DeepSVP

DeepSVP prioritizes structural variants (SVs) implicated in genetic diseases by integrating genomic data, gene function annotations including haploinsufficiency and triplosensitivity, phenotype associations, ontologies, gene expression profiles across cell types and anatomical sites, and machine learning to predict phenotypic consequences.

Key Features:

• Integration of Genomic Information: DeepSVP integrates genomic data with gene function annotations, including haploinsufficiency and triplosensitivity, to assess the impact of SVs on genes and genomic regions.
• Phenotype-Based Prioritization: It links structural variants to phenotypic outcomes using phenotype associations to prioritize candidate disease-causing SVs.
• Utilization of Ontologies and Machine Learning: DeepSVP employs ontologies and machine learning to relate genomic data, gene product functions, gene expression profiles across individual cell types and anatomical sites, and phenotype information for prediction.
• Enhanced Success Rate in Variant Identification: The method improves identification of causative variants in benchmark datasets and is effective at discovering novel pathogenic SVs in consanguineous families.

Scientific Applications:

• Genetic Disease Research: Prioritizing structural variants that may contribute to disease phenotypes to support research into genetic etiology.
• Clinical Genomics: Improving the accuracy of SV interpretation to support clinical variant prioritization and diagnostic investigation.

Methodology:

Integrates genomic data with comprehensive gene function information; incorporates phenotypic data linked to genes and maps relationships using ontologies; applies machine learning algorithms to predict phenotypic consequences of structural variants for prioritization.