# Functional Assays for Variant Interpretation: How to Evaluate Experimental Evidence for ACMG/AMP Criteria (PS3/BS3) Clinical variant interpretation requires integrating multiple evidence types to classify sequence variants. The American College of Medical Genetics and Genomics (ACMG) and Association for Molecular Pathology (AMP) framework assigns strong evidence codes PS3 and BS3 to well-established functional assays that demonstrate abnormal or normal gene or protein function. The original guidelines did not define how to evaluate whether a functional assay qualifies as well-established, and inconsistent application of PS3/BS3 contributes to discordance between laboratories. This article provides a structured approach for researchers and clinical scientists to critically assess functional studies, determine appropriate evidence strength, and apply PS3 or BS3 codes in variant classification. The framework follows recommendations from the Clinical Genome Resource (ClinGen) Sequence Variant Interpretation (SVI) Working Group, which developed a four-step provisional framework for evaluating functional evidence and calibrating evidence strength based on assay validation. ## The Problem of Discordant Functional Evidence Application Variant classification laboratories frequently disagree on whether a functional assay supports pathogenicity or benignity. The ClinGen SVI Working Group identified differences in PS3/BS3 application as a major contributor to variant interpretation discordance between laboratories. The original ACMG/AMP guidelines stated that PS3 should apply when well-established functional studies show a damaging effect and BS3 when well-established functional studies show no damaging effect, but they did not define what constitutes a well-established assay or how to weigh evidence from assays with different validation levels. This ambiguity creates practical problems. A laboratory may apply PS3 to a variant based on a single overexpression assay in a non-physiologic cell line, while another laboratory may require quantitative assays with validated controls before applying any functional evidence. The ClinGen SVI Working Group addressed this by developing recommendations for evaluators regarding the assessment of clinical validity of functional data and a four-step provisional framework to determine the appropriate strength of evidence. The framework requires evaluators to define the disease mechanism, evaluate the applicability of the gene and variant mechanism, assess the assay's validity, and calibrate the strength of evidence based on the number and type of controls tested. For researchers designing functional assays and for laboratory professionals interpreting published functional data, the practical question is whether an assay shows an effect that is specific, quantitative, and relevant to the disease mechanism. A functional assay that demonstrates a measurable difference between wild-type and variant protein activity may support PS3, but only if the assay reflects the variant's mechanism of disease and has been validated with appropriate controls. ## Defining the Disease Mechanism as the First Step The ClinGen SVI Working Group's four-step framework begins with defining the disease mechanism. This step determines whether loss-of-function, gain-of-function, or dominant-negative mechanisms are relevant for the gene in question. The disease mechanism dictates which functional assays are appropriate and how results should be interpreted. For genes where loss-of-function is the established mechanism of disease, assays that measure protein activity, stability, or abundance can provide evidence for pathogenicity when a variant reduces function. For genes with gain-of-function mechanisms, assays that measure increased activity or constitutive activation are more relevant. For dominant-negative mechanisms, assays that measure interference with wild-type protein function may be necessary. The disease mechanism also affects whether the PS3 or BS3 code can be applied at all. For example, in the Alport syndrome genes COL4A3, COL4A4, and COL4A5, the Alport Variant Collaborative noted that well-established functional assays were still mainly research tools, and sequencing and minigene assays were commonly used to confirm splicing variants. The recognition and interpretation of hypomorphic variants in these genes remains a challenge, and the collaborative could not define a minor allele frequency threshold above which variants were considered benign because of different modes of inheritance and the occurrence of hypomorphic variants. When evaluating a functional assay, the first question is whether the assay measures a function that is directly relevant to the disease mechanism. An assay that measures protein-protein interaction may be relevant for a gene where disrupted binding causes disease but irrelevant for a gene where enzymatic activity is the critical function. The ClinGen SVI Working Group recommends that evaluators determine whether the functional assay reflects the gene's mechanism of disease before considering the assay's technical validity. ## Evaluating Assay Applicability to the Variant and Gene The second step in the ClinGen SVI framework is evaluating the applicability of the functional assay to the specific gene and variant under assessment. Not all assays are appropriate for all variant types, and the same assay may have different relevance depending on the variant's location and predicted effect. For missense variants, assays that measure protein function directly are often appropriate. For splice-site variants, RNA-based assays such as minigene assays or RNA sequencing from patient samples are more relevant. The ClinGen SVI Splicing Subgroup addressed this distinction by proposing that PS3/BS3 codes should be applied only for well-established assays that measure functional impact not directly captured by RNA-splicing assays. For splicing evidence, the subgroup proposed repurposing the PVS1_Strength code to capture splicing assay data that provide experimental evidence for variants resulting in RNA transcripts with loss of function, and BP7 to capture RNA results demonstrating no splicing impact for intronic and synonymous variants. This distinction matters for practical variant interpretation. A functional assay that