Applying data science methodologies with artificial intelligence variant reinterpretation to map and estimate genetic...

Area of Science:

• Genomics
• Bioinformatics
• Health Informatics

Background:

• Clinical genetic data is often unstructured and underutilized.
• Data science offers methods to analyze this complex data.
• Artificial intelligence (AI) can aid in genetic variant interpretation.

Purpose of the Study:

• To apply AI and data science to reinterpret clinical genetic variants.
• To associate variants with patient demographics and geographic data.
• To identify actionable genetic findings and estimate disorder prevalence.

Main Methods:

• Utilized Franklin© AI for reinterpretation of genetic variants over a 5-year period.
• Integrated variant data with electronic healthcare records using PowerBI©.
• Mapped variants by ZIP codes to identify geographic hotspots.

Main Results:

• 3,065 variants were identified, with 98% matched to patients.
• AI reinterpretation altered the interpretation of 24% of variants, yielding 156 actionable reinterpretations.
• Identified 739 Mendelian genetic disorders, including specific conditions like Zellweger Spectrum Disorder, Bardet-Biedl syndrome, and Rett syndrome.

Conclusions:

• Developed a database and Exploratory Data Analysis (EDA) methodology for systematic variant reinterpretation.
• Enabled estimation of variant prevalence and identification of conditions treatable with new therapies.
• Localized geographic areas enriched for pathogenic genetic variants, informing public health strategies.