ABSTRACT
Background
While whole genome and exome sequencing is already being implemented in clinics across the globe, there is still very little uptake of these technologies in African healthcare facilities. Significant hurdles remain that impede the translation of these highly effective mutation screening strategies into appropriate and effective clinical services in low‐ and middle‐income countries.
Methods
We developed, validated, and implemented a phenotype‐driven genetic testing platform using a 500‐gene Ion AmpliSeq Inherited Diseases Panel (IDP) within the South African State healthcare system. This panel enables testing for over 1000 Mendelian disorders using a single laboratory workflow, followed by selective, phenotype‐guided analysis to minimize the risk of incidental or secondary findings. A pooled sequencing strategy was used to optimize resource efficiency, achieving an average coverage of 200–400× per sample.
Results
Among 276 patients tested during the validation phase, a molecular diagnosis was identified in 46% (127/276) of cases. All pathogenic and likely pathogenic variants were submitted to ClinVar to support global data sharing and variant interpretation. The high coverage achieved reduced the need for additional validation testing, supporting the panel's efficiency and reliability in a clinical diagnostic setting.
Discussion
This IDP enables us to improve genetic counselling, identify patients with well‐described genetic disorders for whom appropriate interventions are needed and/or available and identify patients that would benefit from whole exome or genome‐based testing, and allow for risk clarification for close relatives. This system has laid the foundation for rare disease genetic testing in the South African public healthcare system and can be used as a point of departure for genetic service implementation in resource‐constrained environments.
We developed and implemented a 500‐gene panel for phenotype‐driven genetic testing of Mendelian disorders in South Africa's public healthcare system, achieving a 46% diagnostic yield. This platform supports scalable, cost‐effective rare disease diagnosis and lays the foundation for broader genetic services in resource‐limited settings.
1. Introduction
A timeous, accurate genetic diagnosis has many benefits beyond improved clinical care. For patients and families affected by Mendelian disorders it enables an appropriate targeted management plan, accurate prognosis, presymptomatic/predictive testing of other at‐risk family members and genetic counselling, which includes risk prediction for future pregnancies (Joseph et al. 2016). Curative treatment solutions are now possible for a small, but growing, group of disorders. These generally require a confirmed molecular diagnosis as they may be gene or even variant specific. For healthcare systems it enables efficient deployment of clinical resources, minimizes expenditure on diagnostic investigations, and creates valuable teaching opportunities, especially for rare diseases.
Clinical genetic testing for Mendelian disorders has advanced at a considerable rate in recent years and much of this progress has been attributed to the adoption of next‐generation sequencing (NGS) into research and routine service. Whole exome‐ and genome sequencing (WES/WGS) are powerful tools to enable and drive precision‐ and genomic medicine on an unprecedented scale. WGS/WES‐based strategies are particularly cost‐effective for paediatric onset disorders (Lavelle et al. 2022; Schofield et al. 2019) and a growing number of clinical testing guidelines recommendations have adopted WES as a first‐tier test (Manickam et al. 2021; Smith et al. 2022; Souche et al. 2022).
However, WES/WGS remains expensive, and the interpretation of results and management of data require specialist genetics and bioinformatics expertise. The implementation of these advancements and guidelines in resource‐limited settings remains unclear, and the majority of patients with Mendelian disorders in low‐ and middle‐income countries (LMICs) have yet to benefit from these developments. These challenges are particularly pronounced in Africa, where overburdened public healthcare systems have yet to integrate NGS‐based genetic testing into the majority of healthcare facilities. Limited access to qualified medical geneticists and/or genetic counselors, lack of genomics infrastructure, and a paucity of African reference data compound some of these challenges (Kamp et al. 2021; Lumaka et al. 2022). Emerging reports from African clinical settings demonstrate that access to targeted NGS can substantially shorten diagnostic odysseys for patients with rare genetic disorders, but also document persistent systemic barriers such as prolonged delays before referral for genetic testing, limited in‐country sequencing capacity, and dependence on externally supported diagnostic pipelines (Flynn et al. 2021; Campbell et al. 2023; Seymour et al. 2024). This is compounded by the limited availability to confirm clinical diagnoses. This increases the risk for misdiagnosis, and at‐risk family members cannot be offered pre‐symptomatic testing, thus reducing opportunities for early intervention and pre‐emptive management.
Appropriate and innovative solutions are urgently needed to leverage existing expertise, networks and infrastructure to create a stronger foothold upon which effective genomic medicine initiatives can be built. We outline our approach to establishing a phenotype‐driven genetic diagnostic testing platform for a comprehensive range of Mendelian disorders, anticipated to vary from common (1 in 500) to very rare within the South African state healthcare system. These disorders are typically clinically distinguishable and recognizable. We share the molecular findings and comment on the diagnostic utility of this panel for 1015 disorders on a cohort of 276 patients with distinct monogenic disorders suspected based on their clinical phenotype. These molecular diagnoses informed patient management and enabled genetic counselling, highlighting the potential of NGS‐based testing to improve outcomes despite resource‐limited settings.
2. Methods
2.1. Panel Design
We developed an inherited diseases panel (IDP) targeting the coding regions and splice‐site junctions of 500 genes to enable mutation screening for a group of heterogeneous but clinically distinguishable Mendelian disorders using the Ion AmpliSeq Designer online tool. Gene selection was guided by the ClinGen framework for the evaluation of gene‐disease associations and we only included genes with a definitive or strong classification (Strande et al. 2017). We used a combination of existing commercially available Mendeliome panels, screening recommendations for germline cancers from the National Comprehensive Cancer Network (NCCN) and the American College of Medical Genetics and Genomics (ACMG) list of clinically actionable genes (Miller et al. 2021) as a point of departure for the panel design. We included genes/disorders that are particularly relevant to genetics clinics serviced by our Division. This included some disorders where we were already screening for common or local founder mutations in single genes. As a general principle we included genes which accounted for more than 5% of documented cases according to GeneReviews for disorders with genetic heterogeneity. Genes only implicated in very rare (< 5%) or isolated cases were excluded unless we had local context to support inclusion. Our selection was restricted to genes that were available on the Ion Torrent On‐demand catalogue at the time. However, the benefit of this restriction was that the IDP assay was wet lab verified and certified to work with the standardized AmpliSeq library preparation protocol, minimizing the need for optimization. During panel validation, coverage analysis confirmed that > 95% of targeted bases were covered at ≥ 20× read depth, meeting standard diagnostic requirements.
