia, OMIM 238700 [C1-P17] and achalasia-addisonianism-alacrimia syndrome, OMIM 231550 [C1-P39]). In contrast, for the 2 class 3 SNVs in MACF1 (C1-P2) and IRF2BPL (C1-P41) that were inherited from mothers with similar clinical presentations, pathogenicity cannot be confirmed, and they remain as VUS.
      Two individuals had dual findings, both from consanguineous pedigrees, and included 1 case with a homozygous INDEL in AP4S1 (spastic paraplegia 52, OMIM 614067) and an STR in CNBP (myotonic dystrophy 2, OMIM 602668) (C1-P81). Interestingly, in this family, the brother of C1-P81 (individual C1-P1) had been referred for GS independently, and only the brother with a more severe phenotype, including ID, harbored the AP4S1 variant. In the second case, 2 homozygous rare missense variants in WIPI2 and BRAT1 were detected, and it was not possible from the clinical presentation to rule out either of the genes (C1-P16).
      Altogether, after follow up, excluding the XY female, 35 individuals harbored pathogenic or likely pathogenic variants that were deemed as causative of ID and in 10 individuals, VUS were reported (Figure 2; Table 2). Hence, the overall diagnostic yield, explaining the primary symptom of ID, in 100 individuals analyzed with GS first was 35%.

      Cohort 2

      In the 129 individuals analyzed with GS as a secondary analysis, 50 individuals (39%) had at least 1 genetic variant reported, encompassing a total of 53 variants subdivided as SNV/INDELs (n = 50) and SVs (n = 1), chromosomal (n = 1), and STRs (n = 1) (Figure 1; Supplemental Table 2). In total, 32 were initially classified as pathogenic and 21 as VUS. The VUS were all SNVs (Supplemental Table 2). Parental samples were available for 29 individuals, of which 12 were VUS cases, and follow up investigations resulted in a reclassification of 6 VUS variants: 3 were downgraded from VUS to likely benign (inherited from a healthy parent) and 3 were upgraded from VUS to pathogenic (2 de novo and 1 maternally inherited SNV on the X-chromosome that was de novo in the mother). After this follow up, a total of 35 variants were classified as pathogenic/likely pathogenic (ACMG/AMP class 4 or 5), 4 were classified as likely benign (ACMG/AMP class 2) and 14 VUS remained uncertain. Parental follow up of pathogenic/ likely pathogenic variants did not lead to any reclassification (Table 2; Supplemental Table 2).
      The single SV detected in cohort 2 was, as expected, below the detection size-limit of CMA (22.5 kb, C2-P59, Supplemental Table 2). The detected STR was a pathogenic FMR1 CGG expansion in a 4-year-old boy (C2-P108) who had not undergone targeted FMR1 STR analysis before GS. A single aneuploidy was detected in cohort 2, which was known before the GS analysis (47,XYY, C2-P39/C3-P62, Supplemental Table 2). After disregarding the previously known aneuploidy, 34 individuals with pathogenic findings remained (Figure 1; Table 2), resulting in an added diagnostic yield of 26% (34/129) through GS with ID panel after CMA with or without FMR1 STR analysis.

      Cohort 3

      In 53 of the 421 individuals in cohort 3 (13%), at least 1 genetic variant was reported after CMA analysis, including 28 DELs, 16 DUPs, 1 unbalanced translocation, 2 CCRs, and 7 aneuploidies (Figure 1; Supplemental Table 3). Parental samples were received for 36 cases. A total of 5 variants in 4 individuals, were downgraded from VUS to likely benign after the parental investigation (2 DELs and 3 DUPs inherited from healthy parents). The remaining variants were classified as pathogenic/likely pathogenic (n = 45) and 4 VUS with a high suspicion of being causative, the latter including 2 individuals (C3-P222 and C3-P283) with the 15q11.2 BP1-BP2 DEL with reduced penetrance was detected in both cases inherited from a healthy father
      • Burnside R.D.
      • Pasion R.
      • Mikhail F.M.
      • et al.
      Microdeletion/microduplication of proximal 15q11.2 between BP1 and BP2: a susceptibility region for neurological dysfunction including developmental and language delay.
      ,
      • Jønch A.E.
      • Douard E.
      • Moreau C.
      • et al.
      Estimating the effect size of the 15Q11.2 BP1-BP2 deletion and its contribution to neurodevelopmental symptoms: recommendations for practice.
      (Supplemental Table 3). One unbalanced translocation between chromosomes 4 and 6 (C3-P108) and 2 complex chromosome rearrangements (CCRs) were detected.
      In the 212 individuals who had undergone FMR1 analysis, a pathogenic repeat expansion was detected in 2 individuals (2/212, 1%), giving a total diagnostic yield of 11% (47/421) in cohort 3 (Figure 2; Table 2).

