ryananeff/superNOVA: Differential coexpression within multiple subgroups of data

Calculates differential coexpression of multiple feature pairs between two or more subgroups of samples, using an ANOVA test for multiple correlations versus a global correlation. This test is based on the Chow statistic, which uses the sum of squared residuals between correlation submodels and a global model of correlation between each feature pair. Useful for determining which genes and gene pairs have been perturbed in gene expression experiments across clinically or biologically relevant subgroups in a large population of samples, such as subtypes of health and disease. Also useful for determining differences between a global gene coexpression network edges and subgroup-specific gene coexpression networks. Can rapidly scale to >billions of feature pairs using both a multicore and multinode parallelization scheme, enabling multi-omic feature pair comparisons for millions of input features.

install.packages("remotes")
remotes::install_github("ryananeff/superNOVA")