2.2. Implementation Strategy
We first performed an in‐house validation using 32 samples that had previously been genotyped or sequenced using existing Sanger‐ or PCR‐based assays in our laboratory. The validation cohort represented a range of disease phenotypes, and all patients reported African ancestry at referral. While detailed ethnicity information was not available, this cohort reflects the typical patient population seen in genetics clinics within the South African state healthcare system. The results from this run showed 100% concordance with previously identified variant calls, demonstrating reliable detection of single nucleotide variants (SNVs) and small insertions/deletions within the targeted regions. Coverage analysis during panel validation confirmed that > 95% of targeted bases were covered at ≥ 20× read depth, meeting standard diagnostic requirements. We then used the IDP to screen for disease‐causing variants in DNA samples from a cohort of patients with a strong clinical suspicion of one or a subset of the Mendelian diseases targeted by our IDP. All samples were processed using the sequencing and variant calling workflow indicated below. This was followed by selective analysis based on the clinical indication to limit incidental or secondary findings and ensure that reporting was appropriate.
2.3. Sequencing and Variant Analysis
DNA was extracted from peripheral blood samples, using a modified version of the salting out method (Miller et al. 1988). Each sample was diluted to 1 ng/μL prior to DNA library preparation. DNA library preparation was performed in batches of 8 using the Ion AmpliSeq Chef DL8 kit on the Ion Chef System according to standard manufacturer protocols (Thermo Fisher Scientific, US). Sixteen barcoded libraries were pooled in equimolar concentration (50 pM each) for sequencing on an Ion 540 chip. This was followed by automated template preparation on the Ion Chef system. Sequencing of the loaded Ion 540 chip was then performed on the Ion GeneStudio S5 Sequencer. Read alignment and variant calling were performed by pre‐installed plug‐ins in the Ion Torrent Suite software 5.14 using the GRCh37/hg19 human reference genome. All further analysis was then restricted to the genes indicated on the test request from a medical geneticist or genetic counsellor. A filter was applied using the Ion Torrent Suite to exclude all genes that were not included in the test request. This produced a VCF file that only contained variants identified in the appropriate target genes for that patient or family. The filtered VCF file was then annotated using the Ensembl Variant Effect Predictor (McLaren et al. 2016). Variant interpretation was performed according to the ACMG‐AMP guidelines (Richards et al. 2015). Visual inspection of shortlisted variants was performed using the Integrated Genome Viewer (IGV) (Robinson et al. 2011). Initial variant curation and interpretation was performed by an experienced medical scientist and reviewed by a second medical scientist before reporting. Complex cases or ‐variants were discussed at an in‐house multidisciplinary meeting with input from medical scientists, genetic counsellors and medical geneticists prior to reporting. Only Class IV (likely pathogenic) and V (pathogenic) variants were reported to the referring clinician. VUS were only reported in cases where they were thought to be potentially reclassifiable based on clinical/segregation analysis.
3. Results
Our IDP of 500 genes enables mutation profiling for a minimum of 1015 OMIM disease phenotypes (Table S1). The panel targets a region of 1.71 Mb using 14,265 amplicons resulting in 99.94% coverage of the regions of interest. The design targeted all protein coding regions including UTRs and intron/exon boundaries (20 bp) for the selected genes. The design file with coverage statistics and amplicon distribution is given in the Supporting Information. The pooling strategy of 16 libraries per Ion 540 chip enabled an average target region coverage of 370× (min. 113×, max 570×).
A total of 276 probands have been sequenced using the IDP in our laboratory thus far. The majority of probands were of African ancestry as indicated on the request form by the referring clinician. Reportable variants were identified for 46% (127/276) of patients thus far (Table 1). Twenty‐six percent (33/127) of the variants identified generated the first ClinVar entry.
TABLE 1.
Reportable variants identified in this study with corresponding ACMG‐AMP classification and ClinVar identifiers.
| Patient | Test request | Variants identified | ACMG‐AMP classification | ClinVar accession number |
|---|---|---|---|---|
| IDP5 | Tuberous sclerosis | TSC2: NM_000548.3: c.5238_5255del (p.His1746_Arg1751del) | Het (Class V) | SCV004034090.1 |