      CCRs

      Disregarding the unbalanced translocation detected in case C3-P108, 2 CCRs were detected (both in cohort 3). In the first individual (C3-P152), the CMA analysis revealed the presence of 3 CNVs on chromosome 18q, 2 DUPs, and 1 DEL separated by NML genomic segments in the following pattern: DUP-NML-DEL-NML-DUP. The 18q CCR was shown to be de novo after NML parental CMA and FISH analysis. As part of the clinical follow-up investigations, GS was performed in patient C3-P152, and the data were used to resolve the derivative chromosome structure. The high-resolution characterization revealed a 10.9 Mb tandem DUP (segment A) followed by a 135 kb DEL (segment C) and a second 2.2 Mb DUP (segment E) that was inserted in an inverted orientation 699 kb upstream of the original segment E and 757 kb downstream of DUP A. Furthermore, a 13 kb NML copy number segment was present 250 kb from the centromeric end of the duplicated segment A, likely representing a DEL in DUP A. Breakpoint junction (BPJ) analysis of the 4 BPJs revealed a 5 nucleotide (nt) insertion (GCCAT) in the tandem DUP BPJ (A-A), a 10 nt insertion containing the same 5 nucleotides (GCCATAGATA) in the DEL A BPJ and for the inverted DUP, a 33 nt insertion was present in 1 BPJ, and the other showed blunt ends. (Figure 3A).
      Figure thumbnail gr3
      Figure 3Overview of identified complex chromosomal rearrangements (CCRs). A. The chromosome 18 CCR detected in patient C3-P152. On top, a schematic illustration of chromosome 18 with the rearranged region marked with a blue rectangle. Below, the array plot is shown, showing the duplicated and deleted genomic segments. Further down, the aberrant genomic segments are outlined on the chr18 reference chromosome (duplication A in red, deletion C in green, and duplication E in gradient red/yellow). A subway plot (in green) shows how each region is connected in the short read genome sequencing data and allows for reconstruction of derivative 18 structure shown below with the same color code as above and the inverted segment marked with an arrow. An arrow in the array plot marks the copy number neutral probe, suggesting a deletion within segment A, also marked with ∗ in the subway plot. At the bottom, screen shots from the integrative genomics viewer and aligned BPJ sequences are shown for each junction. B. Overview of complex 3q-4q rearrangement detected in individual C3-P348. The array data, fluorescence in situ hybridization–analysis validation images, and a schematic drawing of the CCR is shown. BPJ, breakpoint junction; Chr, chromosome; der, derivative; Mb, megabase.
      The second detected CCR was a translocated DUP, in which 5.6 Mb from 3q26.1q26.2 was duplicated and translocated to terminal chromosome 4q (4q35.1) where 4.5 Mb was deleted (C3-P348). The rearrangement was first detected using CMA, and follow up using CMA in the parents revealed that the variant was inherited from the father (with similar clinical symptoms). FISH analysis showed that the segment was inserted into chromosome 4 (Figure 3B).

      Additional genetic tests ordered in cohort 3

      During the 13 months this study was conducted, 38 of the 370 (10%) individuals negative after CMA/FMR1 analysis were referred for further investigation using GS. Of those, 26 individuals underwent ID gene panel analysis and are therefore also part of cohort 2 (patients C2-14, 16, 26, 32, 38, 39, 50, 58, 61, 62, 73, 88, 91, 92, 97, 98, 100, 101, 104, 106, 107, 115-117, 119, and 121), and pathogenic variants were found in 9 of them (Supplemental Table 2). The remaining 12 individuals were investigated using different in silico gene panels, and pathogenic SNVs were detected in 8 of those 12 individuals (Supplemental Table 4).

      Age distribution, TAT, and cost calculations

      The median age was 6 years (range 0-38 years) in cohort 1, 7 years (range 0-39 years) in cohort 2, and 6 years (range 0-62 years) in cohort 3. The mean age followed a similar pattern with cohort 3 being the lowest (7.6 years) followed by cohort 1 (8.4 years) and cohort 2 (8.9 years). These differences were not statistically significant except that a higher age distribution was seen in cohort 3 than in cohort 2 (Kruskal-Wallis non-parametric test and post hoc Wilcoxon rank-sum tests; P = .033) (Supplemental Figure 1; Supplemental Table 5).
      The TAT was higher in cohort 1 and cohort 2 than in cohort 3 (median 50 [range 17-195], 46 [range 19-581], and 27 [range 4-104] days, respectively). The differences were significant using Kruskal-Wallis nonparametric test (P < 2.2 × 1016), and post hoc Wilcoxon rank-sum tests revealed a significantly lower TAT in cohort 3 than in cohort 1 (P < 2.2 x 1016) and 2 (P < 2.2 x 1016) but no significant difference between cohort 1 and cohort 2 (P = .561) (Supplemental Figure 1; Supplemental Table 6).
      For the 521 individuals in cohort 1 and cohort 3 who underwent primary testing, the average cost per analyzed individual, including both patient specific analysis and parental follow-up analysis, was 95% higher with GS first ($4505 vs $2315 for cohort 1 and cohort 3, respectively) (Supplemental Table 7). However, the higher diagnostic yield resulted in a cost per diagnosed patient that was 38% lower ($12,872 and $20,737 for cohort 1 and cohort 3, respectively). In consequence, the cost per each additional diagnosis made with GS first compared with CMA/FMR1 analysis was $9124. Furthermore, the mean cost for parental follow-up analysis was lower with GS first ($1120 in cohort 1 compared with $1742 in cohort 3).

