A comparison of cosegregation analysis methods for the clinical setting

Area of Science:

• Genetics
• Bioinformatics
• Statistical Genetics

Background:

• Genetic variant pathogenicity assessment is crucial for clinical diagnosis.
• Genetics professionals often lack statistical training, leading to reliance on qualitative methods.
• Quantitative cosegregation analysis offers a more robust approach to evaluating genetic variants.

Purpose of the Study:

• To evaluate the clinical utility of quantitative cosegregation analysis.
• To compare the performance of three cosegregation analysis methods: full likelihood Bayes factors (FLB), cosegregation likelihood ratios (CSLR), and counting meioses.
• To encourage the adoption of robust quantitative methods in variant classification.

Main Methods:

• Simulated 1,000 pedigrees for benign and pathogenic variants in BRCA1 and MLH1 using US demographic data.
• Analyzed simulated pedigrees using FLB, CSLR, and counting meioses.
• Developed a website (analyze.myvariant.org) and an R package (CoSeg) to implement these methods.

Main Results:

• Both FLB and CSLR outperformed counting meioses for pathogenic variants.
• FLB and CSLR significantly outperformed counting meioses for benign variants, which counting meioses struggled to assess.
• FLB and CSLR demonstrated similar performance, suggesting their quantitative results can be combined.

Conclusions:

• Quantitative cosegregation analysis, particularly FLB and CSLR, provides valuable evidence for genetic variant pathogenicity assessment.
• Accessible tools (website and R package) are now available to facilitate the use of these robust methods.
• Future variant classification guidelines should incorporate nuanced cosegregation evidence.