| IDP6 | RASopathy | PTPN11: NM_002834.3: c.184T>G (p.Tyr62Asp) | Het (Class V) | SCV004034101.1 |
| IDP7 | Sotos syndrome | NSD1: NM_022455.4: c.4913A>G (p.His1638Arg) | Het (Class IV) | SCV004034112.1 |
| IDP9 | CHARGE | CHD7: NM_017780.2: c.2544del (p.Asp849IlefsTer39) | Het (Class IV) | SCV004034123.1* |
| IDP12 | Trichorhinophalangeal syndrome | TRPS1: NM_014112.5:c.2719del (p.Val907PhefsTer7) | Het (Class IV) | SCV004034128.1* |
| IDP13 | RASopathy | PTPN11: NM_002834.3: c.854T>C (p.Phe285Ser) | Het (Class V) | SCV004034129.1 |
| IDP15 | Osteogenesis imperfecta | COL1A2: NM_000089.4: c.2297G>A (p.Gly766Asp) | Het (Class IV) | SCV004034130.1* |
| IDP16 | Osteogenesis imperfecta | COL1A1: NM_000088.3: c.1426G>A (p.Gly476Arg) | Het (Class V) | SCV004034131.1* |
| IDP19 | Myotonia congenita | CLCN1: NM_000083.3: c.1925C>G (p.Ser642Ter) | Hom (Class V) | SCV004034132.1 |
| IDP20 | Li‐Fraumeni syndrome | TP53: NM_001126114.2: c.528C>A (p.Cys176Ter) | Het (Class V) | SCV004034091.1 |
| IDP21 | Cancer syndrome (Ovarian and colon cancer) | BRCA2: NM_000059.3: c.5213_5216del (p.Thr1738IlefsTer2) | Het (Class V) | SCV004034092.1 |
| IDP22 | Distal arthrogryposis (Freeman Sheldon syndrome) | MYH3: NM_002470.4: c.2096T>A (p.Val699Asp) | Het (Class IV) | SCV004034093.1* |
| IDP23 | Hereditary breast ovarian cancer syndrome | BRCA1: NM_007294.3: c.45dup (p.Asn16Ter) | Het (Class V) | SCV004034094.1 |
| IDP24 | Ectodermal dysplasia | EDA: NM_001399.5: c.706G>A (p.Gly236Ser) | Het (Class IV) | SCV004034095.1* |
| IDP25 | Treacher Collins syndrome | TCOF1: NM_001371623.1: c.158G>A (p.Trp53Ter) | Het (Class IV) | SCV004034096.1* |
| IDP26 | Pycnodysostosis | CTSK: NM_000396.4: c.953G>A (p.Cys318Tyr) | Hom (Class V) | SCV004034097.1 |
| IDP28 | Skeletal dysplasia | TRPV4: NM_021625.4: c.992T>C (p.Ile331Thr) | Het (Class IV) | SCV004034098.1 |
| IDP29 | Marfan syndrome | FBN1: NM_000138.4: c.6670dupA (p.Thr2224AsnfsTer6) | Het (Class IV) | SCV004034099.1* |
| IDP30 | Marfan syndrome | FBN1: NM_000138.4: c.640G>A (p.Gly214Ser) | Het (Class V) | SCV004034100.1 |
| IDP31 | Marfan syndrome | FBN1: NM_000138.4: c.3794G>A (p.Cys1265Tyr) | Het (Class V) | SCV004034102.1 |
| IDP32 | Marfan syndrome | FBN1: NM_000138.4: c.5726T>C (p.Ile1909Thr) | Het (Class V) | SCV004034103.1 |
| IDP35 | Marfan syndrome | FBN1: NM_000138.4: c.3037G>A (p.Gly1013Arg) | Het (Class V) | SCV004034104.1 |
| IDP39 | Cystic fibrosis | CFTR: NM_000492.3: c.614C>G (p.Pro205Arg) | Het (Class IV) | SCV004034105.1 |
| CFTR: NM_000492.3: c.3064_3117del (p.Val1022_Gln1039del) | Het (Class IV) | SCV004034106.1 | ||
| IDP46 | Waardenburg syndrome | PAX3: NM_181458.4: c.602C>A (p.Ser201Ter) | Het (Class IV) | SCV004034107.1* |
| IDP47 | Noonan syndrome | PTPN11: NM_002834.3: c.1391G>C (p.Gly464Ala) | Het (Class V) | SCV004034108.1 |
| IDP48 | Ehlers‐Danlos syndrome | COL5A2: NM_000393.5: c.2662‐2A>G | Het (Class IV) | SCV004034109.1* |
| IDP49 | Tuberous sclerosis | TSC2: NM_000548.3: c.4129C>T (p.Gln1377Ter) | Het (Class V) | SCV004034110.1 |
| IDP50 | Alagille syndrome | JAG1: NM_000214.3: c.2406del (p.Trp803GlyfsTer17) | Het (Class V) | SCV004034111.1* |
| IDP58 | Noonan‐Neurofibromatosis syndrome | PTPN11: NM_002834.3: c.188A>G (p.Tyr63Cys) | Het (Class V) | SCV004034113.1 |
| NF1: NM_001042492.2: c.3114‐2A>G | Het (Class IV) | SCV004034114.1 | ||
| IDP62 | Familial adenomatous polyposis | APC: NM_000038.6:c.3183_3187del (p.Glu1062HisfsTer2) | Het (Class V) | SCV004034115.1 |
| IDP68 | Mucopolysaccharidosis | ARSB: NM_000046.5: c.905G>A (p.Gly302Glu) | Hom (Class IV) | SCV004034116.1* |
| IDP71 | Neurofibromatosis type 1 | NF1: NM_001042492.2: c.1890del (p.Gly631AspfsTer57) | Het (Class IV) | SCV004034117.1* |
| IDP73 | EEC syndrome | TP63: NM_003722.5: c.604T>C (p.Tyr202His) | Het (Class IV) | SCV004034118.1* |
| IDP74 | Lamellar ichthyosis | TGM1: NM_000359.3: c.944G>T (p.Arg315Leu) | Hom (Class V) | SCV004034119.1 |
| IDP79 | Marfan syndrome | FBN1: NM_000138.4: c.23_34del (p.Glu8_Leu11del) | Het (Class IV) | SCV004034120.1* |
| IDP82 | X‐linked adrenoleukodystrophy | ABCD1: NM_000033.3: c.1661G>A (p.Arg554His) | Hemi (Class V) | SCV004034121.1 |
| IDP83 | Cystic fibrosis | CFTR: NM_000492.3: c.2988+1G>A | Hom (Class V) | SCV004034122.1 |
| IDP84 | Antley‐Bixler syndrome | FGFR2: NM_000141.4: c.1024T>C (p.Cys342Arg) | Het (Class V) | SCV004034124.1 |
| IDP88 | Haemophilia A | F8: NM_000132.4: c.683A>C (p.His228Pro) | Hemi (Class IV) | SCV004034125.1* |
| IDP89 | Haemophilia B | F9: NM_000133.3: c.880C>T (p.Arg294Ter) | Hemi (Class V) | SCV004034126.1 |
| IDP90 | RASopathy | LZTR1: NM_006767.4: c.1234C>T (p.Arg412Cys) | Het (Class IV) | SCV004034127.1 |
| IDP92 | RASopathy | NF1: NM_001042492.2: c.1885G>C (p.Gly629Arg) | Het (Class IV) | SCV003840175.1 |