      Discussion

      Our results clearly show that GS is an attractive and feasible option as a first-tier test for individuals with ID/NDD. In this retrospective analysis, we show that a singleton GS approach, with in silico filtering for a broad ID/NDD gene panel, results in 3 times as many patients diagnosed compared with the traditional testing strategies. The diagnostic yield for individuals with ID/NDD tested at our unit with 3 different genetic testing approaches during a 13-month period was 35% with GS first, 26% with GS as a secondary analysis, and 11% using only CMA/FMR1 analysis.
      The standardized genetic analysis for individuals with ID and/or NDD, without a specific clinical diagnosis, has been CMA for more than a decade in our laboratory, often in combination with FMR1 STR analysis, with clinically relevant genetic findings in 10% to 15% of the cases.
      • Miller D.T.
      • Adam M.P.
      • Aradhya S.
      • et al.
      Consensus statement: chromosomal microarray is a first-tier clinical diagnostic test for individuals with developmental disabilities or congenital anomalies.
      The detected yield in cohort 3 (11%, 47 pathogenic/likely pathogenic variants detected in 421 individuals) is in concordance with previously published papers, including a recent paper from our own laboratory in which CMA detected a pathogenic DEL or DUP in 12% of individuals.
      • Lindstrand A.
      • Eisfeldt J.
      • Pettersson M.
      • et al.
      From cytogenetics to cytogenomics: whole-genome sequencing as a first-line test comprehensively captures the diverse spectrum of disease-causing genetic variation underlying intellectual disability.
      Hence, the test performs robustly but is limited to the detection of some variant types, ie, DELs and DUPs >50 kb in size and a CGG repeat expansion in a single gene, FMR1. In contrast, GS may detect a multitude of variants of different types and sizes.
      • Lindstrand A.
      • Eisfeldt J.
      • Pettersson M.
      • et al.
      From cytogenetics to cytogenomics: whole-genome sequencing as a first-line test comprehensively captures the diverse spectrum of disease-causing genetic variation underlying intellectual disability.
      Because >1000 genes have been implicated in monogenic ID and NDDs, both isolated and syndromic,
      • Lindstrand A.
      • Eisfeldt J.
      • Pettersson M.
      • et al.
      From cytogenetics to cytogenomics: whole-genome sequencing as a first-line test comprehensively captures the diverse spectrum of disease-causing genetic variation underlying intellectual disability.
      ,
      • Gilissen C.
      • Hehir-Kwa J.Y.
      • Thung D.T.
      • et al.
      Genome sequencing identifies major causes of severe intellectual disability.
      ,
      • Martínez F.
      • Caro-Llopis A.
      • Roselló M.
      • et al.
      High diagnostic yield of syndromic intellectual disability by targeted next-generation sequencing.
      it is imperative that analysis of SNVs and INDELs in a large gene panel is part of a genetic investigation for individuals with ID. Of note, ES is an attractive alternative with an average diagnostic yield of approximately 26.5% for singletons and approximately 34.3% for trio analysis in unselected rare disease pediatric cases
      • Dragojlovic N.
      • Elliott A.M.
      • Adam S.
      • et al.
      The cost and diagnostic yield of exome sequencing for children with suspected genetic disorders: a benchmarking study.
      and an ES first approach has been proposed as a suitable first-line test for individuals with ID/NDD.
      • Srivastava S.
      • Love-Nichols J.A.
      • Dies K.A.
      • et al.
      Meta-analysis and multidisciplinary consensus statement: exome sequencing is a first-tier clinical diagnostic test for individuals with neurodevelopmental disorders.
      At present, many genetic diagnostic laboratories are transitioning to ES first and extracting CNV information (>50 kb) from the ES data.
      • Gordeeva V.
      • Sharova E.
      • Babalyan K.
      • Sultanov R.
      • Govorun V.M.
      • Arapidi G.
      Benchmarking germline CNV calling tools from exome sequencing data.
      However, as we present in this article, many different genetic investigation strategies are currently being applied sometimes even within the same laboratory. This leads to confusion for the referring physicians, affected individuals, and their families.
      At our unit, singleton GS is the main mucopolysaccharidoses-based genetic test for rare diseases
      • Stranneheim H.
      • Lagerstedt-Robinson K.
      • Magnusson M.
      • et al.
      Integration of whole genome sequencing into a healthcare setting: high diagnostic rates across multiple clinical entities in 3219 rare disease patients.
      and, similar to many other laboratories using ES and GS analysis in the clinic, the GS data are filtered in silico vs clinically relevant gene panels on the basis of the patient’s phenotype to expediate the interpretation and minimize incidental findings. With this approach, pathogenic genetic variants in genes not included in the panel will be missed. Even though the gene panel is updated regularly (4 times per year), a risk of missing newly reported monogenic disorders remains. As the newly reported genes become increasingly rare, this risk will decrease but still reanalysis of negative cases is warranted at regular intervals. To partly overcome this limitation, we systematically apply patient’s phenotype specific Human Phenotype Ontology panels
