Next-generation sequencing has fundamentally changed how laboratories approach genetic disease. Where single-gene Sanger sequencing once dominated clinical diagnostics, labs can now sequence thousands of genes simultaneously — or an entire genome — in a matter of days.
But sequencing the genome is only half the battle.
The raw output of an NGS run is a massive file of short DNA reads. Before a clinician can act on any of it, that data must be aligned to a reference genome, variants must be called, annotated, filtered, classified, and interpreted against clinical guidelines. This transformation from raw data to diagnostic answer is called tertiary analysis — and it is where the real complexity of modern genomics lives.
This is precisely where clinical genomics software has become indispensable.
What Is Clinical Genomics Software?
Clinical genomics software refers to integrated platforms designed to process, annotate, interpret, and report genomic variants in a clinical diagnostic setting. Unlike research-grade bioinformatics tools, clinical genomics software is built to meet the regulatory and reproducibility demands of a diagnostic laboratory.
A robust platform handles the following steps of the NGS workflow:
- Variant annotation: Mapping each variant to known databases (ClinVar, gnomAD, OMIM) and adding biological context
- Variant filtration: Applying rule-based filters to remove benign background noise and surface clinically relevant candidates
- Variant classification: Assigning pathogenicity using frameworks like ACMG/AMP guidelines for germline variants, or AMP/ASCO/CAP guidelines for somatic oncology variants
- CNV analysis: Detecting copy number variants alongside small variants within the same workflow
- Clinical reporting: Generating structured, audit-ready reports that can be delivered to ordering physicians
The distinction between research tools and clinical genomics software comes down to validation, reproducibility, and compliance. A clinical lab cannot simply use whichever bioinformatics script produces interesting results — they need a validated system that produces the same output consistently, across analysts, runs, and months of operation.
The Three Phases of NGS Analysis
To appreciate why purpose-built clinical software matters, it helps to understand how NGS data moves through a laboratory.
Primary Analysis
The sequencing instrument converts optical signals into base calls, producing FASTQ files — raw sequence reads with quality scores. This happens inside the sequencer itself.
Secondary Analysis
Bioinformatics pipelines align those reads to the human reference genome and call variants, producing a VCF (Variant Call Format) file. This step identifies where a patient’s genome differs from the reference. High-performance secondary analysis tools, such as those powered by Sentieon algorithms, can process whole genomes in hours rather than days.
Tertiary Analysis
This is where clinical decision-making begins. The VCF file may contain hundreds of thousands of variant positions. Tertiary analysis software annotates each one, applies clinical filters, runs classification algorithms, and surfaces the small subset of variants that are diagnostic candidates. This step requires deep integration of clinical knowledge databases, guideline-compliant classification logic, and laboratory-specific filter workflows.
Most labs find tertiary analysis to be the largest operational bottleneck — not because the biology is unsolvable, but because the tooling has historically been fragmented, requiring analysts to move between spreadsheets, databases, and disconnected software tools.
ACMG Classification and Why It Matters
One of the most critical functions of clinical genomics software is automating ACMG classification — the scoring framework published by the American College of Medical Genetics and Genomics for classifying germline variants as Pathogenic, Likely Pathogenic, Variant of Uncertain Significance (VUS), Likely Benign, or Benign.
Manual ACMG classification requires a trained analyst to evaluate a variant against 28 criteria, each weighted to contribute to a final pathogenicity score. For a laboratory processing hundreds of cases per month, this is an enormous burden.
Software-driven ACMG classification automates the application of these criteria using integrated annotation data — allele frequencies, computational predictions, functional evidence, and curated clinical databases. When implemented correctly, it dramatically reduces analyst time while improving consistency and audit traceability.
This is not about removing clinical judgment. Human review remains essential, particularly for variants of uncertain significance. But automation ensures that straightforward cases are handled efficiently, freeing analysts to focus their expertise on the genuinely ambiguous ones.
CNV Analysis: The Missing Piece
For many years, clinical NGS panels were designed to detect small variants — single nucleotide variants (SNVs) and small insertions or deletions (indels). Copy number variants, which involve deletions or duplications of larger genomic segments, were typically detected through orthogonal methods such as chromosomal microarray.
Modern clinical genomics software has changed this. By applying statistical models to read depth and coverage data from standard NGS runs, platforms can now call CNVs directly from gene panel, exome, and whole genome data — without requiring a separate assay.
The clinical impact is significant. Many hereditary conditions, including hereditary breast and ovarian cancer (BRCA1/2 large deletions), spinal muscular atrophy (SMN1), and Duchenne muscular dystrophy (DMD), have a meaningful proportion of pathogenic variants that are CNVs rather than point mutations. A lab that cannot detect CNVs from its NGS data is returning incomplete results.
Integrating CNV analysis into the same tertiary analysis workflow — where CNV candidates appear alongside SNVs and indels, annotated and classified in a unified interface — represents a significant clinical and operational improvement over running parallel workflows.