| IDP94 | Congenital myopathy | RYR1: NM_000540.2:c.14524G>A | Het (Class III) | SCV004123104.1 |
| RYR1: NM_000540.3: c.10348‐6C>G | Het (Class V) | SCV004123105.1 | ||
| RYR1: NM_000540.2: c.9170_9171delTT (p.Phe3057TrpfsTer29) | Het (Class V) | SCV004123106.1* | ||
| IDP96 | Tuberous sclerosis | TSC2: NM_000548.3: c.4516delC (p.His1506IlefsTer70) | Het (Class V) | SCV004123042.1 |
| IDP97 | Tuberous sclerosis | TSC2: NM_000548.3: c.4628A>G (p.His1543Arg) | Het (Class IV) | SCV004123053.1 |
| IDP99 | Chondrodysplasia punctata | COL2A1: NM_001844.5: c.905C>T (p.Ala302Val) | Het (Class V) | SCV004123064.1 |
| IDP101 | Polycystic kidney disease (autosomal recessive) | PKHD1: NM_138694.4: c.1880T>A (p.Met627Lys) | Hom (Class V) | SCV004123075.1 |
| IDP103 | Metatropic dysplasia | TRPV4: NM_021625.4: c.2396C>T (p.Pro799Leu) | Het (Class V) | SCV004123086.1 |
| IDP107 | Neurofibromatosis type 1 | NF1: NM_001042492.3:c.4930G>A (p.Asp1644Asn) | Het (Class IV) | SCV004123097.1 |
| IDP109 | Leopard syndrome | PTPN11: NM_002834.3: c.836A>G (p.Tyr279Cys) | Het (Class V) | SCV004123107.1 |
| IDP110 | Charcot–Marie–Tooth disease | MFN2: NM_014874.3: c.1090C>T (p.Arg364Trp) | Het (Class V) | SCV004123108.1 |
| IDP111 | Treacher Collins syndrome | TCOF1: NM_001371623.1: c.3754del (p.Leu1252TrpfsTer18) | Het (Class IV) | SCV004123109.1* |
| IDP112 | Tuberous sclerosis | TSC2: NM_000548.3: c.4375C>T (p.Arg1459Ter) | Het (Class V) | SCV004123043.1 |
| IDP113 | Paraganglioma/Pheochromocytoma syndrome | SDHD: NM_003002.4: c.197_198delinsAA (p.Trp66Ter) | Het (Class IV) | SCV004123044.1* |
| IDP115 | Paraganglioma/Pheochromocytoma syndrome | RET: NM_020975.4: c.1183G>C (p.Val395Leu) | Het (Class III) | SCV004123045.1 |
| IDP117 | Cystic fibrosis | CFTR: NM_000492.3: c.2988+1G>A | Het (Class V) | SCV004034122.1 |
| IDP118 | Alpha thalassemia intellectual disability syndrome | ATRX: NM_000489.5: c.736C>T (p.Arg246Cys) | Hemi (Class V) | SCV004123046.1 |
| IDP121 | Microcephalic osteodysplastic primordial dwarfism, type II | PCNT: NM_001315529.2: c.192_193del (p.Gly65AspfsTer6) | Hom (Class IV) | SCV004123047.1* |
| IDP123 | Xeroderma pigmentosum | XPA: NM_000380.3: c.389G>A (Arg130Lys) | Hom (Class IV) | SCV004123048.1 |
| IDP124 | Peutz‐Jeghers syndrome | STK11: NM_000455.4: c.526del (p.Asp176ThrfsTer111) | Het (Class V) | SCV004123049.1 |
| IDP126 | Waardenburg syndrome | SOX10: NM_006941.4: c.400C>T (p.Leu134Phe) | Het (class IV) | SCV004123050.1* |
| IDP132 | Neurofibromatosis type 1 | NF1: NM_000267.3: c.1721+1G>A | Het (Class V) | SCV004123051.1 |
| IDP136 | Noonan syndrome | MAP2K1: NM_002755.3: c.215G>A (p.Ser72Asn) | Het (Class III) | SCV004123052.1 |
| IDP137 | Familial hypercholesterolaemia | LDLR: NM_000527.5: c.2054C>T (p.Pro685Leu) | Het (Class V) | SCV004123054.1 |
| IDP138 | Familial hypercholesterolaemia | LDLR: NM_000527.5: c.681C>G (p.Asp227Glu) | Het (Class V) | SCV004123055.1 |
| IDP140 | Haemophilia A | F8: NM_000132.3: c.836T>G (p.Met279Arg) | Hemi (Class IV) | SCV004123056.1* |
| IDP141 | Haemophilia A | F8: NM_000132.3: c.2717C>G (p.Ser906Ter) | Hemi (Class IV) | SCV004123057.1* |
| IDP142 | Cystic fibrosis | CFTR: NM_000492.3: c.2988+1G>A | Het (Class V) | SCV004034122.1 |
| IDP144 | Haemophilia A | F8: NM_000132.4:c.745A>T (p.Lys249Ter) | Hemi (Class IV) | SCV004123058.1* |
| IDP146 | Rett syndrome | MECP2: NM_004992.3:c.753dup (p.Gly252ArgfsTer7) | Het (Class IV) | SCV004123059.1 |
| IDP149 | Rett (−like) syndrome | MEF2C: NM_002397.5: c.565C>T (p.Arg189Ter) | Het (Class V) | SCV004123060.1 |
| IDP150 | Cystic fibrosis | CFTR: NM_000492.3: c.2988+1G>A | Het (Class V) | SCV004034122.1 |
| IDP151 | Hereditary pancreatitis | PRSS1: NM_002769.5: c.365G>A (p.Arg122His) | Het (Class V) | SCV004123061.1 |
| IDP154 | Noonan syndrome | PTPN11: NM_002834.4:c.182A>G (p.Asp61Gly) | Het (Class V) | SCV004123062.1 |
| IDP155 | Cornelia De Lange syndrome | NIPBL: NM_133433.4:c.5465A>G (p.Asp1822Gly) | Het (Class IV) | SCV002552105.1 |
| IDP157 | Epidermolysis bullosa | COL7A1: NM_000094.4: c.2927G>A (p.Trp976Ter) | Het (Class V) | SCV004123063.1 |
| COL7A1: NM_000094.4: c.3265C>T (p.Gln1089Ter) | Het (Class V) | SCV004123065.1 | ||
| IDP159 | Achondroplasia | FGFR3: NM_000142.5: c.1620C>A (p.Asn540Lys) | Het (Class V) | SCV004123066.1 |
| IDP160 | Skeletal dysplasia | COL2A1: NM_001844.5 c.3437G>C (p.Gly1146Ala) | Het (Class IV) | SCV004123067.1* |
| IDP163 | Achondroplasia | FGFR3: NM_000142.5: c.1138G>A (p.Gly380Arg) | Het (Class V) | SCV004123078.1 |
| IDP164 | Marfan syndrome | FBN1: NM_000138.5:c.478T>C (p.Cys160Arg) | Het (Class IV) | SCV004123068.1 |
| IDP171 | Noonan syndrome | PTPN11: NM_001330437.2:c.923A>G (p.Asn308Ser) | Het (Class V) | SCV003840157.1 |