      • Köhler S.
      • Doelken S.C.
      • Mungall C.J.
      • et al.
      The Human Phenotype Ontology project: linking molecular biology and disease through phenotype data.
      in cases in which additional clinical symptoms are present and no cause is detected with the ID panel. In our previous paper mentioned earlier, 100 individuals referred for CMA were also analyzed using GS, and the diagnostic yield increased 4% when Human Phenotype Ontology panels were added.
      • Lindstrand A.
      • Eisfeldt J.
      • Pettersson M.
      • et al.
      From cytogenetics to cytogenomics: whole-genome sequencing as a first-line test comprehensively captures the diverse spectrum of disease-causing genetic variation underlying intellectual disability.
      The higher diagnostic yield with GS first in the current study compared with the former (35% vs 27%) may be partially explained by the use of a larger gene panel (1097 genes vs 887 genes) as well as improved bioinformatic workflows
      • Stranneheim H.
      • Lagerstedt-Robinson K.
      • Magnusson M.
      • et al.
      Integration of whole genome sequencing into a healthcare setting: high diagnostic rates across multiple clinical entities in 3219 rare disease patients.
      and databases.
      • Magnusson M.
      • Eisfeldt J.
      • Nilsson D.
      • et al.
      Loqusdb: added value of an observations database of local genomic variation.
      However, the main reason for a lower yield in the previous study is that those 100 cases were consecutive unselected individuals referred for CMA whereas the current study is focused on individuals with ID/NDD.
      The singleton GS first approach used in this study also runs the risk of missing de novo variants. In cohort 1, parental analysis was performed in 26 cases and of those, 14 variants were de novo (54%) of which, 3 were first classified as VUS. This de novo fraction is lower than in a recent study from Finland using a trio-ES approach reporting a de novo rate of 75%,
      • Järvelä I.
      • Määttä T.
      • Acharya A.
      • et al.
      Exome sequencing reveals predominantly de novo variants in disorders with intellectual disability (ID) in the founder population of Finland.
      providing further support that a higher diagnostic yield could be obtained with trios.
      Of note, most variants detected using GS are located within the exome.
      • Stranneheim H.
      • Lagerstedt-Robinson K.
      • Magnusson M.
      • et al.
      Integration of whole genome sequencing into a healthcare setting: high diagnostic rates across multiple clinical entities in 3219 rare disease patients.
      Hence, the vast majority of the detected pathogenic SNVs and INDELs would also have been found by ES. The main advantage of GS over ES is the aforementioned ability to broadly capture many different types of genetic variants the most important being SVs and STRs. This means that after a negative GS, no additional tests are necessary whereas after ES, the negative 73.5%
      • Dragojlovic N.
      • Elliott A.M.
      • Adam S.
      • et al.
      The cost and diagnostic yield of exome sequencing for children with suspected genetic disorders: a benchmarking study.
      should be complemented with at least CMA and FMR1 analysis for individuals with ID/NDD.
      • Srivastava S.
      • Love-Nichols J.A.
      • Dies K.A.
      • et al.
      Meta-analysis and multidisciplinary consensus statement: exome sequencing is a first-tier clinical diagnostic test for individuals with neurodevelopmental disorders.
      Even so, the lower resolution of SV analysis through CMA will result in some variants being missed. In this study, 3 such small SVs were detected in the 229 individuals analyzed using GS, increasing the diagnostic yield by 1.3% (Supplemental Tables 1 and 2). Two of those small SVs were detected in cohort 1, a 3.2 kb MED13L DEL (C1-P61) that was below the detection limit of our standard CMA and a 39.2 kb MYO5A DEL (C1-P64) only detectable on CMA platforms with higher resolution (Supplemental Table 1). The third SV was detected in cohort 2 and was, as expected, below the approximately 50 kb detection limit of CMA in size (22.5 kb, C2-P59, Supplemental Table 2). A more important problem is that the correct complementary targeted tests are not performed in many individuals, such as FMR1 analysis not being requested by the treating physician in individual C2-P108 with an FMR1 CGG expansion (Supplemental Table 2). Furthermore, GS enables expansion testing in many genes, increasing the diagnostic yield by approximately 1% (shown by the CNBP CCTG expansion identified in individuals C1-P1 and C1-P81). Finally, a multistep process also runs the risk of patients being lost to follow up, well-illustrated by our results in which 90% (n = 332) of the individuals in cohort 3 who were negative after CMA/FMR1 analysis had not yet received a referral for additional genetic testing 13 months after the initial referral.
