Personalised analytics for rare disease diagnostics.

Area of Science:

• Genomics
• Bioinformatics
• Medical Genetics

Background:

• Whole genome and exome sequencing are crucial for diagnosing rare genetic disorders.
• Interpreting the large number of variants identified presents a significant diagnostic challenge.
• Prioritizing variants based on disease phenotype is essential for accurate diagnosis.

Purpose of the Study:

• To investigate if linking gene expression patterns across multiple tissues to disease phenotypes can improve the prioritization of disease-causing genetic variants.
• To develop and test computational classifiers for identifying associations between tissue-specific gene expression and disease phenotypes.

Main Methods:

• Utilized the Genotype-Tissue Expression (GTEx) project dataset for gene expression information.
• Employed disease-agnostic variant prioritization tools such as CADD and MetaSVM.
• Constructed machine learning classifiers to learn associations between gene expression and phenotypes.

Main Results:

• Integrating GTEx expression data with variant prioritization methods consistently improved classification accuracy.
• The developed method demonstrated enhanced ability to link genetic variants to specific disease phenotypes.
• This approach offers a novel way to leverage functional genomic data for clinical applications.

Conclusions:

• Linking tissue-specific gene expression data to disease phenotypes is a valuable strategy for prioritizing genetic variants in clinical diagnostics.
• The findings highlight an underutilized approach for improving the diagnostic yield of genetic sequencing.
• This methodology can be extended to incorporate other functional genomic datasets for enhanced variant interpretation.