DeMixT

DeMixT performs deconvolution on transcriptome data from heterogeneous tumor samples with two or three components using expression data from RNAseq or microarray platforms.

Key Features:

• Multi-component deconvolution: DeMixT deconvolves high-dimensional data from mixtures of more than two components, estimating both component-specific proportions and expression profiles for individual samples.
• Iterated conditional mode algorithm: DeMixT implements an iterated conditional mode algorithm for deconvolution.
• Gene-set-based component merging: DeMixT employs a novel gene-set-based component merging approach to improve accuracy.

Scientific Applications:

• Tumor-immune deconvolution: DeMixT deconvolves transcriptome data from heterogeneous tumor samples with immune infiltration.
• TCGA data analysis: DeMixT has been applied to TCGA data for improved deconvolution of tumor transcriptomic data.
• Clinical outcome linkage: Improved deconvolution with DeMixT enables linking tumor transcriptomic data with clinical outcomes.

Methodology:

DeMixT implements an iterated conditional mode algorithm and a novel gene-set-based component merging approach.