updateNets: Synchronize patient set in sample table and network table.
In BaderLab/netDx: Network-based patient classifier

Description Usage Arguments Details Value Examples

Synchronize patient set in sample table and network table.

updateNets(
  p_net,
  pheno_DF,
  writeNewNets = TRUE,
  oldNetDir,
  newNetDir,
  verbose = TRUE,
  ...
)

`p_net`	(matrix) rows are patients, columns are networks. a[i,j] = 1 if patient i occurs in network j, else 0.
`pheno_DF`	(data.frame) patient ID and STATUS.
`writeNewNets`	(logical) if TRUE writes new networks to `newNetDir`.
`oldNetDir`	(char) path to directory with networks to be updated
`newNetDir`	(char) path to directory where updated networks are to be written
`verbose`	(logical) print messages
`...`	passed to pruneNets()

This function is useful in applications with highly missing data or where each patient contributes data points not present in the others; e.g. networks based on individual patient CNVs, which are highly sparse. In such a scenario, any kind of patient subsetting - for example, limiting to training samples - changes the population of eligible networks for analysis. Networks that no longer have samples, or that have one patient with the neighbour removed, have to be excluded. This function updates networks and patients so that each network contains at least two patients and only patients in networks are retained. In other words, it keeps pheno_DF and p_net in sync.

list with updated p_net and pheno_DF. pheno_DF will contain IDs in the updated p_net. p_net will contain only those networks with 2+ patients and those patients present in 1+ network.

data(npheno)
netDir <- system.file("extdata","example_nets",package="netDx")
netmat <- countPatientsInNet(netDir,dir(netDir,pattern='txt$'), npheno[,1])
x <- updateNets(netmat, npheno,writeNewNets=FALSE)