Next-Generation Sequencing

Next generation sequencing (NGS, massively parallel sequencing) reads millions to billions of DNA fragments simultaneously, with library preparation (fragmentation, adaptor ligation, target enrichment by hybrid capture or amplicon), sequencing by synthesis (short reads of 75 to 300 bases) or nanopore and single molecule long read technology (10 to 100 kilobases or longer), alignment to a reference genome, variant calling and annotation.

Clinical applications range from gene panels (10 to 500 genes), clinical exome (about 20 000 genes, 1 to 2% of the genome) and genome sequencing (3 billion bases), to RNA sequencing, tumour profiling, circulating tumour DNA, metagenomic pathogen detection and microbiome studies.

Coverage depth (average reads over each base) is 30 times for genomes, 100 to 150 times for exomes and 500 to 1000 times (or more) for tumour and low level variant detection.

Diagnostic yield depends on the phenotype: 25 to 40% for exome in neurodevelopmental disorders and congenital anomalies, 30 to 50% in selected epilepsies, cardiomyopathy and inherited cancer panels (10 to 15% in unselected breast cancer), and 40 to 60% in rapid genome sequencing in critically ill infants (changing management in 30 to 60%).

In oncology, NGS identifies actionable alterations (EGFR, ALK, KRAS G12C, BRAF V600E, HER2, NTRK, MSI-high, tumour mutational burden) for targeted therapy and immunotherapy. In infectious disease, it identifies resistance mutations and outbreak clusters, and metagenomics detects unculturable or unexpected pathogens.

Indications. Suspected monogenic disease with genetic heterogeneity (neurodevelopmental disorders, epilepsy, cardiomyopathy, arrhythmia, inherited retinal disease, hearing loss, skeletal dysplasia, immunodeficiency, metabolic disease). Undiagnosed disease after non-diagnostic microarray or single gene testing. Critically ill infants with suspected genetic disease (rapid genome).

Hereditary cancer syndromes (BRCA1, BRCA2, mismatch repair genes, TP53). Tumour profiling in advanced cancer and for treatment selection. Haematological malignancy classification and risk. Minimal residual disease and cell free DNA. Pathogen identification and typing. Pharmacogenomics. Reproductive carrier screening and preimplantation testing.

Not for routine screening of healthy people without counselling.

Specimen. Peripheral blood 2 to 5 mL in EDTA (DNA extraction 1 to 3 micrograms), saliva or buccal swab;

tumour tissue (formalin fixed paraffin embedded blocks with at least 10 to 20% tumour cells, pathologist marked, fixed in neutral buffered formalin for 6 to 48 hours), fresh frozen tissue, bone marrow, plasma for cell free DNA (10 mL blood in cell stabilising tubes, processed within 3 to 7 days), cerebrospinal fluid and other fluids for metagenomics, dried blood spots, amniotic fluid and chorionic villi.

Trio samples (proband and both parents) increase yield of de novo and compound heterozygous variants. Consent, phenotype description in standardised terms, and family history are required.

Reference values. No pathogenic or likely pathogenic variant relevant to the clinical question, with adequate coverage of target regions (for example, 95 to 98% of target bases at 20 times depth or more). Results are classified in five tiers: pathogenic, likely pathogenic, variant of uncertain significance (VUS), likely benign and benign, using criteria that combine population frequency, computational prediction, functional data, segregation, de novo status and phenotype match.

Interpretation.

