Is It Rare or Common? A Coalescent Tree Approach to Identify the Genetic Types of Variants Underlying Complex Diseases

• Activity Number: 30
• Type: Contributed
• Date/Time: Sunday, August 1, 2010 : 2:00 PM to 3:50 PM
• Sponsor: Biometrics Section
• Author(s): Kaustubh Adhikari*+ and Christoph Lange
• Companies: Harvard University and Harvard School of Public Health
• Keywords: Complex Diseases ; Rare Variants ; Common Variants ; Ancestral Recombination Graphs ; Coalescent Trees ; Bayesian Modeling

Abstract

An important problem in modern genetic studies is whether a given disease is caused by a few common variants or by several rare variants in a gene. This paper presents a unique method of identifying the type of variants responsible for a disease in a gene segment through SNP genotyping. Based on case-control data of SNP genotypes in a gene segment, we obtain joint posterior distribution of the number of common and rare variants present in the segment (phenocopy allowed). We do a Bayesian modeling using coalescent genealogical trees to model the unknown ancestral history of both the cases and the controls jointly. This method, applied to a candidate-gene segment observed in a COPD study, utilizes the ability of minARGs (Ancestral Recombination Graphs) to model the ancestral structural history (with mutation and recombination) by employing the information present in the SNP markers.