| IDP173 | Neurofibromatosis type 1 | NF1: NM_001042492.3:c.2681T>C (p.Phe894Ser) | Het (Class IV) | SCV004123069.1 |
| IDP177 | Apert syndrome | FGFR2:NM_000141.5: c.755C>G (p.Ser252Trp) | Het (Class V) | SCV004123087.1 |
| IDP178 | Rett syndrome | MECP2: NM_004992.4: c.808C>T (p.Arg270Ter) | Het (Class V) | SCV004123070.1 |
| IDP180 | Haemophilia A | F8: NM_000132.4: c.601G>T (p.Gly201Trp) | Hemi (Class IV) | SCV005420274.2* |
| IDP184 | Crouzon syndrome | FGFR2: NM_000141.5: c.799T>C (p.Ser267Pro) | Het (Class V) | VCV000013290.20 |
| IDP187 | Mucopolysaccharidosis | IDUA NM_000203.5: c.406G>A (p.Ala136Thr) | Het (Class III) | VCV000964816.12 |
| IDP188 | Noonan syndrome | PTPN11: NM_002834.5: c.215C>G (p.Ala72Gly) | Het (Class V) | SCV003840151.1 |
| IDP189 | Noonan syndrome | PTPN11:NM_002834.5 c.184T>G (p.Tyr62Asp) | Het (Class V) | SCV004034101.1 |
| IDP192 | Rett syndrome | MECP2:NM_004992.4:c.880C>T (p.Arg294Ter) | Het (Class V) | VCV000011819.69 |
| IDP194 | Simpson‐Golabi‐Behmel syndrome | GPC3: NM_004484.4: c.190del (p.Val64TyrfsTer20) | Hemi (Class IV) | SCV004123071.1* |
| IDP195 | Pfeiffer syndrome | FGFR2:NM_000141.5:c.870G>T (p.Trp290Cys) | Het (Class V) | SCV004123072.1 |
| IDP198 | Neurofibromatosis type 1 | NF1: NM_000267.3: c.3113+5G>C | Het (Class IV) | SCV004123073.1 |
| IDP199 | Neurofibromatosis type 1 | NF1: NM_000267.3: c.1381C>T (p.Arg461Ter) | Het (Class V) | SCV004123093.1 |
| IDP200 | Xeroderma pigmentosum | XPC: NM_004628.4: c.2251‐1G>C | Hom (Class V) | SCV004123074.1 |
| IDP202 | Joubert syndrome | CPLANE1: NM_001384732.1: c.4482‐1G>T | Het (Class IV) | SCV004123076.1 |
| IDP205 | Achondroplasia | FGFR3: NM_000142.5 c.1031C>G (p.Ser344Cys) | Het (Class IV) | SCV004123077.1 |
| IDP207 | Achondroplasia | FGFR3:NM_000142.5:c.1138G>A (p.Gly380Arg) | Het (Class V) | SCV004123078.1 |
| IDP209 | Sotos syndrome | NSD1: NM_022455.5: c.4823dup (p.Pro1609ThrfsTer9) | Het (Class IV) | SCV004123079.1* |
| IDP212 | Thanatophoric dysplasia type I | FGFR3: NM_000142.5 c.742C>T (p.Arg248Cys) | Het (Class V) | SCV004123085.1 |
| IDP213 | Marfan syndrome | FBN1: NM_000138.5: c.2638G>A (p.Gly880Ser) | Het (Class V) | SCV004123080.1 |
| IDP214 | Neurofibromatosis type 1 | NF1: NM_000267.3: c.1133_1136del (p.Asp378AlafsTer8) | Het (Class V) | SCV004123081.1 |
| IDP215 | Neurofibromatosis type 1 | NF1: NM_001042492.3: c.5609G>A (p.Arg1870Gln) | Het (Class V) | SCV003840144.1 |
| IDP216 | Neurofibromatosis type 1 | NF1: NM_000267.3: c.1A>C (p.Met1Leu) | Het (Class IV) | SCV003840143.2 |
| IDP217 | Haemophilia A | F8: NM_000132.3: c.1120del (p.Met374TrpfsTer44) | Hemi (Class IV) | SCV004123082.1* |
| IDP218 | Haemophilia B | F9: NM_000133.4: c.881G>A (p.Arg294Gln) | Hemi (Class V) | SCV004123083.1 |
| IDP222 | Cornelia de Lange syndrome | NIPBL: NM_133433.4: c.1480_2479del (p.Arg827GlyfsTer2) | Het (Class IV) | SCV004123084.1* |
| IDP225 | Thanatophoric dysplasia | FGFR3: NM_000142.5: c.742C>T (p.Arg248Cys) | Het (Class V) | SCV004123085.1 |
| IDP230 | Pfeiffer syndrome | FGFR2: NM_000141.4: c.755C>G (p.Ser252Trp) | Het (Class V) | SCV004123087.1 |
| IDP231 | Xeroderma pigmentosum | XPC: NM_004628.5: c.2251‐1G>C | Het (Class V) | SCV004123074.1 |
| XPC: NM_004628.5: c.877C>T (p.Arg293Ter) | Het (Class V) | SCV004123088.1 | ||
| IDP232 | Neurofibromatosis type 1 | NF1: NM_001042492.3: c.3721C>T (p.Arg1241Ter) | Het (Class V) | SCV003840164.1 |
| IDP235 | Collagen type 1 disorder | COL1A2: NM_000089.4: c.1901G>T (p.Gly634Val) | Het (Class IV) | SCV004123089.1* |
| IDP236 | Marfan syndrome | FBN1:NM_000138.5:c.3726T>G (p.Cys1242Trp) | Het (Class IV) | SCV004123090.1* |
| IDP240 | Apert syndrome | FGFR2: NM_000141.5: c.755C>G (p.Ser252Trp) | Het (Class V) | SCV004123087.1 |
| IDP241 | Xeroderma pigmentosa | XPC: NM_004628.5: c.2251‐1G>C | Hom (Class V) | SCV004123074.1 |
| IDP242 | Neurofibromatosis type 1 | NF1: NM_001042492.3:c.4600C>T (p.Arg1534Ter) | Het (Class V) | SCV004123091.1 |
| IDP245 | Noonan syndrome | BRAF: NM_004333.6:c.1455G>C (p.Leu485Phe) | Het (Class V) | SCV004123092.1 |
| IDP246 | Neurofibromatosis | NF1: NM_000267.3: c.1381C>T (p.Arg461Ter) | Het (Class V) | SCV004123093.1 |
| IDP254 | CHARGE syndrome | CHD7: NM_017780.4: c.5894+1G>T | Het (Class IV) | SCV004123094.1* |
| IDP255 | Crouzon syndrome | FGFR2: NM_000141.5:c.1025G>A (p.Cys342Tyr) | Het (Class V) | SCV004123095.1 |
| IDP256 | Ectodermal dysplasia | WNT10A: NM_025216.3:c.682T>A (p.Phe228Ile) | Het (Class V) | SCV004123096.1 |
| IDP258 | Duchene muscular dystrophy | DMD: NM_004006.3:c.10223+1G>A | Hemi (Class V) | SCV004123098.1 |
| IDP261 | Bardet‐Biedl syndrome | BBS10: NM_024685:c.728_731del (p.Lys243IlefsTer15) | Hom (Class V) | SCV004123099.1 |
| IDP265 | Neurofibromatosis type 1 | NF1: NM_000267.3: c.625C>T (p.Gln209Ter) | Het (Class V) | SCV004123100.1 |