      Implementing both clinical GS and ES services requires highly specialized components, including access to sequencing platforms and expertise in bioinformatics and variant interpretation.
      • Stranneheim H.
      • Lagerstedt-Robinson K.
      • Magnusson M.
      • et al.
      Integration of whole genome sequencing into a healthcare setting: high diagnostic rates across multiple clinical entities in 3219 rare disease patients.
      At our unit, variant interpretation is made by a multidisciplinary expert team, which adds labor costs but likely increases the quality of data analysis.
      • Srivastava S.
      • Love-Nichols J.A.
      • Dies K.A.
      • et al.
      Meta-analysis and multidisciplinary consensus statement: exome sequencing is a first-tier clinical diagnostic test for individuals with neurodevelopmental disorders.
      ,
      • Sabatini L.M.
      • Mathews C.
      • Ptak D.
      • et al.
      Genomic sequencing procedure microcosting analysis and health economic cost-impact analysis: A report of the association for molecular pathology.
      Even though the costs of genetic analysis vary between laboratories and those reported in this article only represent a single region in a single country (the Stockholm region in Sweden), the overall trends will likely be translatable to other laboratories and health care systems. Although the average cost of the GS first approach is nearly twice that of CMA/FMR1 analysis, the higher diagnostic yield of GS first subsequently leads to a cost per diagnosed patient that is 36% lower than with CMA/FMR1 analysis only. We also observed a lower mean cost for parental follow-up analysis for the GS first cohort, likely explained by the fact that GS often is followed up with polymerase chain reaction and Sanger sequencing whereas after CMA, targeted CMA or FISH is commonly performed. However, because more VUS are observed with GS than with CMA/FMR1, the overall number of variants warranting parental testing is higher with GS first. Hence, an overall higher cost for parental follow-up testing is expected, and it is of importance to limit the number of VUS reported.
      A higher yield as well as detection of more diverse variant types was observed in cohort 1 than in cohort 2 although both cohorts were subjected to GS. This was expected, because DELs, DUPs, and FMR1 expansions had been excluded before GS in cohort 2. The added clinical value of GS after CMA/FMR1 analysis (cohort 2) was primarily because of SNV/INDELs constituting 94% of pathogenic/likely pathogenic variants. Altogether, the combined yield of cohort 2 and cohort 3 amount to roughly the same yield as in cohort 1 (37% vs 35%, respectively). This study also shows that a GS-first approach shortens the diagnostic odyssey for individuals with ID in our region. If ID individuals are first investigated using CMA, diagnosis is delayed approximately 6 to 12 months, including the TAT for genetic analysis as well as time to obtain a new referral for GS (Supplemental Figure 1; Supplemental Table 5). However, for most individuals (90%), no more genetic tests were requested after the negative CMA, and therefore, many individuals who could have received a genetic diagnosis remain undiagnosed. Because the phenotypes of the individuals in cohort 1 and cohort 3 were similar (Table 1), we find it probable that GS first would have led to a genetic diagnosis in more individuals. Assuming a similar yield as in cohort 1 (35%), 147 individuals could theoretically have been diagnosed instead of 47.
      Altogether, in the 521 individuals who underwent primary testing for chromosomal rearrangements either through GS-first analysis (cohort 1) or CMA analysis (cohort 3), 2 CCRs were detected (0.4%). This further supports that CCRs are important pathogenic alleles in rare diseases that need to be considered in individuals with ID. The complex DUP-NML-DEL-NML-DUP rearrangement on chromosome 18q (C3-P152) was characterized by GS as part of the clinical investigation, whereas the other CCR, a translocated DUP (C3-P348) was followed up with FISH and chromosome analysis (Figure 3). In the 18q CCR, the GS characterization was able to fully resolve the derivative chromosome 18 including 4 BPJs, 1 DEL, 1 tandem DUP, and 1 inverted inserted DUP. Replication-based mechanisms (fork-stalling and template-switching)
      • Lee J.A.
      • Carvalho C.M.
      • Lupski J.R.
      A DNA replication mechanism for generating nonrecurrent rearrangements associated with genomic disorders.
      have previously been suggested in similar CCRs, often causing a mixture of DUPs, DELs, and inversions. The presence of the same 5 nt insertion in 2 of the 4 BPJs lends further support to such an underlying mechanism.
      In conclusion, currently 3 different genetic investigation strategies are applied for individuals with ID/NDD at the Karolinska University Hospital; (1) GS first, (2) CMA/FMR1/GS, and (3) CMA/FMR1. Of those, GS-first singleton analysis resulted in the highest overall yield (37%). We suggest that GS first-line singleton analysis should be used as the initial genetic investigation in individuals with ID and suspected ID because it is high performing and allows for a time- and cost-effective genetic diagnostic analysis.

