Germline Classification for Confident Variant Interpretation

It takes, on average, 5.6 years for a rare disease patient to receive a diagnosis^1,2. When a diagnosis is reached it can significantly impact treatment options, lifestyle choices, and disease management. For some, a diagnosis might mean a curative treatment. However, for the 95% of rare diseases that don’t have an approved therapy², receiving an accurate diagnosis is about clarity, guiding appropriate medical care, and enabling patients and their families to make informed decisions about their future. Rare diseases collectively affect over 300 million people worldwide, yet each condition is uncommon, often impacting fewer than 1 in 2,000 individuals². This rarity makes accurate diagnosis challenging, frequently leading to diagnostic delays and uncertainty for patients and clinicians.

Genetic variant interpretation is a critical part of rare disease diagnosis. An estimated 80% of rare diseases have a genetic basis², making variant classification essential for accurate decision-making. However, genetic interpretation does not just inform patient care; it drives research. Identifying and classifying disease-associated variants helps us understand disease mechanisms, inform the development of targeted therapies, and link eligible patients with clinical trials.

The VarSome Germline Classifier, based on the American College of Medical Genetics and Genomics (ACMG) Variant Classification Guidelines³, helps ensure that variant interpretation is systematic, evidence-based, and reproducible, supporting clinical decision-making and genetic research.

Here, we discuss the application of germline classification in the context of three rare diseases, each illustrating a different aspect of this interpretive framework:

1. Fibrodysplasia Ossificans Progressiva – Demonstrating step-by-step germline classification.
2. Alkaptonuria – How evidence strength impacts classification confidence.
3. Hutchinson-Gilford Progeria Syndrome – Resolving conflicting evidence.

Through these examples, we highlight the importance of consistent and transparent variant interpretation in rare disease diagnosis to support more accurate and meaningful clinical outcomes.

Germline Classification in Action

The germline classifier categorizes genetic variants into five classes based on the strength of evidence supporting their impact on disease. Pathogenic variants are confidently linked to disease and directly contribute to its development. Likely Pathogenic variants have strong but not definitive evidence of disease association. Variants of Uncertain Significance (VUS) have insufficient or conflicting evidence, making their clinical impact unclear. Likely Benign variants are unlikely to cause disease but lack definitive proof, while Benign variants are well-documented to have no impact on disease. This classification system provides a structured framework for consistent and evidence-based variant interpretation.

The ACVR1(NM_001111067.4):c.617G>A variant is the most common cause of FOP and is well-known to be pathogenic. This classification is determined using the ACMG guidelines, which apply a structured framework to evaluate multiple lines of evidence. By systematically applying the relevant ACMG rules, we can see how the available data supports a Pathogenic classification, ensuring a consistent and transparent interpretation.

Five rules have been triggered for the ACVR1(NM_001111067.4):c.617G>A variant: PS3, PP3, PM1. PP5, PM2. Descriptions for these rules can be found in Table 1. For a full list of the germline classification rules, see VarSome Germline Classification. 

As these are the only rules triggered, by the ACMG guidelines, this variant is classified as Pathogenic. This has significant implications for how clinicians make decisions and manage patient care. While automation streamlines variant classification, expert interpretation remains essential. The germline classifier is designed to assist, not replace, genetic professionals - gathering evidence, applying ACMG rules, and structuring results to support confident decision-making.

Clinical Implications

This comprehensive evidence base helps clinicians have confidence in their clinical assessment. Confident interpretation is crucial for clinical decision-making in rare diseases like FOP for several reasons:

• It supports a definitive genetic diagnosis, confirming the cause of FOP in affected individuals and eliminating diagnostic uncertainty. This allows clinicians to take proactive measures to manage symptoms, avoid trauma-induced flare-ups, and offer tailored patient care.
• It informs genetic counseling by establishing the autosomal dominant inheritance pattern of the disease associated with this variant, enabling informed risk assessments for family members.
• It supports treatment decisions, including available approved therapies like Palovarotene (Sohonos), and helps determine eligibility for clinical trials investigating novel interventions.

By applying a systematic, evidence-based framework, this classification provides clinicians with the confidence needed to make well-informed, precise medical decisions, ultimately leading to better patient outcomes.

How Does Evidence Strength Impact Classification?

In this section, we will examine the classification of three HGD variants to see how evidence strength impacts its pathogenicity classification and how this influences clinical decision-making.

