A Novel Approach for Deconvolution of Bulk Transcriptomic Data Using Single-Cell RNA-Seq information and Weighted Least Squares (wLS)
Tumors are composed of a complex Tumor Microenvironment (TME) consisting of dynamic interactions between malignant, immune, and stromal cells. The presence and abundance of specific cell types significantly influence tumor progression and response to therapies.
Deconvolution is a mathematical method to estimate cell type proportions from gene expression data in patient samples (RNA-Seq data). These algorithms typically use reference profiles from Single-Cell RNA-Seq (scRNA-Seq) data or predefined signature matrices, averaging gene expression for each cell type. However, current methods have limitations, such as imprecise estimates, estimation of missing cell types, and need for specific data. To address these issues, we propose a novel approach for deconvolution based on Weighted Least Squares (wLS) using Single-Cell RNA-Seq as a reference.
Keywords: Deconvolution regression weighted least squares omics data cancer bioinformatics