      Data Availability

      Detected clinically relevant variants are listed in Supplemental Tables 1-3. The ethical approval did not permit sharing of genome sequencing data. Access to de-identified data that are not provided may be requested via the corresponding author.

      Conflict of Interest

      An.L. received honoraria from Illumina, Inc. All other authors declare no conflict of interest.

      Acknowledgments

      We are very grateful to the participating families. Several authors of this publication are members of the European Reference Network on Rare Congenital Malformations and Rare Intellectual Disability ERN-ITHACA (EU Framework Partnership Agreement ID: 3HP-HP-FPA ERN-01-2016/739516). A.L. was funded by the Swedish Research Council (2017-02936 and 2019-02078), the Stockholm Regional Council, the Strategic Research Area Neuroscience at Karolinska Institutet (StratNeuro), and the Swedish Brain Foundation (FO2020-0351). M.P. was funded by the Swedish Rare Diseases Research Foundation (Sällsyntafonden). A.N. was funded by the Swedish Research Council, Sweden (2018-02652), The Hållsten Research Foundation , the Stockholm Regional Council, and the Swedish Brain Foundation (FO2020-0340). A.H. was funded by the Swedish Rare Diseases Research Foundation (Sällsyntafonden) and Sällskapet Barnavård. Funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

      Author Information

      Conceptualization: An.L., M.J.S., A.N.; Data Curation: An.L., M.E., M.K., B.-M.A., E.B., J.C., J.E., G.G., P.G., A.H., H.T.H., M.H.-P., E.K., K.L.-R., Ag.L., H.L., H.M., D.N., Ev.S., M.Pa., El.S., B.T., E.T., Joh.W., M.W., Jos.W., M.J.S., M.Pe., A.N.; Formal Analysis: An.L., M.E., M.K., J.E., H.T.H., L.-Å.L., M.Pe., A.N.; Funding Acquisition: An.L., A.H., M.Pe., A.N.; Project Administration: An.L., A.N.; Resources: An.L., M.J.S., A.N.; Supervision: An.L.; Visualization: M.E., J.E.; Writing-original draft: An.L., M.E., M.Pe.; Writing-review and editing: An.L., M.E., M.K., B.-M.A., E.B., J.C., J.E., G.G., P.G., A.H., H.T.H., M.H.-P., E.K., K.L.-R., L.-Å.L., Ag.L., H.L., H.M., D.N., Ev.S., M.Pa., El.S., B.T., E.T., Joh.W., M.W., Jos.W., M.J.S., M.Pe., A.N.

      Ethics Declaration

      Ethics approval was given by the Regional Ethical Review Board in Stockholm, Sweden (ethics permit numbers KS 2012/222-31/3 and 2012/2106-31/4). This ethics permit allows for use of clinical samples for analysis of scientific importance as part of clinical development. Our Institutional Review Board approval does not require us to obtain written consent for clinical testing. The research conformed to the principles of the Helsinki Declaration. Written informed consent to participate was obtained to publish clinical information.

      Supplementary Material

      References

        • Maulik P.K.
        • Mascarenhas M.N.
        • Mathers C.D.
        • Dua T.
        • Saxena S.
        Prevalence of intellectual disability: a meta-analysis of population-based studies.
        Res Dev Disabil. 2011; 32 (Published correction appears in Res Dev Disabil. 2013;34(2):729.): 419-436
      1. American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders. 5th ed. American Psychiatric Association Publishing.

        • Dragojlovic N.
        • Elliott A.M.
        • Adam S.
        • et al.
        The cost and diagnostic yield of exome sequencing for children with suspected genetic disorders: a benchmarking study.
        Genet Med. 2018; 20: 1013-1021https://doi.org/10.1038/gim.2017.226
      2. Committee to Evaluate the Supplemental Security Income Disability Program for Children with Mental Disorders, Board on the Health of Select Populations, Board on Children, Youth, and Families, Institute of Medicine; Division of Behavioral and Social Sciences and Education National Academies of Sciences, Engineering, and Medicine. Boat TF, Wu JT, eds. Mental Disorders and Disabilities among Low-Income Children. National Academies Press; 2015.

