Elastic-net regularization approaches for genome-wide association studies of rheumatoid arthritis

Cho, Seoae; Kim, Haseong; Oh, Sohee; Kim, Kyunga; Park, Taesung

doi:10.1186/1753-6561-3-s7-s25

Cited by 67 publications

(71 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The elastic net [19] is a hybrid regularization and variable selection method that linearly combines the L1 and L2 regularization penalties of the lasso [21] and ridge [see, e.g., 22] regression methods in multiple regression. This combination of lasso and ridge penalties provides a more precise prediction than using multiple regression, when SNVs are in high linkage disequilibrium [23] . In addition, the elastic net can accommodate situations in which the number of predictors exceeds the number of observations.…”

Section: I P Imentioning

confidence: 99%

Using Gene Genealogies to Localize Rare Variants Associated with Complex Traits in Diploid Populations

Karunarathna¹,

Graham²

2018

Hum Hered

View full text Add to dashboard Cite

Background and Aims: Many methods can detect trait association with causal variants in candidate genomic regions; however, a comparison of their ability to localize causal variants is lacking. We extend a previous study of the detection abilities of these methods to a comparison of their localization abilities. Methods: Through coalescent simulation, we compare several popular association methods. Cases and controls are sampled from a diploid population to mimic human studies. As benchmarks for comparison, we include two methods that cluster phenotypes on the true genealogical trees: a naive Mantel test considered previously in haploid populations and an extension that takes into account whether case haplotypes carry a causal variant. We first work through a simulated dataset to illustrate the methods. We then perform a simulation study to score the localization and detection properties. Results: In our simulations, the association signal was localized least precisely by the naive Mantel test and most precisely by its extension. Most other approaches had intermediate performance similar to the single-variant Fisher exact test. Conclusions: Our results confirm earlier findings in haploid populations about potential gains in performance from genealogy-based approaches. They also highlight differences between haploid and diploid populations when localizing and detecting causal variants.

show abstract

Section: I P Imentioning

confidence: 99%

Using Gene Genealogies to Localize Rare Variants Associated with Complex Traits in Diploid Populations

Karunarathna¹,

Graham²

2018

Hum Hered

View full text Add to dashboard Cite

show abstract

“…Cho et al [2009] and Niu et al [2009] performed, respectively, a series of single-marker logistic regression and log-linear analyses to screen each SNP separately. Cho et al [2009] included sex as a covariate; Niu et al [2009] did not include any covariates.…”

Section: Methodsmentioning

confidence: 99%

“…Cho et al [2009] included sex as a covariate; Niu et al [2009] did not include any covariates. Cho et al [2009] used only an additive model for their analysis while Niu et al [2009] tested each SNP with three models: dominant, recessive, and additive after using the genomic control method [Devlin and Roeder, 1999] to control for spurious associations induced by population structure. The determination of which SNPs to include at Stage 2 was a practical decision for Cho et al [2009]: they selected the largest number of SNPs that could be handled in their computationally intensive penalized logistic regression analysis in Stage 2.…”

Section: Methodsmentioning

confidence: 99%

Multistage analysis strategies for genome‐wide association studies: summary of group 3 contributions to Genetic Analysis Workshop 16

Neuman

Sung

2009

Genetic Epidemiology

View full text Add to dashboard Cite

This contribution summarizes the work done by six independent teams of investigators to identify the genetic and non-genetic variants that work together or independently to predispose to disease. The theme addressed in these studies is multistage strategies in the context of genome-wide association studies. The work performed comes from Group 3 of the Genetic Analysis Workshop 16 held in St. Louis, Missouri in September, 2008. These six studies represent a diversity of multistage methods of which five are applied to the North American Rheumatoid Arthritis Consortium rheumatoid arthritis case-control data, and one method is applied to the low-density lipoprotein phenotype in the Framingham Heart Study simulated data. In the first stage of analyses, the majority of studies used a variety of screening techniques to reduce the noise of single-nucleotide polymorphisms purportedly not involved in the phenotype of interest. Three studies analyzed the data using penalized regression models, either LASSO or the elastic net. The main result was a reconfirmation of the involvement of variants in the HLA region on chromosome 6 with rheumatoid arthritis. The hope is that the intense computational methods highlighted in this group of papers will become useful tools in future genome-wide association studies.

show abstract

“…With penalized regression, researchers are using different criteria to rank the SNPs, including effect size (Cho et al, 2009), self-developed significance (Wu et al, 2009), etc. To keep our situation simple, we provide a list of top SNPs sorted by effect size from pLASSO and LASSO separately.…”

Section: Real Data Analysismentioning

confidence: 99%

Variable Selection With Prior Information for Generalized Linear Models via the Prior LASSO Method

Jiang¹,

He²,

Zhang³

2016

Journal of the American Statistical Association

View full text Add to dashboard Cite

LASSO is a popular statistical tool often used in conjunction with generalized linear models that can simultaneously select variables and estimate parameters. When there are many variables of interest, as in current biological and biomedical studies, the power of LASSO can be limited. Fortunately, so much biological and biomedical data have been collected and they may contain useful information about the importance of certain variables. This paper proposes an extension of LASSO, namely, prior LASSO (pLASSO), to incorporate that prior information into penalized generalized linear models. The goal is achieved by adding in the LASSO criterion function an additional measure of the discrepancy between the prior information and the model. For linear regression, the whole solution path of the pLASSO estimator can be found with a procedure similar to the Least Angle Regression (LARS). Asymptotic theories and simulation results show that pLASSO provides significant improvement over LASSO when the prior information is relatively accurate. When the prior information is less reliable, pLASSO shows great robustness to the misspecification. We illustrate the application of pLASSO using a real data set from a genome-wide association study.

show abstract

Elastic-net regularization approaches for genome-wide association studies of rheumatoid arthritis

Cited by 67 publications

References 13 publications

Using Gene Genealogies to Localize Rare Variants Associated with Complex Traits in Diploid Populations

Using Gene Genealogies to Localize Rare Variants Associated with Complex Traits in Diploid Populations

Multistage analysis strategies for genome‐wide association studies: summary of group 3 contributions to Genetic Analysis Workshop 16

Variable Selection With Prior Information for Generalized Linear Models via the Prior LASSO Method

Contact Info

Product

Resources

About