measures protein enzymatic activity may not capture the effect of a variant that disrupts splicing, even if the variant is in a coding region. Conversely, an RNA-based assay may not capture the effect of a missense variant that alters protein function without affecting transcript abundance or splicing. Evaluators must match the assay type to the variant's predicted mechanism. The applicability assessment also considers whether the assay uses the appropriate isoform of the gene. Many genes have multiple transcript isoforms, and a functional assay may measure activity of one isoform while the variant primarily affects another. The assay should use the isoform that is most relevant to the disease phenotype. ## Assessing Assay Validity and Controls The third step in the ClinGen SVI framework is assessing the assay's validity, including whether the assay uses appropriate controls and whether the readout is quantitative. The SVI Working Group used curated functional evidence from ClinGen Variant Curation Expert Panel-developed rule specifications and expert opinions to refine the PS3/BS3 criteria over multiple meetings. They estimated the odds of pathogenicity for assays using various numbers of variant controls to determine the minimum controls required to reach moderate level evidence. The number and type of controls is a critical determinant of evidence strength. An assay that has been validated with multiple known pathogenic and benign variants provides stronger evidence than an assay tested with only a single control. The SVI Working Group's analysis of variant controls informed recommendations for the minimum controls required to reach moderate level evidence, meaning that assays with insufficient controls may only support lower levels of evidence strength. Key control considerations include: - Known pathogenic variants should show abnormal function in the assay - Known benign variants should show normal function in the assay - Wild-type controls should establish the baseline for normal function - Empty vector or mock controls should establish the baseline for no function - The assay should distinguish between pathogenic and benign variants with high sensitivity and specificity An example of rigorous assay validation comes from a study of BRCA1 variants using Green Fluorescent Protein (GFP)-reassembly assays. The researchers assessed the effect of variants in the RING and BRCT domains of BRCA1 on binding with UbcH5a or ABRAXAS proteins. Analysis of 26 clinically classified variants, including 13 tested in a previous study, showed 100% sensitivity and specificity in identifying pathogenic and benign variants for both the RING/UbcH5a and the BRCTs/ABRAXAS interactions. The researchers derived the strength of evidence generated by the GFP-reassembly assays as moderate for both PS3 and BS3 criteria assessment. This example illustrates the standard for assay validation. The assay was tested against a panel of variants with known clinical classification, demonstrated perfect discrimination between pathogenic and benign variants, and the evidence strength was calibrated based on this validation. An assay that has not been validated against known variants cannot support the same level of evidence strength. ## Quantitative Readouts and Threshold Determination Functional assays should produce quantitative readouts that allow comparison between variant and wild-type function. Qualitative assays that simply indicate whether a variant has an effect or not are less informative and harder to calibrate for evidence strength. A quantitative assay measures a continuous variable such as enzymatic activity, protein abundance, binding affinity, or cell viability. The readout for a variant is compared to the readout for wild-type and for known pathogenic and benign controls. The threshold for calling a variant damaging should be established based on the distribution of readouts from known pathogenic and benign variants. The ClinGen SVI Working Group's recommendations emphasize that the strength of evidence should be calibrated based on assay validation. An assay that shows clear separation between pathogenic and benign controls can support strong evidence, while an assay with overlapping distributions between controls may only support moderate or supporting evidence. For practical application, the evaluator should ask: - Is the readout quantitative or qualitative? - How many replicates were performed? - What is the variability between replicates? - Is the threshold for calling a variant damaging pre-specified or determined post hoc? - Does the assay distinguish between variants with different levels of functional impact? Post hoc threshold determination is a common problem in functional assay evaluation. If a researcher analyzes the data and then selects a threshold that classifies the variant of interest as damaging, the evidence is weaker than if the threshold was established before testing the variant. The assay should have a pre-specified threshold based on validation with known controls. ## The Four-Step Framework in Practice The ClinGen SVI Working Group's four-step provisional framework provides a structured approach for determining the appropriate strength of evidence from functional assays. The steps are: 1. Define the disease mechanism 2. Evaluate the applicability of the gene and variant mechanism 3. Assess the assay's validity 4. Determine the appropriate strength of evidence Each step requires specific information and decisions. The framework is designed to be applied by variant curation expert panels and individual laboratories when evaluating functional evidence for variant classification. ### Step 1: Define the Disease Mechanism The evaluator must determine whether the gene causes disease through loss-of-function, gain-of-function, dominant-negative, or another mechanism. This determination may come from the literature, from established gene-disease validity evidence, or from expert panel specifications. The disease mechanism determines which functional assays are relevant and how results should be interpreted. For example, if a gene causes disease through haploinsufficiency, an assay that measures total protein abundance may be relevant. If a gene causes disease through a dominant-negative mechanism, an assay that measures interference with