Maintainer: Andrew Chen, chenandr@musc.edu
License: Artistic License 2.0
SiMMR is a distance-based approach for analysis of multiple data modalities. Several test statistics are proposed, which are all implemented in the simmr function. Further details are available in our preprint (see Citations).
The R package can be installed via devtools by running the following code
# install.packages("devtools")
devtools::install_github("andy1764/SiMMR", build_vignettes = TRUE)
SiMMR can be run by first computing distance matrices for each modality, then inputting as a list of distance matrices to simmr. Below is an example call using simulated data:
D <- list(dist(rnorm(10)), dist(rnorm(10)), dist(rnorm(10)))
X <- list("var" = runif(10))
simmr(D, X, "var")
Please cite the following article:
Chen, A. A., Weinstein, S. M., Adebimpe, A., Gur, R. C., Gur, R. E., Merikangas, K. R., Satterthwaite, T. D., Shinohara, R. T., & Shou, H. (2023). Similarity-based multimodal regression. Biostatistics, kxad033. https://doi.org/10.1093/biostatistics/kxad033