        • Vissers L.E.L.M.
        • Gilissen C.
        • Veltman J.A.
        Genetic studies in intellectual disability and related disorders.
        Nat Rev Genet. 2016; 17: 9-18https://doi.org/10.1038/nrg3999
        • Martin A.R.
        • Williams E.
        • Foulger R.E.
        • et al.
        PanelApp crowdsources expert knowledge to establish consensus diagnostic gene panels.
        Nat Genet. 2019; 51: 1560-1565https://doi.org/10.1038/s41588-019-0528-2
        • van der Sanden B.P.G.H.
        • Corominas J.
        • de Groot M.
        • et al.
        Systematic analysis of short tandem repeats in 38,095 exomes provides an additional diagnostic yield.
        Genet Med. 2021; 23: 1569-1573https://doi.org/10.1038/s41436-021-01174-1
        • Plesser Duvdevani M.
        • Pettersson M.
        • Eisfeldt J.
        • et al.
        Whole-genome sequencing reveals complex chromosome rearrangement disrupting NIPBL in infant with Cornelia de Lange syndrome.
        Am J Med Genet A. 2020; 182: 1143-1151https://doi.org/10.1002/ajmg.a.61539
        • Hammarsjö A.
        • Pettersson M.
        • Chitayat D.
        • et al.
        High diagnostic yield in skeletal ciliopathies using massively parallel genome sequencing, structural variant screening and RNA analyses.
        J Hum Genet. 2021; 66: 995-1008https://doi.org/10.1038/s10038-021-00925-x
        • Bramswig N.C.
        • Lüdecke H.J.
        • Pettersson M.
        • et al.
        Identification of new TRIP12 variants and detailed clinical evaluation of individuals with non-syndromic intellectual disability with or without autism.
        Hum Genet. 2017; 136: 179-192https://doi.org/10.1007/s00439-016-1743-x
        • Grigelioniene G.
        • Nevalainen P.I.
        • Reyes M.
        • et al.
        A large inversion involving GNAS exon A/B and all exons encoding Gsα is associated with autosomal dominant pseudohypoparathyroidism type Ib (PHP1 billion).
        J Bone Miner Res. 2017; 32: 776-783https://doi.org/10.1002/jbmr.3083
        • Miller D.T.
        • Adam M.P.
        • Aradhya S.
        • et al.
        Consensus statement: chromosomal microarray is a first-tier clinical diagnostic test for individuals with developmental disabilities or congenital anomalies.
        Am J Hum Genet. 2010; 86: 749-764https://doi.org/10.1016/j.ajhg.2010.04.006
        • Beaudet A.L.
        The utility of chromosomal microarray analysis in developmental and behavioral pediatrics.
        Child Dev. 2013; 84: 121-132https://doi.org/10.1111/cdev.12050
        • Waggoner D.
        • Wain K.E.
        • Dubuc A.M.
        • et al.
        Yield of additional genetic testing after chromosomal microarray for diagnosis of neurodevelopmental disability and congenital anomalies: a clinical practice resource of the American College of Medical Genetics and Genomics (ACMG).
        Genet Med. 2018; 20: 1105-1113https://doi.org/10.1038/s41436-018-0040-6
        • Stranneheim H.
        • Lagerstedt-Robinson K.
        • Magnusson M.
        • et al.
        Integration of whole genome sequencing into a healthcare setting: high diagnostic rates across multiple clinical entities in 3219 rare disease patients.
        Genome Med. 2021; 13: 40https://doi.org/10.1186/s13073-021-00855-5
        • Lindstrand A.
        • Eisfeldt J.
        • Pettersson M.
        • et al.
        From cytogenetics to cytogenomics: whole-genome sequencing as a first-line test comprehensively captures the diverse spectrum of disease-causing genetic variation underlying intellectual disability.
        Genome Med. 2019; 11: 68https://doi.org/10.1186/s13073-019-0675-1
        • Smedley D.
        • Smith K.R.
        • et al.
        • 100,000 Genomes Project Pilot Investigators
        100,000 Genomes pilot on rare-disease diagnosis in health care – preliminary report.
        N Engl J Med. 1868-1880; 2021: 385https://doi.org/10.1056/NEJMoa2035790
        • McKenna A.
        • Hanna M.
        • Banks E.
        • et al.
        The Genome Analysis Toolkit: a MapReduce framework for analyzing next-generation DNA sequencing data.
        Genome Res. 2010; 20: 1297-1303https://doi.org/10.1101/gr.107524.110
        • Abyzov A.
        • Urban A.E.
        • Snyder M.
        • Gerstein M.
        CNVnator: an approach to discover, genotype, and characterize typical and atypical CNVs from family and population genome sequencing.
        Genome Res. 2011; 21: 974-984https://doi.org/10.1101/gr.114876.110
        • Chen X.
        • Schulz-Trieglaff O.
        • Shaw R.
        • et al.
        Manta: rapid detection of structural variants and indels for germline and cancer sequencing applications.
        Bioinformatics. 2016; 32: 1220-1222https://doi.org/10.1093/bioinformatics/btv710
        • Eisfeldt J.
        • Vezzi F.
        • Olason P.
