The power comparison of the haplotype-based collapsing tests and the variant-based collapsing tests for detecting rare variants in pedigrees

Wei Guo,Yin Shugart

doi:10.1186/1471-2164-15-632

Abstract

BackgroundBoth common and rare genetic variants have been shown to contribute to the etiology of complex diseases. Recent genome-wide association studies (GWAS) have successfully investigated how common variants contribute to the genetic factors associated with common human diseases. However, understanding the impact of rare variants, which are abundant in the human population (one in every 17 bases), remains challenging. A number of statistical tests have been developed to analyze collapsed rare variants identified by association tests. Here, we propose a haplotype-based approach. This work inspired by an existing statistical framework of the pedigree disequilibrium test (PDT), which uses genetic data to assess the effects of variants in general pedigrees. We aim to compare the performance between the haplotype-based approach and the rare variant-based approach for detecting rare causal variants in pedigrees.ResultsExtensive simulations in the sequencing setting were carried out to evaluate and compare the haplotype-based approach with the rare variant methods that drew on a more conventional collapsing strategy. As assessed through a variety of scenarios, the haplotype-based pedigree tests had enhanced statistical power compared with the rare variants based pedigree tests when the disease of interest was mainly caused by rare haplotypes (with multiple rare alleles), and vice versa when disease was caused by rare variants acting independently. For most of other situations when disease was caused both by haplotypes with multiple rare alleles and by rare variants with similar effects, these two approaches provided similar power in testing for association.ConclusionsThe haplotype-based approach was designed to assess the role of rare and potentially causal haplotypes. The proposed rare variants-based pedigree tests were designed to assess the role of rare and potentially causal variants. This study clearly documented the situations under which either method performs better than the other. All tests have been implemented in a software, which was submitted to the Comprehensive R Archive Network (CRAN) for general use as a computer program named rvHPDT.

Highlights

Both common and rare genetic variants have been shown to contribute to the etiology of complex diseases
To highlight the importance of haplotype-based approaches, we provide a brief review on the studies detected haplotypes composed of multiple rare variants contributing to disease or traits related to lipid metabolism
Type I error rates and analytical power associated with six tests were assessed using the nuclear families simulated both by forsim and by the direct simulation under the special haplotype setting

Summary

Introduction

Both common and rare genetic variants have been shown to contribute to the etiology of complex diseases. We aim to compare the performance between the haplotype-based approach and the rare variant-based approach for detecting rare causal variants in pedigrees Genetic studies such as genome-wide association studies (GWAS) have typically relied on the common disease/ common variant (CDCV) paradigm, and the GWAS approach has had its share of success with some genetic disorders. Some rare variants were found to play a causal role in human diseases, including psychiatric disorders This led to the shift from the CDCV approach to a common disease-rare variant (CDRV) approach. Given this increased recognition that common disorders may reflect the aggregated effects of many rare variants, genetic analysts investigating this new hypothesis have had to deal with a host of new challenges. The more affordable whole genome sequencing (WGS) technology demands new analytical tools to analyze these data

Objectives

Methods

Results

Conclusion