A feasible roadmap for unsupervised deconvolution of two-source mixed gene expressions

TL;DR

This paper proposes a mathematically grounded, unsupervised method for deconvolving two-source mixed gene expressions from heterogeneous samples, estimating proportions and cell-specific profiles without prior knowledge, validated on simulated and real data.

Contribution

It introduces a novel unsupervised deconvolution framework that does not require prior marker gene knowledge, enabling analysis of mixed gene expressions in heterogeneous tissues.

Findings

Successful deconvolution on simulated data

Effective application to real gene expression data

Potential to replace costly experimental solutions

Abstract

Tissue heterogeneity is a major confounding factor in studying individual populations that cannot be resolved directly by global profiling. Experimental solutions to mitigate tissue heterogeneity are expensive, time consuming, inapplicable to existing data, and may alter the original gene expression patterns. Here we ask whether it is possible to deconvolute two-source mixed expressions (estimating both proportions and cell-specific profiles) from two or more heterogeneous samples without requiring any prior knowledge. Supported by a well-grounded mathematical framework, we argue that both constituent proportions and cell-specific expressions can be estimated in a completely unsupervised mode when cell-specific marker genes exist, which do not have to be known a priori, for each of constituent cell types. We demonstrate the performance of unsupervised deconvolution on both simulation and real gene expression data, together with perspective discussions.