        • Nilsson D.
        • Lindstrand A.
        TIDDIT, an efficient and comprehensive structural variant caller for massive parallel sequencing data.
        F1000Res. 2017; 6: 664https://doi.org/10.12688/f1000research.11168.2
        • Dolzhenko E.
        • Deshpande V.
        • Schlesinger F.
        • et al.
        ExpansionHunter: a sequence-graph-based tool to analyze variation in short tandem repeat regions.
        Bioinformatics. 2019; 35: 4754-4756https://doi.org/10.1093/bioinformatics/btz431
        • Chen X.
        • Sanchis-Juan A.
        • French C.E.
        • et al.
        Spinal muscular atrophy diagnosis and carrier screening from genome sequencing data.
        Genet Med. 2020; 22: 945-953https://doi.org/10.1038/s41436-020-0754-0
        • Boeva V.
        • Popova T.
        • Bleakley K.
        • et al.
        Control-FREEC: a tool for assessing copy number and allelic content using next-generation sequencing data.
        Bioinformatics. 2012; 28: 423-425https://doi.org/10.1093/bioinformatics/btr670
        • Richards S.
        • Aziz N.
        • Bale S.
        • et al.
        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.
        Genet Med. 2015; 17: 405-424https://doi.org/10.1038/gim.2015.30
        • Lindstrand A.
        • Malmgren H.
        • Verri A.
        • et al.
        Molecular and clinical characterization of patients with overlapping 10p deletions.
        Am J Med Genet A. 2010; 152A: 1233-1243https://doi.org/10.1002/ajmg.a.33366
      3. The Swedish Riksbank. Search interest rates & exchange rates. The Swedish Riksbank.
        (Accessed June 6, 2021.)
        • Burnside R.D.
        • Pasion R.
        • Mikhail F.M.
        • et al.
        Microdeletion/microduplication of proximal 15q11.2 between BP1 and BP2: a susceptibility region for neurological dysfunction including developmental and language delay.
        Hum Genet. 2011; 130: 517-528https://doi.org/10.1007/s00439-011-0970-4
        • Jønch A.E.
        • Douard E.
        • Moreau C.
        • et al.
        Estimating the effect size of the 15Q11.2 BP1-BP2 deletion and its contribution to neurodevelopmental symptoms: recommendations for practice.
        J Med Genet. 2019; 56: 701-710https://doi.org/10.1136/jmedgenet-2018-105879
        • Gilissen C.
        • Hehir-Kwa J.Y.
        • Thung D.T.
        • et al.
        Genome sequencing identifies major causes of severe intellectual disability.
        Nature. 2014; 511: 344-347https://doi.org/10.1038/nature13394
        • Martínez F.
        • Caro-Llopis A.
        • Roselló M.
        • et al.
        High diagnostic yield of syndromic intellectual disability by targeted next-generation sequencing.
        J Med Genet. 2017; 54: 87-92https://doi.org/10.1136/jmedgenet-2016-103964
        • Srivastava S.
        • Love-Nichols J.A.
        • Dies K.A.
        • et al.
        Meta-analysis and multidisciplinary consensus statement: exome sequencing is a first-tier clinical diagnostic test for individuals with neurodevelopmental disorders.
        Genet Med. 2019; 21 (Published correction appears in Genet Med. 2020;22(10):1731-1732.): 2413-2421
        • Gordeeva V.
        • Sharova E.
        • Babalyan K.
        • Sultanov R.
        • Govorun V.M.
        • Arapidi G.
        Benchmarking germline CNV calling tools from exome sequencing data.
        Sci Rep. 2021; 11: 14416https://doi.org/10.1038/s41598-021-93878-2
        • Köhler S.
        • Doelken S.C.
        • Mungall C.J.
        • et al.
        The Human Phenotype Ontology project: linking molecular biology and disease through phenotype data.
        Nucleic Acids Res. 2014; 42: D966-D974https://doi.org/10.1093/nar/gkt1026
        • Magnusson M.
        • Eisfeldt J.
        • Nilsson D.
        • et al.
        Loqusdb: added value of an observations database of local genomic variation.
        BMC Bioinformatics. 2020; 21: 273https://doi.org/10.1186/s12859-020-03609-z
        • Järvelä I.
        • Määttä T.
        • Acharya A.
        • et al.
        Exome sequencing reveals predominantly de novo variants in disorders with intellectual disability (ID) in the founder population of Finland.
        Hum Genet. 2021; 140: 1011-1029https://doi.org/10.1007/s00439-021-02268-1
        • Sabatini L.M.
        • Mathews C.
        • Ptak D.
        • et al.
        Genomic sequencing procedure microcosting analysis and health economic cost-impact analysis: A report of the association for molecular pathology.
        J Mol Diagn. 2016; 18: 319-328https://doi.org/10.1016/j.jmoldx.2015.11.010
        • Lee J.A.
        • Carvalho C.M.
        • Lupski J.R.
        A DNA replication mechanism for generating nonrecurrent rearrangements associated with genomic disorders.
        Cell. 2007; 131: 1235-1247https://doi.org/10.1016/j.cell.2007.11.037