Dataset Information

Incorporating prior biologic information for high-dimensional rare variant association studies.

ABSTRACT: Given the increasing scale of rare variant association studies, we introduce a method for high-dimensional studies that integrates multiple sources of data as well as allows for multiple region-specific risk indices.Our method builds upon the previous Bayesian risk index by integrating external biological variant-specific covariates to help guide the selection of associated variants and regions. Our extension also incorporates a second level of uncertainty as to which regions are associated with the outcome of interest.Using a set of study-based simulations, we show that our approach leads to an increase in power to detect true associations in comparison to several commonly used alternatives. Additionally, the method provides multi-level inference at the pathway, region and variant levels.To demonstrate the flexibility of the method to incorporate various types of information and the applicability to high-dimensional data, we apply our method to a single region within a candidate gene study of second primary breast cancer and to multiple regions within a candidate pathway study of colon cancer.

SUBMITTER: Quintana MA

PROVIDER: S-EPMC4058572 | biostudies-literature | 2012

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Incorporating prior biologic information for high-dimensional rare variant association studies.

Quintana Melanie A MA Schumacher Fredrick R FR Casey Graham G Bernstein Jonine L JL Li Li L Conti David V DV

Human heredity 20120101 3-4

<h4>Background</h4>Given the increasing scale of rare variant association studies, we introduce a method for high-dimensional studies that integrates multiple sources of data as well as allows for multiple region-specific risk indices.<h4>Methods</h4>Our method builds upon the previous Bayesian risk index by integrating external biological variant-specific covariates to help guide the selection of associated variants and regions. Our extension also incorporates a second level of uncertainty as t ...[more]

PMID: 23594496

Similar Datasets

Project description:BackgroundThe success of collapsing methods which investigate the combined effect of rare variants on complex traits has so far been limited. The manner in which variants within a gene are selected prior to analysis has a crucial impact on this success, which has resulted in analyses conventionally filtering variants according to their consequence. This study investigates whether an alternative approach to filtering, using annotations from recently developed bioinformatics tools, can aid these types of analyses in comparison to conventional approaches.Methods & resultsWe conducted a candidate gene analysis using the UK10K sequence and lipids data, filtering according to functional annotations using the resource CADD (Combined Annotation-Dependent Depletion) and contrasting results with 'nonsynonymous' and 'loss of function' consequence analyses. Using CADD allowed the inclusion of potentially deleterious intronic variants, which was not possible when filtering by consequence. Overall, different filtering approaches provided similar evidence of association, although filtering according to CADD identified evidence of association between ANGPTL4 and High Density Lipoproteins (P = 0.02, N = 3,210) which was not observed in the other analyses. We also undertook genome-wide analyses to determine how filtering in this manner compared to conventional approaches for gene regions. Results suggested that filtering by annotations according to CADD, as well as other tools known as FATHMM-MKL and DANN, identified association signals not detected when filtering by variant consequence and vice versa.ConclusionIncorporating variant annotations from non-coding bioinformatics tools should prove to be a valuable asset for rare variant analyses in the future. Filtering by variant consequence is only possible in coding regions of the genome, whereas utilising non-coding bioinformatics annotations provides an opportunity to discover unknown causal variants in non-coding regions as well. This should allow studies to uncover a greater number of causal variants for complex traits and help elucidate their functional role in disease.

Dataset Information

Incorporating prior biologic information for high-dimensional rare variant association studies.

Publications

Incorporating prior biologic information for high-dimensional rare variant association studies.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure