Format

Send to

Choose Destination
BMC Proc. 2014 Jun 17;8(Suppl 1 Genetic Analysis Workshop 18Vanessa Olmo):S49. doi: 10.1186/1753-6561-8-S1-S49. eCollection 2014.

A variance component-based gene burden test.

Author information

1
Department of Genetics, Texas Biomedical Research Institute, 7620 NW Loop 410, San Antonio, Texas 78227-5301, USA ; Centre for Genetic Origins of Health and Disease of Western Australia (M409), 35 Stirling Highway, Crawley, WA 6009, Australia.
2
Department of Genetics, Texas Biomedical Research Institute, 7620 NW Loop 410, San Antonio, Texas 78227-5301, USA.

Abstract

We propose a novel variance component approach for the analysis of next-generation sequencing data. Our method is based on the detection of the proportion of the trait phenotypic variance that can be explained by the introduction of a new variance component that accounts for the local gene-specific departure of the empirical kinship relationship matrix, estimated from single-nucleotide polymorphism (SNP) genotypes, from their theoretical expectation based on the genealogical information in the pedigree. We tested our method with simulated phenotypes and imputed SNP genotypes from the Genetic Analysis Workshop 18 data set. We observed considerable variation in the differences between theoretical and gene-specific kinship estimates that proved to be informative for our test and allowed us to detect the MAP4 causal gene at a genome-wide significance level. The distribution of our test statistic show no inflation under the null hypothesis and results from a random set of genes suggest that the detection of MAP4 is both sensitive and specific. The use of 2 different strategies for the selection of the SNPs used to derive the gene-specific empirical kinship relationship matrices provides us with suggestive evidence that our method is performing as an empirical test of linkage.

Supplemental Content

Full text links

Icon for BioMed Central Icon for PubMed Central
Loading ...
Support Center