R/writeMIDAS.R

In CellNOptR: Training of boolean logic models of signalling networks using prior knowledge networks and perturbation data

writeMIDAS <- function(CNOlist, filename, timeIndices=NULL, overwrite=FALSE)
{

    if ((class(CNOlist)=="CNOlist")==FALSE){
         cnolist = CellNOptR::CNOlist(CNOlist)
    }
    else{
        cnolist = CNOlist
    }

    # remove exiisting file if required
    if (overwrite==TRUE){
        if (file.exists(filename)==TRUE){
            file.remove(filename)
        }
    } else {
        if (file.exists(filename)==TRUE){
            stop(paste("File ", filename, "already exists.",  sep=""))
        }
    }

    timeSignals = cnolist@timepoints
    namesCues = colnames(cnolist@cues)
    namesStimuli = colnames(cnolist@stimuli)
    namesInhibitors = colnames(cnolist@inhibitors)
    namesSignals = colnames(cnolist@signals[[1]])

    #nCues = length(namesCues)
    nStimuli = length(namesStimuli)
    nInhibitors = length(namesInhibitors)
    nSignals = length(namesSignals)
    input_nTimes = length(timeSignals)
    output_nTimes = length(timeSignals)
    # if timeIndices provided, the inputs nTimes and output nTimes are different
    if (is.null(timeIndices)==TRUE){
        timeIndices = 1:input_nTimes
    } else{
        output_nTimes = length(timeIndices)
    }

    # First column is cellline
    nCols = 1 + nStimuli + nInhibitors + nSignals * 2 
    nConds = dim(cnolist@signals[[1]])[1]
    nRows = nConds * output_nTimes

    # build the header
    colNames = c("TR:mock:CellLine")
    for (x in namesStimuli){
        colNames = cbind(colNames, paste("TR:", x, sep=""))
    }
    for (x in namesInhibitors){
        colNames = cbind(colNames, paste("TR:", x, "i", sep=""))
    }
    for (x in namesSignals){
        colNames = cbind(colNames, paste("DA:", x, sep=""))
    }
    for (x in namesSignals){
        colNames = cbind(colNames, paste("DV:", x, sep=""))
    }

    data = matrix(NA, nRows, nCols)

    # cellline is made of ones for now
    data[,1] = 1

    # fill the experiments first


    for (time in seq_along(timeIndices)){

            i1 = 1 + (time - 1) * nConds
            i2 = i1 + nConds - 1

            j1 = 2; j2 = j1 + nStimuli - 1

            data[i1:i2, j1:j2] = cnolist@stimuli

            if (nInhibitors>0){
                j1 = j2 + 1; j2 = j1 + nInhibitors - 1
                data[i1:i2, j1:j2] = cnolist@inhibitors
            } 

            j1 = j2 + 1; j2 = j1 + nSignals - 1
            data[i1:i2, j1:j2] = rep(timeSignals[timeIndices[time]], nSignals)

            j1 = j2 + 1; j2 = j1 + nSignals - 1
            data[i1:i2,j1:j2] = cnolist@signals[[timeIndices[time]]] 
    }


    # and save it into a file
    write.table(data, file=filename,
        row.names=FALSE, col.names=colNames,quote=FALSE,sep=",")

}