Assigning Pathogenicity Scores Based on Evidence Strength

The first variant we’ll look at is HGD(NM_000187.4):c.87+1G>A.

As shown above, three rules have been triggered for this variant, resulting in a Pathogenic classification with a score of 11 points. 

Pathogenic Evidence:

• PVS1 (Very Strong, +8 points): This is a null variant occurring within +2 of the splice site, in a gene where loss-of-function is a well-established disease mechanism. The HGD gene has 112 reported pathogenic loss-of-function variants, supporting the conclusion that disruption at this site is highly likely to be disease-causing.
• PP5 (Moderate, +2 points): ClinVar classifies this variant as Pathogenic with a 2-star review, based on submissions from multiple high-confidence sources. Additionally, LOVD classifies the variant as Pathogenic, reinforcing expert consensus on its clinical significance.
• PM2 (Supporting, +1 point): The variant is absent from population databases, including gnomAD genomes and exomes, suggesting that it is not a common benign polymorphism.

The pathogenicity score is calculated through a weighted analysis of each triggered rule, with each rule assigned a score based on the strength of evidence provided. The total score is calculated as the sum of the scores from pathogenic rules minus the sum of the scores from benign rules (Table 2).

The HGD(NM_000187.4):c.87+1G>A. variant is supported by multiple lines of evidence, including the triggering of PVS1, which strongly supports its Pathogenic classification. Let’s compare this with another variant. One with a less definitive classification:  HGD(NM_000187.4):c.946G>T.

As shown above, this variant triggers more rules; however, these rules are supported by cumulatively less robust evidence, resulting in a score of 9 and a Likely Pathogenic classification. 

In this case, the following rules contribute to the overall score:

• PP3 (Strong, +4 points): Computational predictions strongly suggest pathogenicity, with a MetaRNN score of 0.947, which is above the threshold for strong evidence. This high score indicates a significant likelihood of impacting protein function.
• PM1 (Moderate, +2 points): The variant is located in a mutational hotspot of 17 amino acids with a high density of pathogenic variants (5 pathogenic, 4 uncertain, and no benign variants). This suggests the region is sensitive to pathogenic changes.
• PM5 (Moderate, +2 points): An alternative variant at the same residue (p.V316I) is classified as Pathogenic in ClinVar, although it is listed as Uncertain Significance in the germline classifier. This suggests that changes in this position are likely to impact protein function, providing moderate evidence.
• PM2 (Supporting, +1 point): The variant is absent from population databases, including gnomAD. This supports its classification as potentially pathogenic, although rarity alone is not sufficient to confirm pathogenicity. 

A score of 9 places this variant within the Likely Pathogenic category, with insufficient evidence to meet the threshold for Pathogenic classification. Specifically, the absence of very strong functional evidence (PS3) and conflicting classifications for similar variants prevent a more definitive interpretation.

So far, we’ve seen how strong and consistent evidence leads to Pathogenic and Likely Pathogenic classifications. But, not all variants provide such clear conclusions. When the available evidence is limited, conflicting, or lacks robustness, an Uncertain Significance classification is assigned.

Let’s look at HGD(NM_000187.4):c.733G>T as an example where multiple rules are triggered, but the evidence is not definitive enough to support a pathogenic or benign classification.

The figure above shows that this variant is classified as Uncertain Significance, with a total score of 5. Although multiple pathogenic rules are triggered, the supporting evidence is cumulatively weak and inconsistent therefore, its contribution to disease cannot be confirmed.

In this case, the score is based on the following rules:

• PP3 (Moderate, +2 points): Computational predictions suggest a moderate likelihood of pathogenicity, with a MetaRNN score of 0.92 However, this is below the threshold for strong evidence.
• PM2 (Supporting, +1 point): The variant is absent from population databases, including gnomAD, suggesting it is rare. However, rarity does not confirm pathogenicity.
• PP2 (Supporting, +1 point): The variant affects a missense residue where over 96% of non-VUS missense variants in the HGD gene are pathogenic. Although this suggests a trend, it does not confirm pathogenicity for this variant.
• PP5 (Supporting, +1 point): There is supporting evidence from external databases, including LOVD and UniProt, that list this variant as pathogenic. However, the strength of this evidence is limited, and it lacks high-confidence expert reviews.

