FARDEEP

FARDEEP: Robust Immune Cell Deconvolution from Gene Expression Data

FARDEEP estimates immune cell subset composition from gene expression data using an adaptive least trimmed squares deconvolution approach to detect and remove outliers and improve accuracy of cell proportion inference.

Key Features:

• Adaptive Least Trimmed Squares: Applies an adaptive least trimmed squares algorithm to perform robust deconvolution and reduce noise contamination in gene expression profiles.
• Outlier Detection and Removal: Automatically identifies and eliminates outliers to enhance reliability of immune cell subset estimation.
• Absolute and Relative Quantification: Estimates both relative proportions and absolute quantities of immune cell subsets.

Scientific Applications:

• Tumor Immunology: Profiles immune cell infiltration in tumors to identify cold cancers, support biomarker discovery, and inform immunotherapy strategies.
• Oncology Research: Quantifies tumor immune composition to evaluate cancer prognosis and therapeutic response.

Methodology:

FARDEEP performs deconvolution of bulk gene expression data using an adaptive least trimmed squares regression framework that iteratively trims extreme residuals, removes outliers, and computes robust estimates of immune cell subset composition.