Somatic Analysis and Precision Oncology
Germline testing looks for inherited variants present in every cell of the body. Somatic analysis is fundamentally different: it looks for variants that arose in a tumor, distinguishing them from the patient’s germline background.
This requires different algorithms, different databases, and different classification frameworks. Somatic variant interpretation follows AMP/ASCO/CAP guidelines, with variants classified by their therapeutic, prognostic, or diagnostic significance — Tier I through Tier IV.
Clinical genomics software built for oncology must handle low variant allele frequencies (VAFs), tumor heterogeneity, and the need to cross-reference variants against curated oncology knowledgebases that connect specific alterations to FDA-approved therapies, clinical trial eligibility, and prognostic evidence.
The most advanced platforms now include automated oncogenicity classification algorithms that assess somatic variants against curated evidence, substantially reducing the manual review burden in high-volume cancer testing programs.
Exome and Genome Analysis at Clinical Scale
Gene panels covering dozens or hundreds of genes remain the workhorse of most clinical NGS programs. But exome sequencing — covering all ~20,000 protein-coding genes — is increasingly used for rare disease diagnosis, particularly in pediatric and prenatal settings. Whole genome sequencing is expanding further, adding non-coding regulatory regions and enabling detection of structural variants and tandem repeat expansions.
Clinical genomics software must scale across all of these assay types without requiring a different tool for each workflow. The same platform should handle a 50-gene hereditary cancer panel, a whole exome for a rare disease patient, and a whole genome for a prenatal case — with assay-specific filter workflows that surface the right variants in each context.
For enterprise genomics programs, scalability also means managing data across hundreds or thousands of samples. Genomic data management infrastructure — such as enterprise-grade warehousing systems — allows labs to store, search, and re-analyze variant data as databases are updated and guidelines evolve.
Deployment and Data Sovereignty
Clinical laboratories handling patient genomic data operate under strict regulatory frameworks — HIPAA in the United States, GDPR in Europe, and equivalent national regulations elsewhere. Genomic data is among the most sensitive personal information that exists; unlike a password, a person’s genome cannot be changed if it is compromised.
This has implications for how clinical genomics software is deployed. Many institutions require that patient data remain within their own infrastructure — on-premises servers, private cloud environments, or air-gapped networks. A software platform that requires patient data to pass through a vendor’s servers is, for many labs, simply not an option.
Software built for clinical deployment must support local and private cloud installation, with network communication limited to license verification and annotation database updates rather than patient data transmission.
Regulatory certification adds another dimension. Platforms seeking use in medical device contexts in Europe must meet IVDR 2017/746 requirements, including quality management system certification under ISO 13485. CE-marked genomic analysis software provides a higher level of regulatory assurance for diagnostic use.
Choosing the Right Clinical Genomics Software
For laboratory directors and bioinformatics leads evaluating platforms, several criteria consistently separate purpose-built clinical solutions from adapted research tools:
Guideline integration. The platform should implement current ACMG/AMP and AMP/ASCO/CAP classification frameworks directly, with traceable logic that can be audited during inspections.
Unified variant types. SNVs, indels, CNVs, and structural variants should be reviewable within a single interface, not across separate tools that must be manually reconciled.
Annotation currency. Clinical knowledge databases change rapidly — ClinVar submissions are added daily, therapeutic approvals shift the landscape of somatic interpretation. The platform’s annotation infrastructure should update regularly and transparently.
Workflow repeatability. The same filter workflow should produce the same output on the same data, every time, regardless of which analyst runs it. This is a basic requirement for clinical validation.
Reporting integration. Clinical reports should be generated from within the platform, incorporating variant interpretations, ACMG classifications, supporting evidence, and laboratory-specific language — without requiring export to a separate reporting tool.
Deployment flexibility. The platform should support the institution’s infrastructure requirements, whether that means on-premises workstations, private cloud, or high-throughput server deployment.
The Direction of Clinical Genomics
NGS has moved from a research instrument to a frontline clinical tool with remarkable speed. Whole genome sequencing is now a first-tier diagnostic test in some national healthcare systems. Pharmacogenomics programs are scaling rapidly as evidence for gene-drug interactions accumulates. Prenatal genomics is expanding beyond chromosomal microarray.
At each step of this expansion, the limiting factor is not sequencing throughput — it is the capacity of clinical laboratories to interpret the variants that sequencing produces.
Clinical genomics software is the infrastructure that bridges this gap. Platforms like Golden Helix’s VarSeq suite — which integrates tertiary analysis, ACMG classification, CNV analysis, somatic interpretation, and clinical reporting within a single validated environment — represent the operational backbone of modern genomic medicine.
As testing volumes grow and assay complexity increases, the difference between a clinical program built on purpose-built genomics infrastructure and one built on fragmented tools will only become more apparent. The labs that invest in scalable, validated clinical genomics software today are building the foundation for genomic medicine at scale.