| IDP271 | Tuberous sclerosis | TSC2: NM_000548.5:c.4375C>T (p.Arg1459Ter) | Het (Class V) | SCV004123043.1 |
| IDP272 | EEC syndrome | TP63: NM_003722.5:c.1028G>A (p.Arg343Gln) | Het (Class IV) | SCV004123101.1 |
| IDP276 | Fanconi anemia | FANCG: NM_004629.2: c.637_643del (p.Tyr213LysfsTer6) | Het (Class V) | SCV004123102.1 |
| FANCG: NM_004629.2: c.179del (p.Leu60ProfsTer12) | Het (Class V) | SCV004123103.1 |
Note: Variants that generated the first ClinVar entry are indicated with an asterisk (*).
A single curator and reviewer pair was able to interpret and report on a batch of 32 samples in 5–10 workdays. Complex cases took longer to resolve due to the extended data analysis and interpretation time. Batching existing or frequently requested tests with rare diseases on the IDP offers several benefits for both the laboratory workflow as well as for the patient in our resource‐constrained environment. This approach was more cost‐effective than running multiple genetic tests individually since it allowed for the simultaneous analysis of multiple genes and enabled optimal usage of precious and limited resources while simultaneously expanding the scope of testing to cover full or multiple genes where single mutation testing was the only option previously. Additionally, the approach greatly improved the turn‐around time to deliver on patient results.
4. Discussion
Genomic medicine implementation is still lagging in the majority of LMICs and this is particularly evident in the African public healthcare sector, which is burdened with a range of challenges including limited budgets, a high infectious disease load, and limited access to new healthcare technologies (Kamp et al. 2021). These challenges can be addressed through innovation, capacity building, collaboration, and advocacy to ensure the successful integration and sustainability of cost‐effective genomic medicine strategies for patients in LMICs. This will enable earlier disease detection, as well as appropriate and effective therapeutic interventions, resulting in improved patient management and health outcomes for individuals in these regions. NGS is a core precision‐ and genomic medicine technology, and efficient, sustainable strategies are needed to integrate this technology into LMICs to address persistent and emerging healthcare disparities. These strategies need to leverage existing infrastructure and expertise with innovative yet pragmatic solutions.
Single‐gene testing or small gene panels are often most appropriate for conditions with distinctive clinical features and minimal locus heterogeneity. The challenge with implementing these lies in the time and resources needed to validate multiple single tests within a clinical testing service. Smaller laboratories lack the resources to perform the necessary optimization and validation work, and these experiments are not economically feasible for rare disorders where the number of tests ordered might not justify the set‐up costs if each panel is run individually. Many laboratories have therefore opted to implement multigene panels as it allows the genomic profiling of a variety of genetic disorders within a single test set‐up and improves turnaround times because rare tests can be batched with more frequently ordered ones to stabilize batch sizes. These multigene panels or clinical exomes can be very cost effective when paired with selective analysis to generate virtual gene panels, and gene panel curation resources (Martin et al. 2019; Strande et al. 2017) are improving to standardize and ensure that appropriate genes are analyzed and reported on, while limiting incidental and secondary findings. Multigene NGS workflows paired with virtual panel analysis strategies are therefore particularly attractive in low resource settings, especially if they can be implemented on benchtop instruments to ease capital expenditure.
The IDP described here is an effective tool for phenotype‐driven diagnostic testing with limited laboratory and clinical resources and can serve as a point of departure for developing cost‐effective genomic medicine services in LMICs. By consolidating multiple gene analyses into a single assay, it reduces the need for sequential testing and optimizes laboratory and personnel resources. Importantly, this workflow is compatible with existing NGS platforms and standardized bioinformatics pipelines, supporting scalability to other public‐sector laboratories in South Africa and similar low‐resource settings. Our facility already housed an Ion GeneStudio Ion S5 System and we therefore chose to tailor our solution to existing infrastructure. This selection of genes can be used as a point of departure for the development of a customised IDP on most NGS instruments.
