The R package SPRING (Semi-Parametric Rank-based approach for
INference in Graphical model) estimates sparse microbial association
networks using rank-based correlation with sparse graphical modeling
techniques. The corresponding reference is
Yoon G., Gaynanova I. and Müller C.L. (2019) Microbial Networks in SPRING - Semi-parametric Rank-Based Correlation and Partial Correlation Estimation for Quantitative Microbiome Data. Frontiers in Genetics, 10:516.
Latent correlation estimation is provided by the R package latentcor
(Huang Z., Guo M., Müller C.L. and Gaynanova I. (2021) latentcor: An R
Package for Estimating Latent Correlations from Mixed Data
Types. Journal of Open Source
Software, 6(65), 3634).
The package is being submitted to Bioconductor. Until it is available there, install from GitHub:
# install.packages("remotes")
remotes::install_github("IrinaStatsLab/SPRING")library(SPRING)
data("QMP") # 106 samples and 91 taxa
# Apply SPRING on QMP data
fit <- SPRING(QMP, quantitative = TRUE, nlambda = 50, rep.num = 50,
verbose = FALSE)
# StARS-selected optimal lambda index (default threshold = 0.1)
opt.index <- fit$output$stars$opt.index
# Estimated adjacency matrix (1 = edge, 0 = no edge)
adj <- as.matrix(fit$fit$refit$stars)
# Symmetrized MB coefficients at optimal lambda
beta <- as.matrix(fit$fit$est$beta[[opt.index]])
beta_sym <- (beta + t(beta)) / 2