wild-type function may be necessary. If the disease mechanism is unknown, functional evidence may be more difficult to apply. ### Step 2: Evaluate Applicability The evaluator must determine whether the functional assay is applicable to the specific gene and variant. This includes considering whether the assay measures a function relevant to the disease mechanism, whether the assay uses the appropriate isoform, and whether the assay is appropriate for the variant type. A functional assay that measures enzymatic activity may not be applicable to a variant that affects splicing. An assay that measures protein-protein interaction may not be applicable to a variant in a domain that does not participate in that interaction. The evaluator should also consider whether the assay was performed in a relevant cell type or model system. ### Step 3: Assess Assay Validity The evaluator must assess whether the assay is well-established and validated. This includes considering the number and type of controls, the sensitivity and specificity of the assay in distinguishing pathogenic from benign variants, the quantitative nature of the readout, and the reproducibility of the results. The SVI Working Group's analysis of variant controls provides guidance on the minimum controls required to reach moderate level evidence. Assays with fewer controls or with poor discrimination between known pathogenic and benign variants support weaker evidence. ### Step 4: Determine Evidence Strength Based on the first three steps, the evaluator determines whether the functional evidence supports PS3 or BS3 and at what strength. The ACMG/AMP framework allows for evidence to be applied at strong, moderate, or supporting levels. The SVI Working Group's recommendations provide guidance on how to calibrate evidence strength based on assay validation. An assay that is well-validated with multiple controls, shows clear separation between pathogenic and benign variants, and is directly relevant to the disease mechanism may support strong evidence. An assay with limited validation or with less clear separation between controls may only support moderate or supporting evidence. ## At a Glance: Functional Assay Evaluation Checklist | Evaluation Step | Key Questions | Evidence Strength Implication | |---|---|---| | Disease mechanism | Is loss-of-function, gain-of-function, or dominant-negative the established mechanism? | Determines whether PS3/BS3 can be applied and which assay types are relevant | | Assay applicability | Does the assay measure a function relevant to the variant's predicted effect? Is the correct isoform used? | Assays measuring irrelevant functions cannot support PS3/BS3 | | Control validation | How many known pathogenic and benign variants were tested? What was the sensitivity and specificity? | More validated controls support stronger evidence, insufficient controls limit evidence to supporting level | | Quantitative readout | Is the readout quantitative with pre-specified thresholds? Are replicates consistent? | Quantitative assays with pre-specified thresholds support stronger evidence than qualitative assays | | Evidence calibration | Does the assay distinguish pathogenic from benign variants with high accuracy? | High accuracy supports moderate or strong evidence, overlapping distributions support only supporting evidence | ## Splicing Assays and the PS3/BS3 Distinction The ClinGen SVI Splicing Subgroup developed specific recommendations for applying ACMG/AMP codes relating to splicing data and computational predictions. The subgroup utilized empirically derived splicing evidence to determine the evidence weighting of splicing-related data and appropriate criteria code selection for general use. A key recommendation from the Splicing Subgroup is that PS3/BS3 codes should be applied only for well-established assays that measure functional impact not directly captured by RNA-splicing assays. This means that RNA-based assays demonstrating aberrant splicing should not automatically receive PS3. Instead, the subgroup proposed repurposing the PVS1_Strength code to capture splicing assay data that provide experimental evidence for variants resulting in RNA transcripts with loss of function. Conversely, BP7 may be used to capture RNA results demonstrating no splicing impact for intronic and synonymous variants. For missense variants, BP7 may be used if protein functional impact has been excluded. The subgroup also recommended the application of PS1 based on similarity of predicted RNA-splicing effects for a variant under assessment in comparison with a known pathogenic variant. This distinction has practical implications for variant interpretation. A researcher who performs a minigene splicing assay and observes aberrant splicing should not automatically apply PS3. The evidence should be evaluated according to the Splicing Subgroup's recommendations, which may involve applying PVS1_Strength or another code depending on the specific findings. For laboratory professionals, this means that the type of functional assay determines which ACMG/AMP code is appropriate. Protein functional assays that measure activity, binding, or stability may support PS3 or BS3. RNA-based assays that measure splicing effects should be evaluated according to the Splicing Subgroup's recommendations and may support different codes. ## Examples of Functional Assay Evaluation ### Example 1: Enzymatic Activity Assay Consider a missense variant in a gene encoding an enzyme where loss-of-function is the established disease mechanism. A researcher performs an in vitro enzymatic activity assay using recombinant protein expressed in bacteria. The assay measures the rate of substrate conversion for wild-type, the variant, and several known pathogenic and benign variants. The evaluator should first confirm that the disease mechanism is loss-of-function and that enzymatic activity is the relevant function. The evaluator should then assess whether the assay was validated with known pathogenic and benign variants. If the assay shows clear separation between pathogenic and benign controls and the variant shows activity in the pathogenic range, PS3 may be applied at moderate or strong strength depending on the number of controls and the assay's validation. However, if the assay was performed only with the variant of