Accounting for established predictors with the multistep elastic net

Abstract

Multivariable models for prediction or estimating associations with an outcome are rarely built in isolation. Instead, they are based upon a mixture of covariates that have been evaluated in earlier studies (eg, age, sex, or common biomarkers) and covariates that were collected specifically for the current study (eg, a panel of novel biomarkers or other hypothesized risk factors). For that context, we present the multistep elastic net (MSN), which considers penalized regression with variables that can be qualitatively grouped based upon their degree of prior research support: established predictors vs unestablished predictors. The MSN chooses between uniform penalization of all predictors (the standard elastic net) and weaker penalization of the established predictors in a cross‐validated framework and includes the option to impose zero penalty on the established predictors. In simulation studies that reflect the motivating context, we show the comparability or superiority of the MSN over the standard elastic net, the Integrative LASSO with Penalty Factors, the sparse group lasso, and the group lasso, and we investigate the importance of not penalizing the established predictors at all. We demonstrate the MSN to update a prediction model for pediatric ECMO patient mortality.

Food Protein-Induced Allergic Proctocolitis: The Effect of Maternal Diet During Pregnancy and Breastfeeding in a Mediterranean Population

“…P -values ≤ 0.05 were considered statistically significant. To investigate the FPIAP risk in relation to the maternal diet during pregnancy and breastfeeding, the elastic net regression model was applied, which combines Lasso and Ridge regression ( 15 ), using the glmnet package in R (version 4.1-1).…”

Integrating 3D genomic and epigenomic data to enhance target gene discovery and drug repurposing in transcriptome-wide association studies

“…The model utilizes both ridge and lasso penalties, as decided by the tuning parameter . is another tuning parameter that controls the penalty on the essential predictors relative to the non-essential predictors 19 . We assume so that essential predictors are penalized no more than the non-essential predictors.…”

Genetic imputation of kidney transcriptome, proteome and multi-omics illuminates new blood pressure and hypertension targets

“…We use X 1 to represent the set of essential genetic variants, X 2 to denote “non-essential” genetic variants, and E to represent the vector of gene expression levels across multiple individuals. ϕ is a tuning parameter that controls the penalty on the “essential” predictors relative to the non-essential predictors 139 . We assume ϕ≤ 1 so that “essential” predictors are penalised no more than the non-essential predictors.…”