README.md

In yuliangwang/perturb.met: What the Package Does (One Line, Title Case)

perturb.met

perturb-Met can identify transcriptional perturbations centered at hundreds of metabolites in a genome-scale metabolic network, using single cell or single nucleus RNA-seq data. perturb-Met can be used to identify metabolic changes between any two conditions, including using bulk transcriptomic data (RNA-seq or microarray). perturb-Met adapts the Kullback-Leibler (KL) divergence concept from information theory to quantify relative changes in the distribution of transcript abundances for all genes connected to a metabolite between two conditions (e.g., AD vs. control). KL divergence quantifies how a statistical distribution differs from a second, reference distribution.

perturb-Met offers several advantages over existing methods. First, perturb-Met addresses both problems by considering relative changes of transcript abundances at individual metabolites, thus it has higher resolution than pathway-level analysis. Second, unlike the Reproter Metabolite analysis method, perturb-Met does not require a a minimum and a maximum number of genes connected to a metabolite (e.g., minimum 3 genes, maximum 50 genes). This avoid the problem of missing important metabolites connected by only two genes. * Third, even if more than 3 genes were connected to a metabolite, Reporter Metabolite analysis would still require enough of them to be significantly differentially expressed. A metabolite may not be selected if only one of its connected genes is differentially expressed. However, using perturb-Met, it is still considered noteworthy if the magnitude of expression change for one gene significantly alters the relative distribution of transcript abundances of all genes around a metabolite.

To install perturb-Met R package, you need to first install the devltools R package, then run:

library(devtools)
install_github("yuliangwang/perturb.met")
library(perturb.met)

You can check out how to use perturb-Met in the demo on Rpubs. Learn more about perturb-Met and its application on single cell RNA-seq of Alzheimer's disease and metabolic differences across neuronal subtypes in AD by reading my preprint: Identifying neuron subtype-specific metabolic network changes in single cell transcriptomics of Alzheimer's Disease using perturb-Met.

yuliangwang/perturb.met documentation built on Jan. 21, 2021, 12:27 a.m.