interest and wild-type, without known pathogenic and benign controls, the evidence is weaker. The evaluator cannot be confident that the assay distinguishes pathogenic from benign variants, and the evidence may only support a lower level of strength or may not be applied at all. ### Example 2: Protein-Protein Interaction Assay Consider a variant in a gene where protein-protein interaction is critical for function. A researcher performs a co-immunoprecipitation assay to measure binding between the variant protein and its interaction partner. The assay includes wild-type, the variant, and several known pathogenic and benign variants. The BRCA1 GFP-reassembly assay example demonstrates how protein interaction assays can be validated. The researchers tested 26 clinically classified variants and achieved 100% sensitivity and specificity in identifying pathogenic and benign variants for both the RING/UbcH5a and the BRCTs/ABRAXAS interactions. Based on this validation, the researchers derived the strength of evidence as moderate for both PS3 and BS3 criteria assessment. This example shows that protein interaction assays can support PS3 or BS3 when properly validated. The key is the validation against known variants and the demonstration that the assay discriminates between pathogenic and benign variants with high accuracy. ### Example 3: Splicing Assay Consider a synonymous variant near a splice site. A researcher performs a minigene splicing assay and observes that the variant causes exon skipping. According to the ClinGen SVI Splicing Subgroup recommendations, this evidence should not automatically receive PS3. The subgroup proposed that splicing assay data providing experimental evidence for variants resulting in RNA transcripts with loss of function should be captured using the PVS1_Strength code. The evaluator should assess whether the splicing assay was properly validated, whether the observed splicing alteration results in loss of function, and whether the assay reflects the variant's effect in the relevant tissue or cell type. The evidence strength should be calibrated based on the assay's validation and the specificity of the observed effect. ## Common Failure Patterns in Functional Assay Evaluation Several common failure patterns lead to inappropriate application of PS3 or BS3 evidence. Recognizing these patterns helps evaluators avoid errors in variant classification. ### Failure to Match Assay to Disease Mechanism A functional assay that measures a function unrelated to the disease mechanism cannot support PS3 or BS3. For example, an assay measuring protein abundance may not be relevant for a gene where the disease mechanism is gain-of-function. The evaluator must confirm that the assay measures a function directly relevant to the disease mechanism. ### Insufficient Control Validation An assay tested only with wild-type and the variant of interest cannot establish that the assay distinguishes pathogenic from benign variants. The assay must be validated with known pathogenic and benign controls to demonstrate that the readout correlates with clinical classification. The SVI Working Group's analysis of variant controls provides guidance on the minimum controls required to reach moderate level evidence. ### Post Hoc Threshold Determination If the threshold for calling a variant damaging is determined after analyzing the data, the evidence is weaker than if the threshold was pre-specified. The evaluator should check whether the assay had a pre-specified threshold based on validation with known controls. ### Overinterpretation of Qualitative Results Qualitative assays that simply indicate whether an effect is present or absent provide weaker evidence than quantitative assays. The evaluator should consider whether the assay produces a quantitative readout that allows comparison between variants. ### Applying PS3 to Splicing Assays According to the ClinGen SVI Splicing Subgroup, PS3/BS3 codes should be applied only for well-established assays that measure functional impact not directly captured by RNA-splicing assays. Applying PS3 to a splicing assay without considering the Splicing Subgroup's recommendations is a common error. ### Ignoring Assay Limitations Every functional assay has limitations. The evaluator should consider whether the assay uses a relevant cell type, whether the protein is expressed at physiologic levels, whether post-translational modifications are present, and whether the assay reflects the in vivo context. Assays performed in non-physiologic systems may provide weaker evidence. ## Records and Measurements for Functional Assay Evaluation Laboratory professionals evaluating functional evidence should maintain records that document the evaluation process. The following records support consistent and reproducible application of PS3/BS3 evidence: ### Assay Validation Records - Number and identity of known pathogenic variants tested - Number and identity of known benign variants tested - Sensitivity and specificity of the assay in distinguishing pathogenic from benign variants - Quantitative readout for each control variant - Threshold for calling a variant damaging, with justification ### Variant Testing Records - Quantitative readout for the variant of interest - Number of replicates and variability between replicates - Comparison of variant readout to wild-type and control distributions - Interpretation of whether the variant shows abnormal or normal function ### Evidence Application Records - ACMG/AMP code applied (PS3 or BS3) - Strength of evidence (strong, moderate, or supporting) - Justification for the evidence strength based on assay validation - Any limitations or caveats that affect the interpretation These records support transparency and reproducibility in variant classification. They also facilitate re-evaluation when new functional data become available or when assay validation is updated. ## Limitations of Functional Assay Evidence Functional assay evidence has inherent limitations that evaluators must consider. Even well-validated assays may not fully capture the in vivo effect of a variant. The following limitations are relevant for PS3/BS3 application: ### Model System Limitations Functional assays are often performed in non-human systems such as bacteria, yeast, or cultured cell lines. These systems may