PINES: phenotype-informed tissue weighting improves prediction of pathogenic noncoding variants.

Area of Science:

• Genomics
• Computational Biology
• Epigenetics

Background:

• Understanding the noncoding genome is crucial for deciphering gene regulation and disease mechanisms.
• Noncoding variants contribute significantly to human diseases, but their functional impact remains challenging to predict.
• Existing methods often lack the specificity to link noncoding elements to particular phenotypes.

Purpose of the Study:

• To introduce PINES (Phenotype-Informed Noncoding Element Scoring), a novel computational framework.
• To predict the functional impact of noncoding genetic variants by integrating phenotype-dependent epigenetic annotations.
• To provide a flexible and user-friendly tool for researchers studying noncoding variation.

Main Methods:

• Developed the PINES computational framework for variant impact prediction.
• Integrated diverse epigenetic annotations, weighted by phenotype relevance.
• Enabled customization of analyses based on cell type-specific genomic data.
• Validated PINES performance against existing computational methods.

Main Results:

• PINES accurately predicts the functional impact of noncoding variants.
• Phenotype-informed scoring enhances prediction accuracy compared to non-specific methods.
• The framework demonstrates flexibility in incorporating various genomic annotations.
• PINES is accessible via a dedicated web portal for ease of use.

Conclusions:

• PINES offers a powerful and accurate approach for functional characterization of the noncoding genome.
• The phenotype-dependent strategy improves the identification of disease-relevant noncoding variants.
• PINES facilitates a deeper understanding of gene regulation and its links to human diseases.