We present Unico, a unified cross-omics method designed to deconvolve standard 2-dimensional bulk matrices of samples by features into a 3-dimensional tensors representing samples by features by cell types. Unico stands out as the first principled model-based deconvolution method that is theoretically justified for any heterogeneous genomic data.

All necessary scripts used for the analyses reported in the manuscript can be found under folder Rscripts

Unico will be available soon on CRAN as an R package

Install from GitHub

if (!require("devtools", quietly = TRUE))

Version info

Our package is tested on both R 3.6.1 and R 4.1.0, on Windows, Linux and MacOS based machines.


Please head to this vignette for a step by step tutorial on (1) deconvolving a simulated PBMC pseudo-bulk expression dataset and (2) association testing on subsets of publicly available methylation datasets.


This software is developed by Zeyuan Johnson Chen (johnsonchen@cs.ucla.edu) and Elior Rahmani (EliorRahmani@mednet.ucla.edu).


Unico is available under the GPL-3 license.