not reflect the human in vivo context, including tissue-specific expression, post-translational modifications, and protein-protein interactions. An assay that shows a damaging effect in a model system may not accurately predict the effect in a patient. ### Isoform Considerations Many genes have multiple transcript isoforms, and a functional assay may measure activity of one isoform while the variant primarily affects another. The evaluator should confirm that the assay uses the isoform most relevant to the disease phenotype. ### Hypomorphic Variants Some variants reduce but do not eliminate function. These hypomorphic variants may show intermediate readouts in functional assays, making classification difficult. The Alport Variant Collaborative noted that recognition and interpretation of hypomorphic variants in the COL4A3-COL4A5 genes remains a challenge. ### Tissue-Specific Effects A variant may have different functional effects in different tissues. An assay performed in one cell type may not capture the effect in the tissue most relevant to the disease. The evaluator should consider whether the assay system reflects the disease-relevant tissue. ### Quantitative Threshold Uncertainty The threshold for calling a variant damaging is often arbitrary and may not reflect the true functional threshold in vivo. A variant that shows 80% of wild-type activity may be pathogenic in one gene but benign in another. The evaluator should consider whether the assay's threshold is calibrated to the disease mechanism. ## Professional Escalation Criteria Laboratory professionals should escalate functional assay evaluation to appropriate expertise when certain conditions are met. The following criteria indicate that additional review is needed: ### Gene-Specific Expertise Required When the gene has complex disease mechanisms, multiple isoforms, or tissue-specific effects, evaluation by a gene-specific expert panel or variant curation expert panel may be necessary. The ClinGen Variant Curation Expert Panels develop gene-specific rule specifications that may include guidance on functional assay evaluation. ### Conflicting Functional Evidence When different functional assays produce conflicting results for the same variant, the evaluator should escalate to a more comprehensive review. Conflicting evidence may indicate that one assay is not relevant to the disease mechanism or that the variant has complex effects. ### Novel Assay Types When a functional assay uses a novel methodology that has not been previously validated for the gene, the evaluator should seek expert review before applying PS3 or BS3 evidence. The assay's validity should be established before it is used for clinical variant interpretation. ### Borderline Quantitative Results When a variant's quantitative readout falls near the threshold between normal and abnormal function, the evaluator should consider whether the evidence is strong enough to support PS3 or BS3. Borderline results may only support a lower level of evidence strength. ### Discrepancy with Other Evidence When functional evidence conflicts with other evidence in the variant classification, the evaluator should review the functional assay's validity and relevance. A well-validated functional assay may outweigh weaker evidence, but conflicting evidence should be documented and considered in the overall classification. ## Safety and Regulatory Context Functional assay evidence for variant interpretation is used in clinical laboratory settings where results may inform patient management decisions. The application of PS3 or BS3 evidence should follow established guidelines and be documented in the variant classification report. Laboratories performing functional assays for clinical variant interpretation should ensure that assays are validated according to appropriate standards. The ClinGen SVI Working Group's recommendations provide guidance on assay validation and evidence strength calibration, but laboratories should also follow applicable regulatory requirements for laboratory-developed tests. The Alport Variant Collaborative's consensus statement illustrates the importance of gene-specific considerations in applying ACMG criteria. The collaborative refined the ACMG criteria for variant assessment for the Alport genes, identified mutational hotspots, and considered that well-established functional assays were still mainly research tools. This gene-specific approach reflects the broader principle that functional assay evaluation must be tailored to the gene and disease. For researchers developing functional assays, the goal should be to produce evidence that meets the standards for clinical variant interpretation. This means validating assays with known pathogenic and benign variants, using quantitative readouts with pre-specified thresholds, and documenting the assay's sensitivity and specificity. ## Practical Steps for Implementing Functional Assay Evaluation Laboratory professionals can implement the ClinGen SVI Working Group's framework through the following practical steps: ### Step 1: Establish Gene-Specific Context Before evaluating any functional assay, establish the gene's disease mechanism, the relevant isoform, and the types of variants that are typically pathogenic. Review gene-specific rule specifications from ClinGen Variant Curation Expert Panels if available. ### Step 2: Collect Functional Assay Information Gather all available functional assay data for the variant, including the assay type, the readout, the controls tested, and the validation status. For published studies, review the methods and results carefully to assess the assay's quality. ### Step 3: Apply the Four-Step Framework Apply the ClinGen SVI Working Group's four-step framework: define the disease mechanism, evaluate applicability, assess assay validity, and determine evidence strength. Document the assessment for each step. ### Step 4: Determine the Appropriate ACMG/AMP Code Based on the framework assessment, determine whether the functional evidence supports PS3, BS3, or neither. Consider whether the assay measures functional impact directly or whether it captures splicing effects that should be evaluated according to the Splicing Subgroup's recommendations. ### Step 5: Calibrate Evidence Strength Calibrate the evidence strength based on the assay's