run.lean: Run the LEAN approach

In LEANR: Finds "Local Subnetworks" Within an Interaction Network which Show Enrichment for Differentially Expressed Genes

Description

Apply the LEAN approach to a given network and a list of pvalues

Usage

run.lean(ranking, network, ranked = F, 
    add.scored.genes = F, keep.nodes.without.scores = F, 
    verbose = F, n_reps = 10000, bootstrap = F, ncores = NULL)

Arguments

ranking	Either a file containing gene p-values or a named numerical vector of p-values with names matching node names used in the network
network	Either a file containing the network in sif format or an igraph graph object representing the network
ranked	whether to transform input p-values into a uniformly distributed list of p-values based on the genes' rank before p* calculation
add.scored.genes	whether to create one singleton node for each gene with a score but not occurring in the graph
keep.nodes.without.scores	whether to keep nodes of the graph that have no recorded score. For those nodes it is still possible to compute enrichment scores if at least one of their network neighbors has a recorded score.
verbose	whether to print additional status messages
n_reps	the number of samples each background distribution should consist of. Largely influences the run-time, but higher values needed for meaningful empirical pvalues !
bootstrap	whether to draw the pvalues of the background distributions with or without replacement
ncores	number of cores to be used in parallel computation. Default (NULL) leads to automatic guessing of max number of cores to be used (depending on operating system).

Value

A list object containing the results of the LEAN run. The list encompasses the following elements:

restab	Result table of applying LEAN to the real data
randtab	Result table of applying LEAN to a permuted p-value list
indGraph	igraph graph representing the input network after adapting it according to parameters <add.scored.genes>, <keep.nodes.without.scores> and the presence of gene scores in the input scores
nhs	The extracted local subnetworks. Encoded as a named (by protein/gene ids) list of igraph node indices detailing each evaluated local subnetwork
gene.scores	The gene p-values extracted from the input scores. Encoded as a numeric vector named with protein/gene ids

Author(s)

Frederik Gwinner

References

Gwinner et al., Network-based analysis of omics data: The LEAN method, MS submitted to Bioinformatics

See Also

LEANR-package

Examples

## Simple use case starting from a test network and p-value list
## Not run:  
# compute LEAN p-values starting from a p-value file and a network file
rank_file<-system.file('extdata/pvals_red.txt.gz', package='LEANR')
net_file<-system.file('extdata/g_red.sif.gz', package='LEANR')
system.time(res<-run.lean(ranking=rank_file, network=net_file, 
  add.scored.genes=T, verbose=T, n_reps=1000, ncores=3))

# compute LEAN p-values starting from a list of gene scores and a graph
data(pvals_red)
data(g_red)
system.time(res2<-run.lean(ranking=pvals_red, network=g_red, verbose=T, 
   n_reps=1000, ncores=3))


## End(Not run) 

LEANR documentation built on May 2, 2019, 3:46 p.m.