VarMeter2: An enhanced structure-based method for predicting pathogenic missense variants through Mahalanobis...

Area of Science:

• Genomics
• Computational Biology
• Rare Disease Diagnostics

Background:

• Predicting missense variant pathogenicity is vital for diagnosing rare genetic diseases.
• Existing tools require validation and improvement for broader clinical application.

Purpose of the Study:

• To enhance the predictive accuracy of variant pathogenicity assessment.
• To develop an improved computational tool, VarMeter2, for classifying missense variants.

Main Methods:

• Analysis of structural features (nSASA, mutation energy, pLDDT) from AlphaFold models for 296 pathogenic and 240 benign variants.
• Development of VarMeter2 using Mahalanobis distance for variant classification.
• Experimental validation of a novel pathogenic variant (Q365P) in N-sulphoglucosamine sulphohydrolase (SGSH).

Main Results:

• VarMeter2 achieved 82% accuracy on the ClinVar dataset, outperforming the original VarMeter (74%).
• VarMeter2 showed 84% accuracy on SGSH variants, identifying a novel pathogenic variant Q365P.
• The Q365P variant exhibited loss of enzymatic activity, mislocalization, and reduced protein stability, consistent with Sanfilippo syndrome A.

Conclusions:

• VarMeter2 demonstrates improved predictive power and versatility for missense variant pathogenicity assessment.
• The tool aids in the diagnosis of rare diseases by accurately classifying genetic variants.
• Findings support the clinical utility of VarMeter2 in identifying disease-causing mutations.