validation. Assays with more validated controls and clear separation between pathogenic and benign variants support stronger evidence. Document the justification for the evidence strength. ### Step 6: Document and Review Document the functional assay evaluation in the variant classification record. Include the assay validation data, the variant's quantitative readout, the evidence strength applied, and any limitations or caveats. Review the evaluation with appropriate expertise if needed. ## A Practical Decision Framework for Assigning PS3 and BS3 Evidence Strength The ClinGen SVI Working Group framework provides the conceptual structure for evaluating functional assays, but applying it in daily practice requires a concrete decision procedure. The following framework translates the four steps into sequential decisions that produce a defensible evidence assignment. This approach is designed for laboratory professionals who need to document why a particular functional study received strong, moderate, or supporting weight. ### Step 1: Confirm the Assay Measures the Disease-Relevant Function Before any other consideration, verify that the assay readout corresponds to the function that is disrupted in disease. For a loss-of-function gene where haploinsufficiency causes the phenotype, an assay measuring residual enzymatic activity is directly relevant. For a gene where the pathogenic mechanism is dominant-negative interference, an assay measuring total protein abundance may be misleading because it does not capture the interference effect. The ClinGen SVI Working Group recommendations require evaluators to define the disease mechanism as the first step. This determination should come from established gene-disease validity evidence, not from the functional assay itself. If the disease mechanism is unknown or contested, the functional evidence cannot be applied at strong strength because the evaluator cannot confirm the assay measures the relevant function. Document the disease mechanism determination in the variant record with a citation to the evidence source. If the mechanism is established through ClinGen gene-disease validity classifications or expert panel specifications, note that source. If the mechanism is inferred from limited literature, reduce the evidence strength accordingly. ### Step 2: Determine Whether the Assay Type Matches the Variant Class The assay must be appropriate for the variant type under assessment. A missense variant in a coding region may be evaluated with a protein functional assay. A synonymous or intronic variant near a splice site requires an RNA-based assay. The ClinGen SVI Splicing Subgroup explicitly proposed that PS3 and BS3 codes apply only for well-established assays that measure functional impact not directly captured by RNA-splicing assays. For splicing evidence, the subgroup recommended using PVS1_Strength for experimental evidence of loss-of-function transcripts and BP7 for RNA results demonstrating no splicing impact in intronic and synonymous variants. This means a laboratory should not apply PS3 to a minigene assay result showing exon skipping. The correct code depends on whether the resulting transcript produces loss of function. Create a simple matching table for the variant class and assay type: | Variant Class | Appropriate Assay Type | Applicable Code | |---|---|---| | Missense | Protein activity, binding, stability | PS3 or BS3 | | Nonsense or frameshift | Protein abundance or transcript degradation | PVS1 or PS3 with caution | | Synonymous near splice site | RNA splicing assay | PVS1_Strength or BP7 | | Intronic | RNA splicing assay | PVS1_Strength or BP7 | | Regulatory region | Reporter gene or expression assay | PS3 or BS3 with validation | ### Step 3: Score the Assay Validation Against Known Variants The SVI Working Group estimated the odds of pathogenicity for assays using various numbers of variant controls to determine the minimum controls required to reach moderate level evidence. This analysis provides the basis for a scoring system that laboratories can apply consistently. Assign points based on the validation panel: - One known pathogenic variant tested and showing abnormal function: 1 point - Two to four known pathogenic variants tested and showing abnormal function: 2 points - Five or more known pathogenic variants tested and showing abnormal function: 3 points - One known benign variant tested and showing normal function: 1 point - Two to four known benign variants tested and showing normal function: 2 points - Five or more known benign variants tested and showing normal function: 3 points - Wild-type control included in each experiment: 1 point - Empty vector or negative control included: 1 point - Assay demonstrates complete separation between pathogenic and benign control distributions: 2 points - Assay shows partial overlap between control distributions: 1 point A total score of 8 or higher supports strong evidence. A score of 5 to 7 supports moderate evidence. A score of 3 to 4 supports supporting evidence. A score below 3 means the assay does not meet the threshold for PS3 or BS3 application. This scoring system is a practical operationalization of the SVI Working Group guidance. The specific point thresholds should be calibrated by each laboratory based on its validation requirements and documented in the laboratory's standard operating procedures. ### Step 4: Evaluate the Quantitative Readout and Threshold The assay must produce a quantitative readout that allows comparison between the variant and controls. The threshold for calling a variant damaging should be pre-specified based on the distribution of known pathogenic and benign controls. For each assay, document: - The quantitative metric measured (for example, percent of wild-type activity, binding affinity, or transcript abundance) - The number of replicates performed for each sample - The standard deviation or confidence interval for each measurement - The pre-specified threshold for calling a variant damaging - Whether the variant readout falls clearly above or below the threshold or within the borderline zone A variant readout that falls within the borderline zone between the pathogenic and benign control distributions should not receive strong evidence. The uncertainty in the measurement means the evaluator cannot