Abstract: FR-PO0057
Functional Classification of PKD2 Missense Variants Using Saccharomyces cerevisiae
Session Information
- ADPKD and Cystic Kidney Disease - 2
October 23, 2026 | Location: Exhibit Hall A, Convention Center
Abstract Time: 10:00 AM - 12:00 PM
Category: Genetic Diseases of the Kidneys
- 1201 Genetic Diseases of the Kidneys: Cystic (Monogenic)
Authors
- Guerriero, Christopher J., University of Pittsburgh, Pittsburgh, Pennsylvania, United States
- van Loggerenberg, Warren, University of Pittsburgh, Pittsburgh, Pennsylvania, United States
- Ononiwu, Rita N., University of Pittsburgh, Pittsburgh, Pennsylvania, United States
- Sharp, Katherine G., University of Pittsburgh, Pittsburgh, Pennsylvania, United States
- Sarangi, Srikant, Paradigm4, Waltham, Massachusetts, United States
- Roth, Frederick P., University of Pittsburgh, Pittsburgh, Pennsylvania, United States
- Brodsky, Jeffrey L., University of Pittsburgh, Pittsburgh, Pennsylvania, United States
Background
Single amino acid variants in PKD2, which encodes polycystin-2 (PC2), can result in autosomal dominant polycystic kidney disease (ADPKD), yet ~93% are currently classified as variants of uncertain significance (VUS). Therefore, functional data are needed to reclassify VUS and improve clinical variant interpretation, but in most cases, this requires expensive and time-consuming studies.
Methods
To overcome these hurdles, we developed a yeast-based complementation assay that measures the functional consequences of PC2 variants. Specifically, yeast lacking two plasma membrane potassium transporters (trk1Δ/trk2Δ) were engineered to express PC2 or a variant, and cell growth was measured under selective (0 mM KCl) or permissive (100 mM KCl) conditions. To sensitize the assay, PC2 also contained two lower gate mutations, L677A and N681A (“PC2_2A”), which augment PC2 function.
Results
To calibrate the assay, 22 ClinVar-annotated variants were chosen as a “truth-set” to validate the assay. Activity scores were calculated by normalizing growth to PC2_2A. Using optimized thresholds (loss-of-function ≤0.10; partial function >0.10<0.625 and wild-type-like ≥0.625), variants were classified. Notably, results from the yeast assay matched ClinVar annotations better than AlphaMissense, with balanced precision-recall scores of 0.96 vs 0.84, respectively. We then assayed 136 PC2 missense variants curated from various genomic databases, including ClinVar, gnomAD, and the UK Biobank, and identified 68 variants that compromised yeast growth and had a normalized activity score <0.625. Because disease-associated mutations in some protein conformational diseases can be rescued by cell growth at low temperatures, yeast growth was also measured at 26°C. Under these conditions, the growth of 45 deleterious variants was rescued, suggesting that these alleles might be amenable to pharmacological correction.
Conclusion
Finally, we used our data to benchmark four variant effect predictors (VEPs): Rhapsody, Rhapsody 2, AlphaMissense, and ESM2. Rhapsody 2 yielded the highest overall accuracy of 80%, while the other VEPs had an accuracy of ~70%. Together, our work represents a first step toward predicting the pathogenicity of all PC2 VUS and support utilization of the yeast assay as an evidence source for PKD2 VUS classification.
Funding
- NIDDK Support