Environmental health studies are increasingly measuring multiple pollutants to characterize the joint health effects attributable to exposure mixtures. However, the underlying dose-response relationship between toxicants and health outcomes of interest may be highly nonlinear, with possible nonlinear interaction effects. Hierarchical integrative group least absolute shrinkage and selection operator (HiGLASSO) is a general framework to identify noteworthy nonlinear main and interaction effects in the presence of group structures among a set of exposures.
higlasso can be installed via Github using
devtools
# install.packages("devtools")
devtools::install_github("umich-cphds/higlasso")You’ll need a working C++11 compiler, which can obtained by installing Xcode on MacOS, and RTools on Windows.
higlasso can be slow, so it may may be beneficial to
tweak some its settings (for example, nlambda1 and
nlambda2) to get a handle on how long the method will take
before running the full model.
library(higlasso)
set.seed(48109)
X <- as.matrix(higlasso.df[, paste0("V", 1:10)])
Y <- higlasso.df$Y
Z <- matrix(1, nrow(X))
# This can take a bit of time
fit <- cv.higlasso(Y, X, Z)
print(fit)
## 'cv.higlaso' fit:
## Average cross validation error for each (lambda1, lambda2)
## l2.1 l2.2 l2.3 l2.4 l2.5 l2.6 l2.7 l2.8
## l1.1 31.27646 30.16180 26.95437 27.53877 29.02558 30.07576 26.47052 31.89449
## l1.2 29.41992 29.10689 27.10846 26.87978 27.47474 27.63711 26.47278 26.85553
## l1.3 27.54288 29.05916 26.93971 27.08178 28.45151 29.09442 24.79442 21.60319
## l1.4 26.54363 28.53124 26.53870 25.96099 26.54686 28.31096 30.42201 28.16428
## l1.5 26.81162 27.25271 24.09822 25.15356 22.58817 23.14864 27.22033 30.09574
## l1.6 27.16120 27.58498 23.66292 22.17489 22.00758 24.25419 26.21524 29.39816
## l1.7 27.22267 27.51427 23.86357 22.68843 22.47261 25.06126 29.36773 29.38315
## l1.8 27.38411 27.67919 24.78215 22.76672 21.47550 24.58022 31.35179 29.82478
## l1.9 27.50484 27.74833 24.55036 23.56705 22.28636 24.74533 31.95273 36.06778
## l1.10 27.55453 27.81106 25.47948 24.48003 22.49150 23.45593 30.46539 30.89732
## l2.9 l2.10
## l1.1 30.40880 29.62060
## l1.2 26.55717 30.37125
## l1.3 25.07946 30.30343
## l1.4 27.52339 27.48728
## l1.5 30.18009 30.65290
## l1.6 28.36495 28.98296
## l1.7 29.70256 26.97493
## l1.8 31.19091 36.82322
## l1.9 28.48208 25.13074
## l1.10 29.82692 33.11209
## Lambda min:
## 0.292944 0.9037191
## Lambda 1 SE:
## 0.7951724 0.9037191If you encounter a bug, please open an issue on the Issues tab on Github or send us an email.
For questions or feedback, please email Jonathan Boss at bossjona@umich.edu or Alexander Rix alexrix@umich.edu.
A Hierarchical Integrative Group LASSO (HiGLASSO) Framework for Analyzing Environmental Mixtures. Jonathan Boss, Alexander Rix, Yin-Hsiu Chen, Naveen N. Narisetty, Zhenke Wu, Kelly K. Ferguson, Thomas F. McElrath, John D. Meeker, Bhramar Mukherjee. 2020. arXiv:2003.12844