confidently assign the variant to either category. ### Step 5: Apply the Evidence Strength and Document the Rationale Based on the validation score and the quantitative assessment, assign the evidence strength and document the rationale. The documentation should include the validation score calculation, the control variants used, the variant readout, and the threshold determination. The BRCA1 GFP-reassembly assay example illustrates this process. The researchers tested 26 clinically classified variants and achieved 100% sensitivity and specificity in identifying pathogenic and benign variants for both the RING/UbcH5a and the BRCTs/ABRAXAS interactions. Based on this validation, they derived the evidence strength as moderate for both PS3 and BS3 criteria assessment. The high sensitivity and specificity supported moderate evidence, but the researchers did not claim strong evidence, likely because the assay measures protein interaction instead of the full range of BRCA1 functions. ### Troubleshooting Common Decision Points When the framework produces an ambiguous result, the following troubleshooting steps can help resolve the uncertainty. **Problem: The assay shows a damaging effect but the disease mechanism is uncertain.** Resolution: Reduce the evidence strength to supporting or do not apply PS3. Functional evidence cannot compensate for an unknown disease mechanism because the evaluator cannot confirm the assay measures the relevant function. Document the mechanism uncertainty in the variant record. **Problem: The assay was validated with pathogenic controls but no benign controls.** Resolution: The assay can demonstrate that a variant behaves like known pathogenic variants, but it cannot demonstrate that a variant behaves like benign variants. This asymmetry means the assay may support PS3 but cannot support BS3. The evidence strength for PS3 should be limited to moderate because the lack of benign controls means the assay's specificity is unknown. **Problem: The variant readout falls between the pathogenic and benign control distributions.** Resolution: Do not apply PS3 or BS3 at strong strength. The overlapping distributions mean the assay cannot reliably classify the variant. Consider whether the variant may be hypomorphic, showing partial reduction of function. The Alport Variant Collaborative noted that recognition and interpretation of hypomorphic variants in the COL4A3-COL4A5 genes remains a challenge. Document the borderline result and consider whether other evidence types can resolve the classification. **Problem: Different functional assays produce conflicting results for the same variant.** Resolution: Evaluate each assay independently using the framework. Determine whether one assay is more relevant to the disease mechanism or has better validation. If the assays measure different functions, the variant may affect one function but not another. Document the conflict and consider whether the variant has complex effects that require gene-specific expertise. **Problem: The assay was performed in a non-physiologic system such as bacterial overexpression.** Resolution: The model system limitation reduces the evidence strength. An assay performed in a non-human system may not reflect the human in vivo context, including tissue-specific expression, post-translational modifications, and protein-protein interactions. Reduce the evidence strength by one level unless the assay has been extensively validated against known variants and the function measured is conserved across species. ### Records for the Decision Framework Maintain a standardized record for each functional assay evaluation that includes: - The assay name and methodology - The disease mechanism determination with citation - The variant class and matching assay type - The validation score with the point breakdown - The control variants used and their readouts - The variant readout with replicates and variability - The pre-specified threshold and whether the variant falls clearly above or below - The final evidence strength assignment with justification - Any limitations or caveats that affect the interpretation These records support transparency and reproducibility in variant classification. They also facilitate re-evaluation when new functional data become available or when assay validation is updated. The records should be reviewed by a second evaluator or a variant curation committee to ensure consistency in evidence application. ### Escalation Criteria for the Decision Framework Escalate the functional assay evaluation to gene-specific expertise or a variant curation expert panel when: - The validation score falls between 4 and 6, creating uncertainty about whether the evidence supports moderate or supporting strength - The variant readout falls within the borderline zone between control distributions - The disease mechanism is uncertain or contested in the literature - The assay uses a novel methodology that has not been previously validated for the gene - Different functional assays produce conflicting results that cannot be resolved by the framework - The gene has multiple isoforms and the assay may not measure the disease-relevant isoform The ClinGen Variant Curation Expert Panels develop gene-specific rule specifications that may include guidance on functional assay evaluation. Consulting these specifications when available can resolve ambiguity in the framework application. ## Frequently Asked Questions ### What qualifies as a well-established functional assay for PS3 or BS3 application? A well-established functional assay is one that has been validated with known pathogenic and benign variants, produces a quantitative readout with pre-specified thresholds, and measures a function directly relevant to the disease mechanism. The ClinGen SVI Working Group recommends that assays demonstrate high sensitivity and specificity in distinguishing pathogenic from benign variants. An assay tested only with wild-type and the variant of interest, without known controls, does not qualify as well-established. ### How many control variants are needed to support moderate level evidence? The ClinGen SVI Working Group estimated the odds of pathogenicity for assays using various numbers of variant controls to determine the minimum controls required to reach moderate level evidence. The