• Pathogenic or likely pathogenic variant in a gene matching the phenotype with the right inheritance pattern (heterozygous for dominant, homozygous or compound heterozygous for recessive, hemizygous for X-linked): molecular diagnosis, with recurrence risk counselling and targeted management.
• Variant of uncertain significance: not actionable; present in 20 to 40% of tests; parental studies, segregation, functional work and periodic re-analysis can reclassify (10 to 15% reclassified over time).
• Negative result: does not exclude genetic disease (about 50 to 75% of exome negative cases remain unexplained), because of variants in non-coding regions, structural variants, repeat expansions, mitochondrial DNA, mosaicism, poor coverage, and novel genes.
• Secondary (incidental) findings: pathogenic variants in actionable genes (for example, 80 to 100 genes, such as BRCA1, BRCA2, Lynch syndrome genes, familial hypercholesterolaemia genes, cardiomyopathy genes) are reported only with consent.
• Tumour profiling: variant allele frequency indicates clonality (germline variants cluster near 50% or 100%); actionable variants are matched to approved or investigational therapies; tumour mutational burden and microsatellite instability predict immunotherapy response.
• Circulating tumour DNA: detection of mutations in plasma shows tumour burden and resistance mutations (EGFR T790M), with sensitivity about 60 to 80% and specificity over 95%.
• Metagenomic sequencing: reads matching pathogens, interpreted with abundance, negative controls and clinical context; contamination is common.
• Copy number variants, structural variants and loss of heterozygosity can be called from exome and genome data.
• Mosaic variants appear with allele fractions under 30% and need deep coverage.

Critical values and findings needing immediate action.

• Rapid genome sequencing in a critically ill infant: results within 2 to 5 days to direct treatment (for example, treatable metabolic disease, immunodeficiency, mitochondrial disease), redirection of care, or transplantation.
• Pathogenic variant in an actionable cardiac gene (long QT, hypertrophic cardiomyopathy, Marfan syndrome) or cancer predisposition gene: family cascade testing, surveillance and preventive measures.
• Actionable tumour alteration (EGFR, ALK, ROS1, BRAF V600E, NTRK, HER2, KRAS G12C, RET): targeted therapy.
• Resistance mutations in infectious disease sequencing (Mycobacterium tuberculosis, HIV): adjust treatment.
• Incidental finding of a high risk cancer predisposition variant: genetic counselling, surveillance and family testing.
• Pre-implantation or prenatal finding of a serious condition: genetic counselling and informed decision making.

Interferences and pitfalls.

• Short read sequencing has limits in repetitive regions, pseudogenes (for example SMN1/SMN2, CYP21A2, PMS2, GBA), GC-rich regions, large structural variants and repeat expansions; long read sequencing and targeted tests are used.
• Poor DNA quality (formalin damage causing C to T artefacts), low tumour content, contamination, sample swaps and index hopping lead to errors; quality control metrics (coverage, uniformity, contamination, sex check) are required.
• Variant interpretation is subject to bias in reference populations (under representation of African, Asian and other ancestries) and reclassification over time; reports should state date and database versions.
• Clonal haematopoiesis in older adults produces blood variants (DNMT3A, TET2, ASXL1) that may be mistaken for tumour or germline variants in cell free DNA.
• Germline versus somatic distinction needs matched normal samples.
• Incidental findings, privacy, data sharing, insurance issues and the right not to know require consent processes, and results in children raise ethical questions about adult onset conditions.
• Turnaround (2 to 8 weeks for exome, 24 hours to 5 days for rapid genome, 7 to 14 days for tumour panels) and cost limit access, particularly in low- and middle income settings.
• Bioinformatics pipelines, validation, storage of large data and clinical expertise (multidisciplinary review) are essential.
• Over testing, such as large panels without a defined phenotype, increases uncertain findings.

Special groups.

• Children with developmental delay and congenital anomalies: exome or genome as first line test after microarray in many centres.
• Critically ill neonates: rapid whole genome sequencing in neonatal and paediatric intensive care.
• Patients with cancer: germline and somatic testing together, with reflex testing on histology.
• Families with consanguinity: homozygosity mapping combined with exome.
• Prenatal: exome in foetuses with structural anomalies and normal microarray (yield 8 to 10%, 30% with skeletal or neurological anomalies).
• Regions: sequencing capacity is expanding in Africa, Asia and Latin America with local genomic initiatives, but reference data from these populations are limited.

Follow up tests. Segregation analysis in parents and relatives, Sanger confirmation of key variants, copy number and structural variant analysis (microarray, MLPA, long read), repeat expansion testing, mitochondrial DNA sequencing, RNA sequencing for splice variants, functional and biochemical tests (enzyme assays, metabolic screen), multidisciplinary review,

cascade testing of at risk relatives, genetic counselling before and after testing, periodic re-analysis of negative or uncertain results, and research study referral for actionable tumour variants.