The diagnostic yield of 46% achieved using this IDP is a result of several key factors, including careful gatekeeping, appropriate virtual panel selection, and effective referral processes. Careful gatekeeping is essential and should be the first step for NGS testing. This process was conducted by medical geneticists and/or genetic counsellors within our Division. It involved screening and evaluating the appropriateness of the genetic test requested, based on reviewing the clinical presentation of the patient and their family history. This approach allowed for relevant genetic testing, particularly to cases where a molecular diagnosis was likely to influence management and allow for downstream testing of other relatives. Patients were carefully evaluated for suitability for inclusion on the IDP, thus ensuring the appropriate and efficient allocation of our resources. The testing process in our laboratory selected the most appropriate genes related to the suspected condition for analysis. This virtual gene panel strategy increased the likelihood of identifying the causative variants and thus enabled us to achieve the high diagnostic yield.
The transition from Sanger sequencing to the IDP for single‐gene testing has provided numerous advantages, particularly within the context of diagnostic services. This shift has enhanced the coverage of genes associated with specific disorders, allowing for comprehensive genomic analysis rather than focusing solely on targeted hotspots or regions harboring common founder mutations. As a result, the diagnostic yield has significantly improved, especially in cases where clinical presentations are ambiguous and individual tests have limited detection capabilities. This approach has facilitated more accurate diagnoses and better‐informed patient management. Additionally, it has enabled the laboratory to rapidly expand its testing capabilities, increasing the number of genetic disorders analyzed from approximately 40 to over a thousand.
The incorporation of the IDP into routine diagnostic workflows has proven to be both cost‐effective and efficient, reducing the need for multiple single‐gene tests or exon‐specific analyses. Its broader, yet targeted, design has resulted in a higher rate of positive diagnoses and the identification of numerous novel genetic variants, reflecting the under‐representation of African populations in global reference datasets (Lumaka et al. 2022). This lack of representation complicates variant interpretation and increases the risk of misclassification, highlighting the urgent need for African‐ancestry genomic data in the public domain to improve diagnostic accuracy. As anticipated, regions with limited background variation also yielded a number of variants of uncertain significance (VUS). Additionally, although copy number variation (CNV) analysis was not included in the initial validation phase, its integration could further enhance the panel's diagnostic yield if reliably implemented (Louw et al. 2023).
The success of implementing NGS technology into the diagnostic setting is highly dependent on effective referral processes, which ensure that patients undergoing testing are appropriate candidates. In this study, patients referred for testing via the IDP were predominantly referred by specialized healthcare professionals, such as medical geneticists and genetic counsellors, thus ensuring that the appropriate tests were performed. This underscores the importance of genomic medicine education for referring clinicians, ensuring that they possess the requisite knowledge for making effective and informed referrals.
The effective implementation of this panel now needs to be complemented with long‐term efforts to ensure that it remains appropriate and to ensure that we can derive its full potential in our low resource setting. To ensure continued clinical relevance, the IDP gene content will be reviewed periodically as part of formal departmental service evaluations. Gene–disease validity will be reassessed using established curation frameworks with multidisciplinary review guiding the addition of newly validated genes and the removal of those with downgraded evidence. As sequencing costs decrease and bioinformatics infrastructure becomes more accessible within the South African public healthcare system, exome‐ or genome‐based approaches with virtual panel analysis may become increasingly appropriate. The IDP is therefore positioned as a pragmatic and sustainable entry point to genomic diagnostics in the current resource‐constrained environment, with flexibility to transition to broader sequencing strategies as capacity evolves.
Clinical variant interpretation is a dynamic process and efforts are underway to implement a regular variant reanalysis workflow. There are, however, many practical considerations and resources needed for this (Appelbaum et al. 2023) that can be challenging in a LMIC laboratory with limited bioinformatics infrastructure and skilled staff. In the interim we plan to reinterpret and reclassify variants previously classified by us as they are identified again in the NGS service. Regular review of yields and gene panel content and redesign using input from the clinic, local research, literature and gene panel curation resources will be important to ensure that the content remains relevant and appropriate. Data from IDP will allow us to gather population‐level data that could highlight local recurring pathogenic variants and help us to understand mutation profiles in understudied populations and disorders.
Data sharing is of vital importance as it enables consolidating genetic knowledge. Promoting and encouraging data sharing, particularly from diverse populations helps to achieve a comprehensive representation of the global population. Diverse datasets support reliable clinical variant classification and accurate interpretation, particularly for rare or novel variants present in African populations who have a unique and diverse genetic architecture (Lumaka et al. 2022). This will ultimately facilitate the development of therapies and advancing genomic medicine.
5. Conclusion
Clinical genomics relies on close cooperation between medical geneticists, genetic counsellors, laboratory scientists, and referring clinicians to ensure appropriate reporting and that NGS resources are used optimally. It is also critical to the implementation of complex interventions. This can be a considerable challenge in the fragmented, under‐resourced South African public healthcare system. Continued efforts are needed to implement pragmatic and locally feasible NGS testing solutions in LMICs to address barriers to genomic medicine and prevent health disparities. Coupled with a focus on data sharing, this can significantly improve clinical variant interpretation resources.
Our IDP allows us to use the same laboratory workflow for > 1000 Mendelian disorders, followed by selective analysis based on the clinical indication to ensure that reporting is appropriate. This enables us to accurately identify (1) patients with well‐documented genetic disorders for whom appropriate interventions are needed and/or available and (2) patients that would benefit from additional phenotyping and/or WES/WGS‐based testing (3) define risk and offer testing for other family members. The high diagnostic yield of this IDP justifies the cost in our setting.
This panel has enabled us to leverage limited resources effectively and to lay the foundation for a precision genomic medicine service in the South African public healthcare system.
Author Contributions
Nadia Carstens: conceptualization, funding acquisition, methodology, formal analysis, investigation, supervision, writing – original draft, project administration. Maria Mudau: investigation, data curation, writing – review and editing. Fahmida Essop: investigation, data curation, writing – review and editing. Amanda Krause: methodology, clinical validation, supervision, writing – review and editing.
Funding
This work is based on the research supported by the National Research Foundation of South Africa (Grants 129779 and 106948).
Ethics Statement
The study was approved by the Human Research Ethics Committee of the University of the Witwatersrand (M170761).
