APOGEE 2: multi-layer machine-learning model for the interpretable prediction of mitochondrial missense variants

Area of Science:

• Genetics
• Molecular Biology
• Bioinformatics

Background:

• Mitochondrial dysfunction, often caused by mitochondrial DNA mutations, presents diverse clinical symptoms, complicating variant pathogenicity interpretation.
• Accurate assessment of mitochondrial genetic variants is crucial for diagnosing and understanding mitochondrial diseases.

Purpose of the Study:

• To introduce APOGEE 2, an advanced ensemble method for improved pathogenicity prediction of mitochondrial missense variants.
• To enhance the accuracy and reliability of interpreting variants within the mitochondrial genome.

Main Methods:

• APOGEE 2 utilizes a mitochondrially-centered ensemble approach with an improved machine learning model and a curated training dataset.
• Incorporates region-wise genome fragility assessments and mechanistic analyses of amino acid impacts on protein structure.
• Developed following joint consensus recommendations from the American College of Medical Genetics and Genomics/Association for Molecular Pathology.

Main Results:

• APOGEE 2 demonstrates enhanced performance metrics for pathogenicity prediction of mitochondrial missense variants.
• Provides region-specific genome fragility insights and detailed mechanistic explanations for variant effects.
• Pathogenicity scores and probabilities are precompiled and accessible via the MitImpact database.

Conclusions:

• APOGEE 2 is a valuable tool for addressing the complexities of interpreting mitochondrial missense variants.
• Facilitates more accurate genetic diagnoses and research into mitochondrial disorders.
• Improves the understanding of genotype-phenotype correlations in mitochondrial diseases.