High-dimensional covariate-augmented multi-study nonlinear factor model
=========================================================================
We introduce a novel modeling strategy for analyzing single-cell multi-omics data sourced from various studies. Our approach is driven by the surge in single-cell multi-omics research in biological and medical fields, exemplified by the RNA-protein simultaneous sequencing case-control study on human peripheral blood mononuclear cells (PBMC). To capture the nonlinear relationships between diverse molecular types, including genes and proteins, and to identify both study-shared and study-specific features while accounting for the count nature of the data, we propose a multi-study nonlinear factor model incorporating study-shared and study-specific factors, along with augmented omics variables as covariates.
Check out the Package Website for a more complete description of the methods and analyses.
Installation
“MultiCOAP” depends on the ‘Rcpp’ and ‘RcppArmadillo’ package, which requires appropriate setup of computer. For the users that have set up system properly for compiling C++ files, the following installation command will work.
Method 1：
if (!require(“remotes”, quietly = TRUE))
remotes::install_github(“feiyoung/MultiCOAP”)
Method 2: install from CRAN
install.packages(“MultiCOAP”)