Consent
Informed consent has been obtained from the patients and/or parents.
Conflicts of Interest
The authors declare no conflicts of interest.
Supporting information
Data S1: Provides the design of the IDP targeted sequencing panel.
Table S1: Lists the genes included in the IDP panel, together with relevant annotations.
Data Availability Statement
The variants described here were submitted to ClinVar and can be viewed under Organization ID 508172. The gene panel design is shared in Ion Reporter format in the Supporting Information.
References
- Appelbaum, P. S. , Berger S. M., Brokamp E., et al. 2023. “Practical Considerations for Reinterpretation of Individual Genetic Variants.” Genetics in Medicine 25, no. 5: 100801. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Campbell, L. , Fredericks J., Mathivha K., et al. 2023. “The Implementation and Utility of Clinical Exome Sequencing in a South African Infant Cohort.” Frontiers in Genetics 14: 1277948. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Flynn, K. , Feben C., Lamola L., et al. 2021. “Ending a Diagnostic Odyssey‐The First Case of Takenouchi‐Kosaki Syndrome in an African Patient.” Clinical Case Reports 9, no. 4: 2144–2148. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Joseph, L. , Cankovic M., Caughron S., et al. 2016. “The Spectrum of Clinical Utilities in Molecular Pathology Testing Procedures for Inherited Conditions and Cancer: A Report of the Association for Molecular Pathology.” Journal of Molecular Diagnostics 18, no. 5: 605–619. [DOI] [PubMed] [Google Scholar]
- Kamp, M. , Krause A., and Ramsay M.. 2021. “Has Translational Genomics Come of Age in Africa?” Human Molecular Genetics 30, no. R2: R164–R173. [DOI] [PubMed] [Google Scholar]
- Lavelle, T. A. , Feng X., Keisler M., et al. 2022. “Cost‐Effectiveness of Exome and Genome Sequencing for Children With Rare and Undiagnosed Conditions.” Genetics in Medicine 24, no. 6: 1349–1361. [DOI] [PubMed] [Google Scholar]
- Louw, N. , Carstens N., Lombard Z., and for DDD‐Africa as Members of the H3Africa Consortium . 2023. “Incorporating CNV Analysis Improves the Yield of Exome Sequencing for Rare Monogenic Disorders—An Important Consideration for Resource‐Constrained Settings.” Frontiers in Genetics 14: 1277784. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Lumaka, A. , Carstens N., Devriendt K., et al. 2022. “Increasing African Genomic Data Generation and Sharing to Resolve Rare and Undiagnosed Diseases in Africa: A Call‐To‐Action by the H3Africa Rare Diseases Working Group.” Orphanet Journal of Rare Diseases 17, no. 1: 1–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Manickam, K. , McClain M. R., Demmer L. A., et al. 2021. “Exome and Genome Sequencing for Pediatric Patients With Congenital Anomalies or Intellectual Disability: An Evidence‐Based Clinical Guideline of the American College of Medical Genetics and Genomics (ACMG).” Genetics in Medicine 23, no. 11: 2029–2037. [DOI] [PubMed] [Google Scholar]
- Martin, A. R. , Williams E., Foulger R. E., et al. 2019. “PanelApp Crowdsources Expert Knowledge to Establish Consensus Diagnostic Gene Panels.” Nature Genetics 51, no. 11: 1560–1565. [DOI] [PubMed] [Google Scholar]
- McLaren, W. , Gil L., Hunt S. E., et al. 2016. “The Ensembl Variant Effect Predictor.” Genome Biology 17, no. 1: 122. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Miller, D. T. , Lee K., Chung W. K., et al. 2021. “ACMG SF v3.0 List for Reporting of Secondary Findings in Clinical Exome and Genome Sequencing: A Policy Statement of the American College of Medical Genetics and Genomics (ACMG).” Genetics in Medicine 23, no. 8: 1381–1390. [DOI] [PubMed] [Google Scholar]
- Miller, S. A. , Dykes D. D., and Polesky H. F.. 1988. “A Simple Salting Out Procedure for Extracting DNA From Human Nucleated Cells.” Nucleic Acids Research 16, no. 3: 1215. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Richards, S. , Aziz N., Bale S., et al. 2015. “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.” Genetics in Medicine 17, no. 5: 405–424. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Robinson, J. T. , Thorvaldsdóttir H., Winckler W., et al. 2011. “Integrative Genomics Viewer.” Nature Biotechnology 29, no. 1: 24–26. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Schofield, D. , Rynehart L., Shresthra R., White S. M., and Stark Z.. 2019. “Long‐Term Economic Impacts of Exome Sequencing for Suspected Monogenic Disorders: Diagnosis, Management, and Reproductive Outcomes.” Genetics in Medicine 21, no. 11: 2586–2593. [DOI] [PubMed] [Google Scholar]
- Seymour, H. , Feben C., Nevondwe P., et al. 2024. “Mutation Profiling in South African Patients With Cornelia de Lange Syndrome Phenotype.” Molecular Genetics & Genomic Medicine 12, no. 1: e2342. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Smith, L. , Malinowski J., Ceulemans S., et al. 2022. “Genetic Testing and Counseling for the Unexplained Epilepsies: An Evidence‐Based Practice Guideline of the National Society of Genetic Counselors.” Journal of Genetic Counseling 32, no. 2: 266–280. [DOI] [PubMed] [Google Scholar]
- Souche, E. , Beltran S., Brosens E., et al. 2022. “Recommendations for Whole Genome Sequencing in Diagnostics for Rare Diseases.” European Journal of Human Genetics 30, no. 9: 1017–1021. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Strande, N. T. , Riggs E. R., Buchanan A. H., et al. 2017. “Evaluating the Clinical Validity of Gene‐Disease Associations: An Evidence‐Based Framework Developed by the Clinical Genome Resource.” American Journal of Human Genetics 100, no. 6: 895–906. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data S1: Provides the design of the IDP targeted sequencing panel.
Table S1: Lists the genes included in the IDP panel, together with relevant annotations.
Data Availability Statement
The variants described here were submitted to ClinVar and can be viewed under Organization ID 508172. The gene panel design is shared in Ion Reporter format in the Supporting Information.