Degrees of separation as a statistical tool for evaluating candidate genes

Area of Science:

• Genetics
• Bioinformatics
• Systems Biology

Background:

• Candidate gene selection is vital for dissecting complex genetic architectures.
• Gene networks offer valuable insights for identifying candidate genes.
• Current validation methods using gene network connectedness in Genome Wide Association (GWA) and Quantitative Trait Locus (QTL) studies can yield misleading results if not properly validated.

Purpose of the Study:

• To present a robust method and tool for validating gene pairs identified through GWA studies within their co-occurring gene network context.
• To ensure that proposed gene interactions and associations are not statistical artifacts arising from gene network architecture.
• To provide a reliable approach for leveraging the increasing volume of gene network information for candidate gene validation.

Main Methods:

• Developed the CandidateBacon R package for calculating the average degree of separation (DoS) between gene pairs.
• Utilized empirical estimates of average gene connectedness from existing gene networks.
• Compared the connectedness of candidate gene pairs against the average connectedness within the gene network.

Main Results:

• CandidateBacon offers an efficient method to calculate the average DoS between gene pairs across various gene networks.
• Demonstrated the utility of empirical DoS estimates for validating candidate gene pairs from GWA studies.
• Showcased how comparing gene pair connectedness to network averages supports proposed interactions.

Conclusions:

• The CandidateBacon package provides a reliable method for validating candidate gene pairs from GWA studies.
• This approach mitigates the risk of false positives by accounting for gene network topology.
• Validating gene interactions through network connectedness enhances the interpretability of genetic association studies.