Variable selection in varying coefficient models for mapping quantitative trait loci
Public DepositedAdd to collection
You do not have access to any existing collections. You may create a new collection.
Downloadable Content
Download PDFCitation
MLA
Gong, Yi. Variable Selection In Varying Coefficient Models for Mapping Quantitative Trait Loci. Chapel Hill, NC: University of North Carolina at Chapel Hill, 2011. https://doi.org/10.17615/tpy5-yj62APA
Gong, Y. (2011). Variable selection in varying coefficient models for mapping quantitative trait loci. Chapel Hill, NC: University of North Carolina at Chapel Hill. https://doi.org/10.17615/tpy5-yj62Chicago
Gong, Yi. 2011. Variable Selection In Varying Coefficient Models for Mapping Quantitative Trait Loci. Chapel Hill, NC: University of North Carolina at Chapel Hill. https://doi.org/10.17615/tpy5-yj62- Last Modified
- March 21, 2019
- Creator
-
Gong, Yi
- Affiliation: Gillings School of Global Public Health, Department of Biostatistics
- Abstract
- The Collaborative Cross (CC), a renewable mouse resource that mimics the genetic diversity in humans, provides great data sources for mapping Quantitative Trait Loci (QTL). The recombinant inbred intercrosses (RIX) generated from CC recombinant inbred (RI) lines have several attractive features and can be produced repeatedly. Many quantitative traits are inherently complex and change with other covariates. To map such complex traits, phenotypes are measured across multiple values of covariates on each subject. In the first topic, we propose a more flexible nonparametric varying coefficient QTL mapping method for RIX data. This model lets the QTL effects evolve with certain covariates, and naturally extends classical parametric QTL mapping methods. Simulation results indicate that the varying coefficient QTL mapping has substantially higher power and higher mapping precision compared to parametric models when the assumption of constant genetic effects fails. We model the time-varying genetic effects with functional approximation using B-spline basis. We apply a nested permutation method to obtain threshold values for QTL detection. In the second topic, we extend the single marker QTL mapping to multiple QTL mapping. We treat multiple QTL mapping as a model/variable selection problem and propose a penalized mixed effects model. We apply a penalty function for the group selection of coefficients associated with each gene. We propose new selection procedures for tuning parameters. Simulations showed that the new mapping method performs better than the single marker analysis when multiple QTL exist. Last, in the third topic, we extend the multiple QTL mapping method to longitudinal data. We pay special attention to modeling the covariance structure of repeated measurements. Popular stationary assumptions on variance and covariance structures may not be realistic for many longitudinal traits. The structured antedependence (SAD) model is a parsimonious covariance model that allows for both nonstationary variance and correlation. We propose a penalized likelihood method for multiple QTL mapping using the SAD model. Simulation results showed the model selection method outperforms the single marker analysis. Furthermore, the performance of multiple QTL mapping will be affected if the covariance model is misspecified.
- Date of publication
- December 2011
- DOI
- Resource type
- Rights statement
- In Copyright
- Note
- "... in partial fulfillment of the requirements for the degree of Doctor of Philosophy in the Department of Biostatistics."
- Advisor
- Zou, Fei
- Degree granting institution
- University of North Carolina at Chapel Hill
- Language
- Publisher
- Place of publication
- Chapel Hill, NC
- Access right
- Open access
- Date uploaded
- March 18, 2013
Relations
- Parents:
This work has no parents.
Items
Thumbnail | Title | Date Uploaded | Visibility | Actions |
---|---|---|---|---|
2019-04-10 | Public | Download |