Although these rules cumulatively lead to a score of 5, which places it at the upper limit of the VUS category, the evidence is neither strong nor consistent enough to confidently classify it as Likely Pathogenic. This illustrates how the germline classifier carefully balances multiple lines of evidence, ensuring that only robust and consistent data result in definitive classifications. However, classification is not the same as interpretation. While the classifier aggregates data, applies ACMG criteria and structures the results, it is ultimately the clinician or researcher who assesses the broader clinical context and determines the significance of the findings.

Clinical Implications

The Uncertain Significance classification means clinicians should interpret this result with caution. In clinical practice, this generally leads to conservative decision-making, avoiding unnecessary interventions or changes in patient management. Importantly, a VUS classification should not be used in isolation for diagnostic or predictive purposes.

However, this classification can still provide value by guiding further investigations, such as functional studies or segregation analysis within affected families. Additionally, as new data emerges—such as population frequency updates or new functional evidence—the classification may be revisited, potentially leading to reclassification as Likely Pathogenic or Likely Benign.

While Likely Pathogenic variants can support diagnostic hypotheses and influence treatment decisions when correlated with clinical findings, Uncertain Significance classifications require a more cautious approach. A VUS should not be used as the basis for clinical decisions but rather as an indicator that further research or family studies may be necessary to clarify its significance. This distinction is critical in genetic counseling and patient management, ensuring that only well-supported genetic findings guide clinical interventions.

This example demonstrates the importance of maintaining clinical caution and highlights the dynamic nature of variant interpretation, where new evidence can shift the classification and its clinical implications.

By examining these three HGD variants, we can see how the strength and consistency of evidence directly influence germline classification. This structured approach ensures that only robust and reliable data lead to definitive clinical interpretations, highlighting the importance of comprehensive evidence evaluation in variant classification.

Resolving Conflicting Evidence in Classification 

This section will show how the LMNA(NM_170707.4):c.1824C>T variant is classified as Pathogenic, despite conflicting computational predictions that suggest a benign effect. This case highlights the importance of functional evidence and expert curation in resolving ambiguous classifications, ensuring that clinically significant variants are correctly interpreted for informed decision-making.

Balancing Pathogenic and Benign Evidence

The LMNA(NM_170707.4):c.1824C>T variant is an example of how the germline classification framework balances conflicting evidence to reach a final pathogenicity assessment (Figure 5). Despite being a synonymous variant, which typically do not alter protein function, this variant is classified as Pathogenic with a total score of 10. This classification is driven by strong pathogenic evidence that outweighs the benign computational prediction (BP4).

Evidence Supporting Pathogenicity:

• PS3 (Very Strong, +8 points): Functional studies have demonstrated that this variant affects RNA splicing, leading to an altered protein product that contributes to HGPS.
• PP5 (Moderate, +2 points): ClinVar and LOVD classify this variant as Pathogenic, with multiple supporting literature references and expert-reviewed submissions reinforcing its disease association.
• PM2 (Supporting, +1 point): The variant is absent from population databases, including gnomAD, supporting its pathogenicity.

Conflicting Benign Evidence:

• BP4 (Supporting, -1 point): Computational tools predict that, as a synonymous variant, this change should not affect protein function. Additionally, conservation-based metrics suggest that this position is not highly conserved across species, which typically lowers the likelihood of pathogenicity.

Clinical Implications

Here, we have highlighted a crucial aspect of germline classification: computational predictions alone are not enough to determine pathogenicity. While BP4 suggests a benign effect, the strong functional evidence (PS3) and expert-reviewed pathogenic classifications (PP5) outweigh computational predictions, leading to a Pathogenic classification.

This example reinforces the importance of functional validation and expert curation in variant interpretation. The classifier streamlines the evaluation process by aggregating data and applying ACMG criteria, but final interpretation requires expert review to consider clinical context and determine the true impact of a variant. Without functional studies, this variant might have been misclassified as Likely Benign or Uncertain Significance based on computational predictions alone.

This highlights the need for a structured, evidence-based approach to variant classification, ensuring that clinically significant findings are not overlooked due to conflicting in silico predictions. It also demonstrates how the classifier enhances consistency and efficiency while keeping expert oversight at the core of variant interpretation. As our understanding of variant impact continues to advance, cases like LMNA(NM_170707.4):c.1824C>T highlight the dynamic nature of genetic interpretation and the necessity of ongoing research to refine classification frameworks.