specific number depends on the assay's sensitivity and specificity. Assays with more validated controls and clear separation between pathogenic and benign variants support stronger evidence. Evaluators should consult the SVI Working Group's recommendations for specific guidance on control numbers. ### Can RNA-based splicing assays be used to apply PS3? According to the ClinGen SVI Splicing Subgroup, PS3/BS3 codes should be applied only for well-established assays that measure functional impact not directly captured by RNA-splicing assays. The subgroup proposed repurposing the PVS1_Strength code to capture splicing assay data that provide experimental evidence for variants resulting in RNA transcripts with loss of function. BP7 may be used to capture RNA results demonstrating no splicing impact for intronic and synonymous variants. ### How should conflicting functional assay results be handled? Conflicting functional assay results should be evaluated carefully to determine whether one assay is more relevant to the disease mechanism or better validated than another. The evaluator should consider the assay type, the validation status, and the relevance to the disease mechanism. Conflicting evidence may indicate that the variant has complex effects or that one assay is not appropriate for the gene. Escalation to gene-specific expertise may be necessary. ### What is the difference between PS3 and BS3 evidence strength? PS3 and BS3 are both strong evidence codes in the ACMG/AMP framework, but they apply to different findings. PS3 applies when well-established functional studies show a damaging effect on gene or protein function, supporting pathogenicity. BS3 applies when well-established functional studies show no damaging effect, supporting benignity. The ClinGen SVI Working Group's recommendations provide guidance on calibrating the strength of evidence based on assay validation, allowing for moderate or supporting level evidence when the assay is not fully validated. ### How should functional assay evidence be documented in variant classification records? Functional assay evidence should be documented with the assay type, the validation data including the number and identity of controls, the variant's quantitative readout, the threshold for calling a variant damaging, and the evidence strength applied. The documentation should include any limitations or caveats that affect the interpretation. This documentation supports transparency and reproducibility in variant classification. ### Are functional assays always required for PS3 or BS3 application? Functional assays are not always required for PS3 or BS3 application, but the evidence must come from well-established functional studies. The original ACMG/AMP guidelines stated that PS3 and BS3 apply to well-established functional assays, and the ClinGen SVI Working Group refined this by providing guidance on assay validation and evidence strength calibration. If no well-established functional assay is available for the gene, PS3 or BS3 may not be applicable. ### How do gene-specific variant curation expert panels affect functional assay evaluation? ClinGen Variant Curation Expert Panels develop gene-specific rule specifications that may include guidance on functional assay evaluation. These specifications may define which assays are considered well-established for the gene, how evidence strength should be calibrated, and how functional evidence should be integrated with other evidence. Evaluators should consult gene-specific rule specifications when available. ## Related Bioinformatics Guides - [Proteomics Analysis Tools: A Comparative Guide for Functional Interpretation](/knowledge/bioinformatics/proteomics-analysis-tools-a-comparative-guide-for-functional-interpretation) - [Spatial Transcriptomics Methods: A Guide to Experimental Approaches](/knowledge/bioinformatics/spatial-transcriptomics-methods-a-guide-to-experimental-approaches) - [Functional Annotation of Metagenomes: A Guide to Databases and Pipelines](/knowledge/bioinformatics/functional-annotation-of-metagenomes-a-guide-to-databases-and-pipelines) - [Lipidomic Analysis: A Beginner's Guide to Workflows and Data Interpretation](/knowledge/bioinformatics/lipidomic-analysis-a-beginner-s-guide-to-workflows-and-data-interpretation) - [Pathway Enrichment Analysis for Proteomics: Tools and Interpretation](/knowledge/bioinformatics/pathway-enrichment-analysis-for-proteomics-tools-and-interpretation) ## References and Further Reading - [NCBI Data Resources](https://www.ncbi.nlm.nih.gov/). National Center for Biotechnology Information. - [EMBL-EBI Training](https://www.ebi.ac.uk/training). European Bioinformatics Institute. - [Bioconductor](https://bioconductor.org/). Bioconductor Project. - [Galaxy Training Network](https://training.galaxyproject.org/). Galaxy Project. - [nf-core Documentation](https://nf-co.re/docs). nf-core. - [The Carpentries Lessons](https://carpentries.org/lessons). The Carpentries. - [Recommendations for application of the functional evidence PS3/BS3 criterion using the ACMG/AMP sequence variant interpretation framework.](https://pubmed.ncbi.nlm.nih.gov/31892348). Genome medicine, 2019. - [Using the ACMG/AMP framework to capture evidence related to predicted and observed impact on splicing: Recommendations from the ClinGen SVI Splicing Subgroup.](https://pubmed.ncbi.nlm.nih.gov/37352859). American journal of human genetics, 2023. - [APPLICATION OF THE ACMG/AMP FRAMEWORK TO CAPTURE EVIDENCE RELEVANT TO PREDICTED AND OBSERVED IMPACT ON SPLICING: RECOMMENDATIONS FROM THE CLINGEN SVI SPLICING SUBGROUP.](https://pubmed.ncbi.nlm.nih.gov/36865205). medRxiv : the preprint server for health sciences, 2023. - [Consensus statement on standards and guidelines for the molecular diagnostics of Alport syndrome: refining the ACMG criteria.](https://pubmed.ncbi.nlm.nih.gov/33854215). European journal of human genetics : EJHG, 2021. - [Refinement of the assignment to the ACMG/AMP BS3 and PS3 criteria of eight BRCA1 variants of uncertain significance by integrating available functional data with protein interaction assays.](https://pubmed.ncbi.nlm.nih.gov/37168384). Frontiers in oncology, 2023. > This article is educational and does not replace validated analysis plans, institutional policy, clinical interpretation, or specialist review.