Cohort-driven variant burden analysis and pathogenicity identification in monogenic autoinflammatory disorders

Area of Science:

• Genetics
• Genomics
• Computational Biology

Background:

• Monogenic autoinflammatory diseases (AIDs) involve nearly 50 identified pathogenic genes and hundreds of variants.
• Understanding the pathogenicity mechanisms of many genes and variants in AIDs remains a challenge.

Purpose of the Study:

• To investigate pathogenic variants in monogenic AIDs using variant burden analysis.
• To identify consistent characteristics for improved pathogenicity classification.

Main Methods:

• Performed variant burden analysis on the Genome Aggregation Database cohort.
• Analyzed allelic signatures and deleterious predictions of variants.
• Utilized a Random Forest model with cohort-based features for pathogenicity prediction.

Main Results:

• Variant pathogenicity classification requires more than allele frequency and deleterious predictions alone.
• Distinct allelic signature patterns and deleterious variant distributions were observed across pathogenicity categories.
• The developed cohort-based features achieved superior prediction performance for monogenic AIDs pathogenicity.

Conclusions:

• Variant frequency and deleterious predictions within a cohort effectively classify pathogenicity.
• This approach enhances the accuracy of current variant pathogenicity classification.
• Cohort-based features show potential for broader applications in disease classification.