Conclusions

Accurate genetic variant classification is fundamental to rare disease diagnosis and clinical decision-making. The ACMG-driven classification framework discussed here provides a structured, evidence-based approach to interpreting genetic data, ensuring that classifications remain consistent, transparent, and clinically meaningful. Through three rare diseases, we have shown how this framework is applied across different classification challenges.

Key Takeaways:

The ACVR1 variant involved in FOP (ACVR1(NM_001111067.4):c.617G>A) illustrated how ACMG rules are systematically applied to classify a well-documented Pathogenic variant, reinforcing the importance of a structured approach.

The three variants associated with Alkaptonuria showed how the strength and breadth of evidence impact classification, distinguishing between Pathogenic, Likely Pathogenic, and Uncertain Significance variants. This highlights the need for comprehensive evidence evaluation in genetic interpretation.

The variant associated with HGPS (LMNA(NM_170707.4):c.1824C>T) demonstrated how the classifier integrates both pathogenic and benign criteria, ensuring that strong functional evidence (splicing studies) outweighs computational predictions that suggest a benign effect. This underscores the importance of functional validation and expert curation in resolving ambiguous cases.

These examples illustrate the challenges of variant classification, particularly in rare, severe, and largely untreatable conditions like FOP, Alkaptonuria, and HGPS. While many rare diseases now have emerging treatments, others remain difficult to manage, with limited therapeutic options. For these conditions, an accurate genetic diagnosis offers clarity to patients and their families, informing clinical care, and guiding future research efforts.

While well-characterized variants benefit from established classifications, many rare disease cases involve novel or poorly understood variants. In these cases, the classifier helps structure and weigh the available evidence, providing a framework for expert review. This approach ensures that even in cases of limited data, variant classification remains systematic and transparent. Often, a clinician may only come across a particular variant once in their career. A platform like VarSome makes it possible to share expertise and knowledge concerning disease-causing rare variants, around the world. Particularly when it comes to newly identified or rarely seen variants. 

We have highlighted the critical role of variant interpretation in conditions where clinical decisions rely heavily on genetic evidence. In the absence of curative treatments, precise classification ensures that patients receive the right information, appropriate management, and access to genetic counseling or clinical trials.

On Rare Disease Day, it is essential to recognize that every rare disease diagnosis matters, whether it leads to an immediate treatment or simply provides long-awaited answers. Germline classification plays an important role in ensuring that variant classification is consistent, transparent, and where possible, actionable, helping clinicians, researchers, and families navigate the challenges of rare disease genetics.

References

1. NHS. Diagnostic odyssey in rare disease. National Genomics Education Programme. February 13, 2024. Accessed February 19, 2025. https://www.genomicseducation.hee.nhs.uk/genotes/knowledge-hub/the-diagnostic-odyssey-in-rare-disease/
2. The Lancet Global Health. The landscape for rare diseases in 2024. Lancet Glob Health. 2024;12(3):e341. doi:10.1016/S2214-109X(24)00056-1
3. 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(5):405-424. doi:10.1038/gim.2015.30
4. National Organization for Rare Disorders (NORD). Fibrodysplasia Ossificans Progressiva. Rare Diseases. June 26, 2024. Accessed February 19, 2025. https://rarediseases.org/rare-diseases/fibrodysplasia-ossificans-progressiva/
5. Wang RN, Green J, Wang Z, et al. Bone Morphogenetic Protein (BMP) signaling in development and human diseases. Genes Dis. 2014;1(1):87-105. doi:10.1016/j.gendis.2014.07.005
6. National Organization for Rare Disorders (NORD). Alkaptonuria. Rare Diseases. June 26, 2017. Accessed February 19, 2025. https://rarediseases.org/rare-diseases/alkaptonuria/
7. NHS. Alkaptonuria. Health A - Z. March 10, 2022. Accessed February 19, 2025. https://www.nhs.uk/conditions/alkaptonuria/
8. National Organization for Rare Disorders (NORD). Hutchinson-Gilford Progeria Syndrome. Rare Diseases. January 4, 2021. Accessed February 19, 2025. https://rarediseases.org/rare-diseases/hutchinson-gilford-progeria/
9. Lamis A, Siddiqui SW, Ashok T, Patni N, Fatima M, Aneef AN. Hutchinson-Gilford Progeria Syndrome: A Literature Review. Cureus. Published online August 31, 2022. doi:10.7759/cureus.28629
10. Progeria Research Foundation. Progeria 101 / FAQ. 2025. Accessed February 20, 2025. https://www.progeriaresearch.org/progeria-101faq/