Objective:

The goal of this study was to accurately predict the pathogenicity of all possible protein-coding genetic variation in genes causing sarcoglycanopathies. 

Background:

Limb-girdle muscular dystrophies R3/R4/R5 are due to recessive mutations in SGCA, SGCB and SGCG.  Genetic variation identified within genetic test reports leads to diagnostic uncertainty.  Only 14% of missense variants in the sarcoglycan genes have been asserted as either benign or pathogenic; with the vast majority being "variants of unknown significance (VUS)."  VUS in LGMD genes leads to an under recognition of these diseases that have treatments within the pre-clinical pipeline. 

Design/Methods:

To help resolve VUSs, we performed Deep Mutational Scanning (DMS) to create lentiviral libraries of all possible coding variations in SGCA, SGCB and SGCG.  Using these libraries we performed high throughput and orthogonal functional screens in cells followed by sequencing to determine a functional score for each unique variant.  

Results:

We tested the effect of all possible coding variants in SGCB to obtain a functional score that predicts the pathogenicity of each genetic variant. The measured scores were perfectly predictive of known pathogenic and benign variants and out-performed all tested bioinformatic prediction algorithms. Validation studies of questionably pathogenic vs benign variants identified in patients support the robustness of our results.  Additionally, patient acquired clinical data demonstrated that our DMS functional score was quantitatively correlated with with disease onset and severity. 

Conclusions:

In summary, we have generated a valuable resource or "look up table" of functionally defined pathogenic variants in sarcoglycan genes.  This type of approach can be utilized for other LGMD genes.  In the future these mutational libraries can be used to perform high throughput analysis in the presence and absence of small molecules that may elucidate variant specific therapies.