⚠️ This repository is based on the very early version of DNB, please refer to the latest paper for your research.

An R package to calculate Dynamical Network Biomarkers (DNB).

For more information, see:

Chen L, Liu R, Liu Z-P, Li M, Aihara K. Detecting early-warning signals for sudden deterioration of complex diseases by dynamical network biomarkers. Scientific Reports. 2012;2:342. doi:10.1038/srep00342

devtools::install_github("gpli/DNB")

library(DNB)

# Load data
gene_expr <- read.csv(system.file("extdata", "gene_expr.csv", package = "DNB"), row.names = 1)
sample_info <- read.csv(system.file("extdata", "sample_info.csv", package = "DNB"), row.names = 1)

# New DNB object
# all-zero genes at any time point will be removed
dnb <- new_DNB(data = gene_expr,
               time = as.factor(sample_info$time))

# Calulate correlation and coefficient of variation
dnb <- cal_cor(dnb)
dnb <- cal_cv(dnb)

# Search DNB
dnb <- search_candidates(dnb, min_size = 20, max_size = 100)
dnb <- cal_final(dnb)

# Get results
dnb_genes <- get_DNB_genes(dnb)
candidates <- get_candidates(dnb)
plot_DNB(candidates)

final <- get_final(dnb)
plot_DNB(final)

Set group in new_DNB to define normal samples (control group).

dnb <- new_DNB(data = gene_expr,
               time = as.factor(sample_info$time),
               group = as.factor(sample_info$group))

By default, control group will be considered during searching DNB, but you can still change your mind by set with_ctrl = F in search_candidates and cal_final. This is useful for tuning parameters.

Calculating correlation are often memory consuming if there are too many genes, and difficult to load them all if there are too many time points. You can set use_bigcor = T to calculate correlation with propagate::bigcor and/or set data_dir to store correlation data to local, i.e.

cal_cor(dnb, use_bigcor = T, data_dir = "tmp")

It is recommended to pre-filter low expression genes, which will improve calculation speed.

plot_DNB use ggplot2 to draw results. You can set outfile to write into a image file. Further arguments were passed to ggplot2::ggsave.

plot_DNB(final, "DNB_final.pdf")