make.decision.tree: Creates a decision tree to classify samples using the...
In Pigengene: Infers biological signatures from gene expression data

Description Usage Arguments Details Value Note See Also Examples

A decision tree in Pigengene-package uses module eigengenes to build a classifier that distinguishes the different classes. Briefly, each eigengene is a weighted average of the expression of all genes in the module, where the weight of each gene corresponds to its membership in the module.

make.decision.tree(pigengene, Data, 
  Labels = structure(pigengene$annotation[rownames(pigengene$eigengenes),
          1], names = rownames(pigengene$eigengenes)),
  testD = NULL, testL = NULL, selectedFeatures = NULL,
  saveDir = "C5Trees", minPerLeaf = NULL, useMod0 = FALSE, 
  costRatio = 1, toCompact = NULL, noise = 0, noiseRepNum = 10, doHeat=TRUE,
  verbose = 0, naTolerance=0.05)

`pigengene`	The pigengene object that is used to build the decision tree. See `pigengene-class`.
`Data`	The training expression data
`Labels`	Labels (condition types) for the (training) expression data. It is a named vector of characters. `Data` and `pigengene` will be subset according to these names.
`testD`	The test expression data, for example, from an independent dataset. Optional.
`testL`	Labels (condition types) for the (test) expression data. Optional.
`selectedFeatures`	A numeric vector determining the subset of eigengenes that should be used as potential predictors. By default ("All"), eigengenes for all modules are considered. See also `useMod0`.
`saveDir`	Where to save the plots of the tree(s).
`minPerLeaf`	Vector of integers. For each value, a tree will be built requiring at least that many nodes on each leaf. By default (`NULL`), several trees are built, one for each possible value between 2 and 10 percent of the number of samples.
`useMod0`	Boolean. Wether to allow the tree(s) to use the eigengene of module 0, which corresponds to the set of outlier, as a proper predictor.
`costRatio`	A numeric value effective only for 2 groups classification. The default value (1) considers the misclassification of both conditions as equally disadvatageous. Change this value to a larger or smaller value if you are more interested in the specificity of predictions for condition 1 or condition 2, respectively.
`toCompact`	An integer. The tree with this `minPerLeaf` value will be compacted (shrunk). Compacting in this context means reducing the number of required genes for the calculation of the relevant eigengenes and making the predictions using the tree. If `NULL` (default), the (persumably) most general proper tree (corresponding to the largest value in the `minPerLeaf` vector for which a tree could be constructed) is compacted. Set to `FALSE` to turn off compacting.
`noise, noiseRepNum`	For development purposes only. These parameters allow investigating the effect of gaussian noise in the expression data on the accurracy of the tree for test data.
`doHeat`	Boolean. Set to `FALSE` not to plot the heatmaps for faster comoutation.
`verbose`	The integer level of verbosity. 0 means silent and higher values produce more details of computation.
`naTolerance`	Upper threshold on the fraction of entries per gene that can be missing. Genes with a larger fraction of missing entries are ignored. For genes with smaller fraction of NA entries, the missing values are imputed from their average expression in the other samples. See `check.pigengene.input`.

This function passes the inut eigengenes and appropriate arguments C5.0 function from C50 package.

A list with following elements:

`call`	The call that created the results
`c5Trees`	A list, with one element of class `C5.0` for each attempted `minNodesperleaf` value. The list is named with the corresponding values as characters.
`minPerLeaf`	A numeric vector enumerating all of the attempted minPerLeaf values.
`compacted`	The full output of `compact.tree` function if `toCompact` is not `FALSE`
`heat`	The output of `module.heatmap` function for the full tree if `doHeat` is not `FALSE`
`heatCompact`	The output of `module.heatmap` function for the compacted tree if `toCompact` is not `FALSE`
`noisy`	The full output of `noise.analysiy` function if `noise` is not 0. For development and evaluation purposes only.
`leafLocs`	A matrix reporting the leaf for each sample on 1 row. The columns are named according to the correspoding `minNodesperleaf` value.
`toCompact`	Echos the `toCompact` input argument
`costs`	The cost matrix
`saveDir`	The directory where plots are saved in

For faster computation in an initial, explanatory run, turn off compacting, which can take a few minutes, with toCompact=FALSE.

Pigengene-package, compute.pigengene, compact.tree, C5.0, Pigengene-package

     ## Data:
     data(aml)
     data(mds)
     data(pigengene)
     d1 <- rbind(aml,mds)

     ## Fiting the trees:
     trees <- make.decision.tree(pigengene=pigengene, Data=d1,
       saveDir="trees", minPerLeaf=14:15, doHeat=FALSE,verbose=3,
       toCompact=15)